Introduction

Plastic pollution has become one of the most significant challenges currently facing the marine and coastal environments (Barnes et al. 2009; Lebreton et al. 2018; Lebreton and Andrady 2019). It has become a serious cause of mortality for marine animals and has contributed to the degradation of marine and coastal ecosystems (Allsopp et al. 2006; GESAMP 2015). In 2021, global production was 390.7 million tonnes (Plastics Europe 2022), an increase of 4% on the previous year. In the absence of any coordinated and strategic efforts aimed at decreasing production, it is probable that the 2015 estimates of 4.8 and 12.7 million tonnes of plastic waste entering the oceans annually (Jambeck et al. 2015) have been surpassed. Findings by The Pew Charitable Trusts & SYSTEMIQ (2020) estimate that the amount of plastic waste entering the oceans could triple by 2040.

Of particular concern are microplastics (MPs), small plastic particles less than 5 mm in diameter, first defined by Thompson et al. (2004), and mesoplastics (MEPs) (5–25 mm) (GESAMP 2015). When plastics of this size enter the environment, they present a risk to a wide variety of marine and coastal organisms through ingestion (Lavers et al. 2014; Provencher et al. 2017). Increasing concern about the impacts of these plastics has been the catalyst for an increase in environmental awareness campaigns and research (Filho et al. 2020; Ho 2020). Despite this, there is still much we don’t know about the scale, distribution, sources and pathways of MP pollution in the ocean and on coastlines (Browne et al. 2010; Imhof et al. 2017; Reisser et al. 2013; Ryan et al. 2009). Furthermore, public understanding and concern of MPs, and the implications of MPs with regards to the environment and human health remains low (Deng et al. 2020; Henderson and Green 2020; Khatibi et al. 2021).

The GESAMP (2015) report recommended a need to identify the main sources and categories of plastics and MPs entering the ocean, and a requirement to promote a greater awareness of their impacts. These recommendations remain extant. One reason is the limitations of our research capacity (Horton 2022) and the scale of the pollution. Small teams of researchers, capable of gathering moderate amounts of data over long periods of time in geographically limited areas, are falling behind the ever-increasing scale of the problem. These shortcomings and a growing worldwide motivation to tackle the plastic pollution problem (Hunt 2017) present an opportunity for the development of a citizen science (CS) based approach that could simultaneously collect useful scientific data and raise awareness.

CS research can be a cost-effective way to augment the information collected through more traditional scientific research. It can also play a role in providing positive benefits for volunteers and communities through public engagement, enhancing scientific literacy, developing more sustainable lifestyles, and changing attitudes (Anderson and Wentworth 2014; Cigliano et al. 2015; Cooper et al. 2007; Pocock et al. 2014). Although more frequently associated with terrestrial research, CS has been used successfully in several marine environmental research programmes. In Japan for example, the International Pellet Watch (IPW) organisation has been utilising citizen scientists to gather pre-production pellets from coastlines around the world. IPW use these to monitor the pollution status of the oceans focussing on persistent organic pollutants (POPs) (Ogata et al. 2009). In Australia, Smith and Edgar (2014) worked alongside 300 volunteer divers to document the scale of sub-tidal marine debris in 120 sites along the coast of New South Wales, Australia. At the time when this research project commenced there were no CS projects investigating MPs and MEPs in coastal areas, although there are now others in operation such as the Australian Microplastic Assessment Project (https://ausmap.org) and the European Union (EU) funded Plastic Pirates project (https://www.plastic-pirates.eu), which aims to empower young people to collect data about plastic and microplastic pollution across Europe.

Whilst there are clear benefits to CS research, there are also many challenges associated with it (Druschke and Seltze 2012). Data produced by volunteers are still often viewed with scepticism by some scientists and accuracy, consistency, and reliability are essential if there is to be confidence in the results (Aceves-Bueno et al. 2017; Burgess et al. 2017; Riesch and Potter 2014). In order to have an impact and be successful, CS projects need one or more outcomes that include collecting useful scientific data, providing participants with new skills or knowledge, influencing policy, or having a positive effect on communities (Shirk et al. 2012). Strategies for volunteer engagement, communication, the type and quality of data, and barriers to participation also need to be considered (Dickinson et al. 2012; Haklay et al. 2006).

In this study we designed, developed, and implemented a CS project to enable volunteers to undertake coastal surveys for MPs and MEPs. We adapted existing protocols in a way that would simultaneously meet the scientific requirements and address the numerous challenges and limitations of CS research. The protocol was tested in pre-launch field trials conducted on a local site on the south coast of the UK before an initial launch to a wider global audience. Some samples from that initial launch were re-analysed by the research team and a volunteer survey was conducted to examine participant motivations and potential areas for project improvement. The method was subsequently adjusted prior to digitisation and final project implementation. The project became a long-term, global CS research programme known as the ‘Big Microplastic Survey’ (BMS). This paper describes the development of the project and assesses its success.

Materials and methods

Protocol design

A review of existing literature regarding MP and MEP sampling methods was undertaken which highlighted myriad approaches, a lack of consistency and no standard operating protocol (Claessens et al. 2013; Hidalgo-Ruz et al. 2012; Smith and Edgar 2014). Additionally, many of the existing processes required specialist training, equipment and techniques that are often unsuitable for volunteers (Geoghegan et al. 2016; Masura et al. 2015; Ryan et al. 2009). This literature review was used to produce a project design framework (Fig. 1) and to develop our protocol with regards to sampling methodology, to suit the context of a CS based approach.

Fig. 1
figure 1

Project design framework schematic, showing key phases (grey) and inputs into the process (orange) leading to final implementation (black)

Full size image

Sampling methodology

The starting point for the sampling methodology was the recommendations of the European Union’s Marine Strategy Framework Directive’s (MSFD) Guidance of Monitoring of Marine Litter in European Seas (Hanke et al. 2013). The MSFD protocol, suitable for both MPs and MEPs, recommended a minimum sampling requirement of five replicate 0.5 m2 quadrat samples to a depth of 5 cm. Samples are collected from the strandline with each being separated by at least 5 m across a 100 m stretch of coastline. The coordinates of each quadrat are recorded, and samples are removed for further laboratory processing and analysis. Löder and Gerdts (2015) adapted the MSFD methodology in their case study by replacing the 0.5 m quadrat samples with five 10 × 10 cm within every 1.0 m2 quadrat of sediment, sampled to a depth of 1 cm (Fig. 2).

Fig. 2
figure 2

Representation of adapted MSFD recommendation based on the Löder and Gerdts (2015) case study

Full size image

The variation in sample depth between the MSFD method and the case study of Löder and Gerdts (2015) was reflected in other sampling methods that showed a range of sample depths from 2 to 32 cm (Besley et al. 2017; Hidalgo-Ruz et al. 2012; Lee et al. 2015). In the absence of any consensus, a decision was made to collect samples to a depth of 2 cm for the project protocol as this would: (1) reduce the time taken for sample collection which was an important consideration; (2) account for the dynamic nature of coastal sediments and ensure that some buried MPs and MEPs would be collected; and (3) reduce the impact of buried marine vegetation that was known to be a feature at some of the sample locations of the field trial.

Sample processing

An on-site, sea-water density separation method was used to separate MPs and MEPs from the sediment (Hidalgo-Ruz et al. 2012). This method was decided on for two reasons: (1) laboratory facilities for the separation and analysis of samples were unlikely to be available to volunteers; and (2) the removal of sediment from the trial study location was not permitted due to its Site of Special Scientific Interest (SSSI) designation. Samples were collected and sediment was placed into a bucket containing seawater and stirred. Floating plastic debris and any organic material was then passed through a 1 mm mesh sieve. The contents of each sample were placed into a metal container before being dried and counted. GPS coordinates of the first quadrat were recorded. Each sample took approximately 30 min to complete. Once dry, organic material was removed from the samples by hand and remaining MPs and MEPs were counted, and their characteristics recorded.

Sample analysis

MP studies typically focus on quantifying abundance in marine and coastal environments (Fauziah et al. 2015; Jambeck et al. 2015; Thompson 2015), yet additional characteristics can be easily described. In a comprehensive analysis of 68 studies, Hidalgo-Ruz et al. (2012) identified key characteristics of source, type, shape, erosion, colour, and size. The ‘source’ of MPs was defined as either primary (industrial pre-production pellets also known as ‘nurdles’) or secondary (fragments of consumer products), however this definition excludes other types of primary plastic. In the GESAMP (2016) definition, the distinction between primary and secondary MPs relates to whether the plastic piece is of commensurate size to when it was manufactured; this distinction between primary and secondary was retained for the sampling protocol developed in this study.

Hidalgo-Ruz et al. (2012) further defined plastics by a ‘type’ characteristic (fragments, pellets, filaments, foamed plastic, granules and styrofoam) whereas Provencher et al. (2017) grouped both source and type characteristics into the same categories. We aimed to use the pre-launch trial to determine categories for “type” suitable for describing plastic within the study area. Seven easily recognisable and distinct categories were identified: pre-production pellets, bio-beadsFootnote 1, expanded polystyrene (balls and pieces), monofilament fishing line, cotton bud sticks and straws; these were retained as permitted values for the “type” characteristic.

Hidalgo-Ruz et al. (2012) further identified shape as a useful characteristic to record as it could provide information to the researcher about the fragmentation process and the amount of time the plastic might have been in the environment. These data were not the focus of this study, however, during the pre-launch trial it was noted that spherical, cylindrical, disk and cube shaped pre-production pellets were regularly encountered at the field trial location. These shapes were included in the sample categorisation as a ‘sub-characteristic’ of pre-production pellets. Erosion was considered too subjective and was not included in the analysis (Karapanagioti 2018; Veerasingam et al. 2016).

Whilst some research has regarded colour as incidental, Hidalgo-Ruz et al. (2012) and Provencher et al. (2017) considered it a useful characteristic worth recording as it allowed for some preliminary identification opportunities and could also provide insights into ingestion patterns. Rather than using complex colour permutations or colour wheels, it was decided that colours would be limited to those identified during the pre-launch trial.

Size characteristics are generally accepted within the scientific community as < 5 mm for MPs and 5–25 mm for MEPs. The minimum size of 1 mm was deemed appropriate for volunteers as it reduced the need for specialised equipment and was comparable to the size of the mesh found in household sieves (Geoghegan et al. 2016; Riesch and Potter 2014). It was decided that the maximum size of plastics for data would be 25 mm as this coincided with existing data on the size of plastics that can be easily ingested by a significant number of coastal and marine animals (Arthur et al. 2009; English et al. 2015; Van Franeker et al. 2011).

Pre-launch field trail

To provide proof of concept for the sampling and analysis process, a pre-launch field trial was conducted at the entrance of Chichester Harbour, which forms a coastal border between the counties of West Sussex and Hampshire on the south coast of the United Kingdom. The harbour is highly designated and is classified as an Area of Outstanding Natural Beauty (AONB), Ramsar site, SSSI, Special Area of Conservation (SAC) and a Special Protection Area (SPA).

Eight locations for sampling were selected based on ease of access for volunteers and sediment suitability (Fig. 3). Participants for the trial were recruited through social media, and by word of mouth from local community groups and marine science students from the University of Portsmouth, providing an engaged volunteer base (Eveleigh et al. 2014; Seymour and Haklay 2017). A project-specific information webpage was produced by the principal investigator (PI) and published on a marine conservation charity website (https://justoneocean.org/chichester-harbour-microplastic-survey). Volunteer registration was undertaken using the ‘Eventbrite’ online events website. On the survey dates, volunteers were directed to specific sampling locations where they collected samples from visible strandlinesFootnote 2. MPs are generally found at the strandlines created by subsequent high tides (Fig. 3) with the highest mean particle number and weight being found at the highest strandline on the shore, often referred to as the wrack zone (Lee et al. 2015). On two of the survey dates there were sufficient volunteers and low enough tides to collect samples from four strandlines across the intertidal range (Fig. 4). On the other two dates samples were collected from a single wrack zone at the top of the beach.

Fig. 3
figure 3

Survey locations for the pre-launch field trial undertaken within Chichester Harbour AONB

Full size image
Fig. 4
figure 4

Source: Author’s own

Visible strandlines and the wrack zone at one of the pre-launch trial sites

Full size image

Volunteers were provided with the equipment needed to conduct the survey and clearly written guidelines. A demonstration of the sampling protocol was given on the days of the surveys to ensure standardisation (Freitag et al. 2016). Sampling was undertaken over four dates covering a three-month period (14 March – 12 July 2018). Analysis of the samples was undertaken by some of the volunteers and the PI.

Initial launch

Data collection

Following the pre-launch field trial, the BMS project was presented to a wider audience in the initial launch phase. A social media campaign directed volunteers to a purpose-built project website (https://microplasticsurvey.org). The site provided additional supporting videos, information, and an online registration form. Project specific social media pages for Facebook, Twitter and Instagram were developed and used for volunteer recruitment and the dissemination of project information. Volunteers who registered through the website were sent a project resource pack by email that included sampling methods and data forms to be completed and returned. Volunteers returned the completed data forms by email and through the postal system. Volunteers were also asked to submit a photograph of their sample for verification and invited to send their samples to the research team for further analysis.

Analysis of volunteer samples

A total of 86 samples were received from volunteers and 40 of these were randomly selected and analysed for accuracy. This was undertaken by examining percentage agreement of the submitted data against verified counts using source, type, shape, and colour variables (Aceves-Bueno et al. 2017). The count data were not normally distributed, and a Wilcoxon signed ranks test was used to establish whether there was a significant difference in the abundances of different types of MPs counted by volunteers and the verifier. In addition to accuracy testing, 30 samples were analysed using a Bruker benchtop Attenuated Total Reflectance (ATR) Fourier Transform Infrared (FTIR) scanner for raw material verification.

Volunteer survey analysis

An online survey was produced using the Survey Monkey website (Survey Monkey Inc, n.d.) to examine participant motivations and identify areas for project improvement. The survey was sent to 587 participants and comprised 24 questions grouped into four main categories: pre-existing knowledge; motivation; future intentions; and socio-demographic data. Questions were multiple choice except for one question that asked how participants felt the project might inspire them to undertake more surveys in the future. The results of this open question were analysed using NVivo 12 analytical software (QSR International) to develop response themes (Braun and Clarke 2006).

Final project design and further digitisation

Analyses of the pre-launch trial analysis of volunteer samples and volunteer survey led to changes being made to the protocol that could be incorporated into the final project design:

  • The colour ‘other’ was removed (high degree of subjectivity).

  • Colour options were simplified; participants were directed to choose a colour that most closely represented the colour of the plastic.

  • Rope, fishing line and cotton bud sticks were removed as a specific characteristic due to data being submitted that was inconsistent with the size characteristics of MP and MEP options. These items could still be submitted if they were within the size range and were further characterised as secondary MPs or MEPs.

  • Size categories were clearly defined as MPs (1–5 mm) and MEPs (5–25 mm) in any one direction.

  • Primary and secondary MEPs were added to the characteristics.

  • The resource pack was updated to reflect changes in the data characteristics.

  • The BMS website was updated with the addition of a downloadable version of the updated resource pack. Additional supporting videos were also added.

  • An online data submission portal was introduced that enabled volunteers to upload their data directly to a geographical information system (GIS) database. This included a requirement to submit an image of the sample.

  • A real time ‘results’ page was developed on the website with interactive maps that provided a visual representation of the MP and MEP data collected.

Results

Pre-launch field trial

A total of 102 surveys were undertaken during the pre-launch field trial by 96 volunteers. The volunteers collected a total of 1550 plastic pieces (Table 1). There were very similar quantities of primary and secondary plastics found (786:764) with pre-production pellets making up most of the primary plastics (n = 679) and other plastics accounting for the majority of secondary MPs and MEPs (n = 640). Clear was the most common colour encountered. In one survey location (survey site number 28) there were particularly high concentrations of MPs and MEPs, a total count 685 plastic pieces, which included 416 pre-production pellets.

Table 1 Summary of total counts and colours submitted by volunteers during the pre-launch trial phase. Samples collected from total of 5.1 m3 of sediment
Full size table

Data were collected from the intertidal zone on two of the survey dates with 92 samples being taken (Table 2). Total counts of MP and MEPs from four strandlines across the intertidal zone showed that abundance increased with distance from the sea.

Table 2 Data showing total MP and MEP counts from samples taken across the intertidal zone. Zone A was the highest strandline (wrack zone) and Zone D was the lowest strandline. Total sediment volume collected was 4.6 m3
Full size table

Initial launch phase

Following the pre-launch trial, the initial launch saw a total of 579 online registrations to the project from 53 different countries during the 12 months of the initial launch phase. Of those registrations, 65 were from universities and other educational institutions. There were also registrations from government agencies, environmental NGOs, commercial organisations, water sports clubs, scouting organisations, and individuals. A total of 53 volunteers (individuals and organisations) undertook one or more surveys. A total of 151 surveys were collected and analysed over the period in 18 different countries collecting a total of 5,849 pieces of plastic. Samples were extracted from a total sediment volume of 7.55 m3. Secondary MPs and MEPs were the most common types of plastic pollution encountered followed by pre-production pellets (Table 3). The most common colours were white and clear, followed by blue and green. The colour option described as ‘other’ resulted in numerous variations; colour descriptions were often highly subjective and included teal, turquoise, green/brown, mixed, and striped. The categorisation of rope, cotton buds and fishing line appeared to cause some confusion amongst volunteers with lengths of rope and fishing line longer than 25 mm being included in the data because they were less than 25 mm in diameter. This also applied to cotton bud sticks and straws.

Table 3 Summary of total counts and colours submitted by volunteers during the initial launch phase
Full size table

Quality assurance results

Of the 40 samples that were re-analysed for verification purposes, 99% produced identical counts to those submitted by the volunteers. There were no statistically significant differences in the total abundances or the abundances of the different types of plastics (p ≥ 0.66), except for ‘other’ primary and secondary MPs (p ≤ 0.044), where volunteers occasionally misidentified plastic types. The most common discrepancy was misidentification of source (primary or secondary plastics) which occurred 11 times. Four of the samples differed due to subjective errors in determining the shape of pre-production pellets. Other discrepancies were due to organic matter being incorrectly identified as a plastic, and an oversized piece of plastic being included in the data.

Of the 30 samples that were analysed by FTIR scanner for raw material verification, the plastics collected were predominantly expanded polystyrene (203 pieces), polyethylene (221 pieces) and polypropylene (79 pieces). All three types of polymers have a lower density than seawater and could be expected to be found in strandlines. In addition, there were three pieces of polyester, five pieces of nylon monofilament fishing line and one piece of PVC, all of which had a density greater than seawater. One pre-production pellet was identified as polybutene. A total of 30 pieces could not be identified due to uncertain scanner results or excessive contamination of the samples.

Volunteer survey results

A total of 224 returns (38.16% response rate) were received from the volunteer survey. A proportion of those respondents submitted socio-demographic details (Table 4).

Table 4 Socio-demographic information submitted by respondents to volunteer survey
Full size table

Most of the volunteers were female (81%), married (52%) and with a university level of education (undergraduate 40%, postgraduate 41%), or in full-time employment (45%). Half of the respondents (50%) were in the 35–54 age group.

A complete summary of responses to the other questions in the survey are available in SI, however some key data stood out. Most of the volunteers were either extremely or moderately aware of the MP issue (85%) prior to being involved in the BMS project. Only 8% were either unaware or only slightly aware. Most respondents had been aware of the problem for between 2 and 5 years (61%) and 73% felt that there was not enough information in the media about the MP problem. The impact of MPs on the marine environment was reported as the most important concern of participants (92%). Participants were more concerned with the impacts on marine organisms (birds, mammals, fish, and molluscs) (82–92%%), than they were about impacts on human health (65%).

The social media campaign was responsible for recruiting 45% of participants and word of mouth accounted for 17%. A total of 88 respondents also listed ‘other’ as the reason for volunteering and an analysis of their responses showed that 35% were actively searching for citizen science projects, or MP research to be involved in. Motivation for involvement included concern for the environment (82%), being interested in the subject (61%), and wanting to make a difference (60%). Only 20% listed personal development as a motivation. At the time the volunteer survey was sent out, 57% had not yet conducted a BMS survey. When questioned, 6% of those who responded had forgotten to do it, 15% had not found the time and 23% were planning to complete it in the future. Only 33% of participants submitted their survey results, although 21% of the remaining registrants indicated that they planned to do so later.

Design questions were also focussed on identifying areas for improvement. Most participants (83%) agreed that the resource pack and supporting videos provided enough information to plan a survey, and 61% found the videos helpful. When asked how the resource pack could be improved, 48% felt it was adequate, but 42% felt that a digital option should be provided, and 11% indicated that the resource pack needed to be available in different languages.

The final design question was an open-ended question that asked, “What could we do to inspire you to do more surveys?” There was a total of 103 responses and four key themes were identified: (1) Feedback and communication (n = 15). Comments included the need for reminders and email notification, and recognition that the data had been received; (2) Publication and sharing of data (n = 12). For example, one volunteer wrote, “Have a map of the data, so you can see the impact your surveys are having in contributing to the knowledge of MPs.”; (3) Easier data submission (n = 6). Respondents requested electronic or online methods to record data; and (4) Different protocols (n = 7). Respondents identified the need to develop protocols for other types of sediments, in particular pebbles and shingle shorelines. Some had not understood that the methods could be used for riverbanks and lakesides. Other suggestions included providing the equipment, providing certificates of completion for younger volunteers, introducing volunteer networks, and translating the guidelines into other languages. After participating in the BMS study, 72% of respondents felt they knew more about MPs and their impact on the environment than they did previously and 86% said they intended to carry out additional sampling. Nearly all respondents (98%) said that they would recommend the project to friends and colleagues.

Final project implementation (BMS)

As of 1 July 2023, there have been 948 registrations from 64 countries. This includes 114 educational establishments and 214 other types of organisations including government agencies, environmental NGOs, commercial organisations, water sports clubs, and scouting organisations. A total of 910 samples have been collected and analysed by 174 volunteers (individuals and organisations) who undertook one or more surveys from 32 countries, including locations overseen and regulated by the Antarctic Treaty System. A total of 51,852 plastic pieces have been collected (Table 5) with almost half of that number being pre-production pellets. White was the most common colour found and again a high proportion (> 50%) of those white plastics were pre-production pellets. Cube pre-production pellets were the least common type of plastics found.

Table 5 Summary of total counts and colours submitted by volunteers. Total sediment volume collected was 4.55 m3
Full size table

In many locations, data have been collected over an extended period at a variety of locations providing opportunities for further temporal and spatial analyses. For example, in Koh Tao Island, Thailand, 109 surveys have been undertaken around the island over a three-year period. Regional patterns and variations are also beginning to emerge, such as a higher prevalence of EPS found in countries within SE Asia. Out of 146 surveys in the region there were 8,047 EPS balls (89.5% of the total) and 1,731 EPS pieces (61.5% of the total) found.

Discussion

The development of CS projects generally follows a common structure: define the project aims; design and test protocols; promote the project and collect the data; report the results (Pocock et al. 2014). This research followed a similar structure, underpinning the process with a user-centred design approach(McKinley et al. 2017). That said, the success or failure of a CS research project is dependent on many things and there are numerous challenges (Druschke and Seltze 2012). From a scientific perspective, accurate, consistent, and reliable results that can be verified are essential to overcome scepticism amongst the scientific community and ensure there is confidence in the methodology and the data collected (Riesch and Potter 2014; Sheppard and Terveen 2011). Various methods have been used to achieve this in the past such as crowdsourcing (Haklay 2017; Watson and Floridi 2018), and instrumental observation using specialised scientific instruments. This project relied on a process-oriented approach; a method often used in conservation projects. Supervision and training, or the provision of clear guidelines and task simplification are needed with this type of approach (Riesch and Potter 2014) and in this research project we relied on the latter, developing clear protocols, simple and logical data submission formats, and educational guidelines through videos, briefings and supporting documents (Bonney et al. 2014).

These challenges associated with CS projects led to the decision to undertake a pre-launch trial as part of the design process (Druschke and Seltze 2012). The purpose was to provide proof of concept, to: (1) determine whether large amounts of MP and MEP data could be collected by volunteers without the need for specialised training and equipment in a reasonably short space of time; (2) determine whether the protocol was suitable for collecting and analysing plastic samples in coastal sediments; and (3) assess the enthusiasm amongst volunteers to participate in this type of research.

During the pre-launch trial, we succeeded in collecting a large amount of data (1,550 plastic pieces from 102 surveys; Table 1) using volunteers (n = 96), over a short timeframe and with minimal training and instruction. Volunteers were able to collect samples for analysis without the need for the removal of large amounts of sediment or laboratory equipment. The two volunteer surveys within the pre-launch trial provided the opportunity to collect data from the intertidal zone and our results supported the findings that microplastics tend to accumulate higher up the beach and in strandlines, although patterns within a single site can be inconsistent (Table 2) (Browne et al. 2010; Joint Research Centre Institute for Environment and Sustainability 2011). Additionally, the pre-launch trial provided information that was used to identify discrete categories for selected characteristics of MP and MEPs that were subsequently used during the initial launch phase.

Overall, volunteer recruitment trends demonstrated enthusiasm amongst the public to be involved in CS projects of this type. When asked why they had registered for the project the majority expressed concern for the environment and interest in the subject and a desire to be involved. The pre-launch trial generated a large amount of local media interest and was featured in newspapers and on television news, publicity that is invaluable for volunteer recruitment and retention and, ultimately, the continued success of CS projects over time. This publicity, coupled with the development of a dedicated website and social media sites (Twitter, Facebook and Instagram) was instrumental in recruitment for the initial launch.

The initial launch provided an intermediary step between the pre-launch trial and the final project design and implementation that provided information for possible improvement in the project. During this phase of the project, we had 579 registrations from 53 different countries within a 12-month period and subsequently 151 surveys were undertaken by 53 volunteers (either individuals or organisations) providing a registration (recruitment) to engagement rate of 9.1%. Given the high-profile nature and increased awareness surrounding the plastic pollution issue (Gell 2019; Hunt 2017; Joyce 2018), it could be argued that this figure is relatively low. However, traditionally within CS, a large proportion of results are provided by a small number of participants (Eveleigh et al. 2014). This participation inequality usually results in what Nielsen (2006) referred to as the 90-9-1 rule, where 90% of participants contribute nothing or very little, 9% of the volunteers provide 10% of the contributions and 1% account for most of the data.

In the initial launch we re-analysed and compared 40 samples submitted by volunteers and by comparison found that the total count aligned 99% of the time. The main variations in precision were due to subjective analysis of the shape of pre-production pellets (20% of variation between volunteers and re-analysis) and misidentification of primary and secondary materials (55% of variation between volunteers and re-analysis). By comparisonAceves-Bueno et al. (2017) found that 55.2% of results submitted had an agreement level between data collected by volunteers and scientists of 80% or higher. The high level of agreement in our study was probably due to the simplicity of the sample characteristics being recorded (identified during the pre-launch trial, thus minimising the opportunities for subjective assessment. Variations occurred in those characteristics where there were opportunities for ambiguity and subjectivity. The detail of those differences resulted in changes to the data characteristics of the protocol including a simplification of colour options, the removal of certain types of plastics (rope, fishing line and cotton buds) and the separation of MP and MEP data for both primary and secondary plastics.

The volunteer survey provided useful information regarding the motivation of volunteers, socio-demographic details, and design considerations. There were a higher number of female (81%) than male respondents which is contrary to the findings of Pateman et al. (2021) and may have been due to variations in employment between men and women (Fetterolf 2017).We found that 35% of participants were actively searching for citizen science or microplastic projects that they could be involved in, and that over 50% of respondents found the survey through social media and extremely important recruitment tool for CS projects as previously found by Oliveira et al. (2021).

The level of knowledge of MPs (85% were extremely or moderately aware) was higher than that found in members of the public. In their research Deng et al. (2020) found that 67% of the public had not heard of MPs. This significant difference in results may just reflect variations in the level of awareness between the volunteers in this study and those in Shanghai. Alternatively, it may have been the result of high number of graduates and postgraduates that participated in the project, or a reflection of the number of registrations (n = 65) from educational establishments. The precise reason is uncertain from the data available, however, given the motivation responses amongst participants it may be intuitive that a highly motivated sub-population will be more knowledgeable on the subject matter than the general population.

Following survey responses, the resource pack, which 83% of respondents felt provided adequate information, was updated to reflect the changes in data characteristics previously mentioned. A total of 42% of participants expressed a desire for a digital resource pack option and there were also comments from respondents (n = 12) expressing a desire for the publication and sharing of data. In addition, of the 23 comments received on how the resource pack could be improved, 10 of the respondents said they did not receive the resource pack. An online download option was included, and a digital data submission process with a real-time results portal displayed using a Geographical Information System (GIS) database developed by the PI.

Following final implementation of the BMS project, registration surpassed expectations: with 948 registrations from 64 countries. The total number of surveys undertaken was 910 which at 30 min per survey (para. 2.1.2) is 455 man-hours of data collection, or 56 days working on an eight-hour day. Even without considering the wide geographical area such an approach can access, the benefits of CS for data collection are evident. The actual number of participants who completed one or more surveys when compared to total registrations was 174 (18.7%) which is an increase on the results of the initial launch phase and an improvement on Nielsen’s (2006) norms, but lower than the engagement rate of 35% experienced by De Moor et al. (2019) in their study of the dynamics of engagement within citizen science projects. The reason for this increase cannot be stated with certainty but is probably the result of an increased social media presence, of which De Moor is an avid supporter, and the simplification that the digitisation of the process provided. Although this conversion rate may appear low, attracting high numbers of volunteers remains important as it serves to increase the profile of the project and provides active volunteers an opportunity to compare their results and activities to others. ‘Word of mouth’ is also an important consideration and our survey found it was the reason 17% of respondents joined the project. As De Moor et al. (2019) state there is a need for a “critical mass” of volunteers for the “core” to function.

The data that have been collected to date are already providing some interesting results. In contrast to the initial launch the most frequent type of plastic pollution found to date has been pre-production pellets. On analysis it is apparent that the main reason for this difference was a total count of 18,825 pre-production pellets collected during 18 surveys around Harlingen in the Netherlands. These results followed the MCS Zoe shipping incident (https://www.bbc.co.uk/news/world-europe-46746312) when lost cargo included shipping containers carrying pre-production pellets. The most common colours from the total count were white and clear which can also be accounted for to some extent by the loss of cargo from the MCS Zoe. Another pattern emerging is the higher prevalence of polystyrene in Southeast Asia than other parts of the world. Full analysis of the temporo-spatial trends in these data is outside the scope of this paper, but future hypotheses to explain trends in these data may provide insight to the shifting patterns of global microplastics following marine and industrial accidents, natural disasters, or other matters of interest. The data collected in such a manner can then be a fulcrum upon which the negative externalities of microplastic pollution can be fully understood from a public health and/or economic perspective.

The aim of this research was to design, develop, and implement a CS project to enable volunteers to undertake coastal surveys for MPs and MEPs and subsequently increase understanding of the issues. Volunteers have contributed significantly to our knowledge of the distribution, types, and sources of coastal MPs and MEPs and they have gained new skills and a better understanding of the plastic pollution issue in the process. This type of public engagement can have a positive impact and increase awareness of the marine and coastal plastic pollution issue in general (Jones et al. 2013; Locritani et al. 2019; Wiggins et al. 2014). However, the BMS project has a wider relevance within environmental research. The Driver-Pressure-State-Impact-Responses (DPSIR) framework (Smeets et al. 1999) introduced by the European Environment Agency as a means of monitoring the progress of environmental policy targets through indicators remains relevant today. Environmental CS projects serve to highlight the relationship between humans and their environment and have relevance within the DPSIR framework by monitoring (pressure and state) and identifying long term trends through data collection (impact). This wider relevance is evident in the use of the BMS protocol and data for scientific reports, undergraduate and postgraduate research projects and the numerous NGOs that are now working in collaboration with the BMS project. The project is by no means perfect; further research needs to be conducted to examine retention rates amongst volunteers now that there is some temporal data emerging. We also need to understand better the discrepancy between registration numbers and survey participants (MacLeod and Scott 2021) and there are still some areas of ambiguity so further work is required. Additionally, there is a need to compare the method with others used in CS studies that have since been developed, as well as accepted scientific methods. This will allow us to assess whether the data produced are comparable and accurate. CS is ultimately a trade-off between data quality and data quantity, so it is almost inevitable that there is going to be a degree of subjectivity when volunteers are describing MPs and MEPs. However, those errors should not preclude the use of citizen science approaches from further scientific study. To do so is missing an opportunity for wider public engagement and future scientific discovery (Lukyanenko et al. 2016); the crucial thing is to reduce the errors as much as possible. Given the amount of data collected and the high level of public engagement, we believe this research has been a success (Cox et al. 2015) and as registrations and the number of samples being submitted to the database continue to grow, the BMS project will remain a useful resource for scientific research into the future.

Conclusion

Plastic pollution continues to have a detrimental impact on marine and coastal ecosystems, biodiversity and many other socio-economic aspects of the communities living in coastal environments (Brouwer et al. 2017). However, despite an increase in research and awareness there are still large gaps in our knowledge and a large proportion of the public that remains unaware, or unconcerned (Henderson and Green 2020; Khatibi et al. 2021). This research has demonstrated that a CS approach to collecting MP and MEP data could be used to address these shortcomings. The project used a structured, user-centred approach in development, and the adaptation of existing sampling methodologies to suit the limitations and challenges of CS. An initial launch phase was critical to the success of this project, providing information about data characteristics, volunteer motivation and design considerations that allowed for improvement of the final design. Another important aspect has been social media which has been crucial in the recruitment of volunteers and remains so moving forward. The BMS protocol is now being used around the world and data from the project is augmenting that collected by more traditional scientific methods. Additionally, the public engagement has been far reaching and supports the existing evidence that CS can provide a win-win situation, where local communities can benefit, and scientific knowledge can be advanced on a scale that might not otherwise be achievable. With no end-date, the opportunities that this project presents for further research are considerable.