The generation of marine litter in Mediterranean island beaches as an effect of tourism and its mitigation

The Mediterranean Sea and its coastal systems are threatened by intense anthropogenic pressures including rapid accumulation of marine litter by diverse human activities. The region, which is the world’s leading touristic destination, has to face a seasonal increase of waste generation due to the seasonal influx of visitors. The beaches, extremely crowded during the summer, are particularly vulnerable since they are proven to be concentrated accumulation zones and one of the main gateways of litter to enter the marine system. We found that the accumulation rates of marine litter on Mediterranean island beaches follow a seasonal pattern, increasing up to 4.7 times during the high season, representing a daily load of (40.6 ± 11.5) 106 items/day extrapolated to all the islands of the region. We developed an accumulation index to assess the dynamics of marine litter and support efficient mitigation strategies by local authorities. To limit marine litter production attributable to recreational activities, a series of pilot actions implemented during the high touristic season, demonstrated a substantial reduction (up to 52.5%). The implementation towards an efficient and sustainable tourism business model is urgently required.


Scientific Reports
| (2020) 10:20326 | https://doi.org/10.1038/s41598-020-77225-5 www.nature.com/scientificreports/ The cleanliness of the beaches being one of the main factors considered by the tourists, along with the scenery, the safety, the facilities and the water quality 23 , marine litter can represent, in addition to its negative environmental effects, a shortfall 24 for local economies which often depend on this financial windfall. For example, in many countries, including those from the Mediterranean region, the absence of litter dictates visitors' choice (see Ref. 25 and references herein) and the probability to return to a given beach is strongly associated to the quality of the coastal environment 26 .
In this general context, it is then crucial to understand the seasonal dynamics of the waste generated by the visitors on the beaches in order to design effective solutions to mitigate this issue and move toward more sustainable tourism. Here we present the results of marine litter monitoring surveys performed on 24 beaches, going from remote to highly touristic sites, of Mediterranean islands during both the low and high touristic season of 2017. The results are compared to monitoring surveys performed on 11 of the same sites in 2019 after the implementation of pilot actions aiming at reducing the waste generated by the visitors.

Seasonal variation of beach litter in 2017.
In order to evaluate the seasonal variation of marine litter as an effect of tourism on sandy beaches of Mediterranean islands, 147 surveys were conducted in 2017 during both the low and high touristic season, from February to November. For each of the eight participating islands (Mallorca-Spain, Sicily-Italy, Rab-Croatia, Malta-Malta, Crete, Mykonos and Rhodes-Greece, and Cyprus), three different beaches were selected: a very touristic beach (T beach ), a beach mainly used by locals (L beach ) and a remote beach (R beach ) with limited use by humans ( Fig. 1 and "Methods"). For each beach, a periodic monitoring was performed on the same fixed 100 m portion 27 , covering the area going from the water line to the back of the beach 28 (Supplementary Fig. S1). Here, any item found was collected, characterized and properly disposed of. We included the smaller size classes of mesoplastics (MePs, 0.5-2.5 cm), large microplastics (MPs, 0.1-0.5 cm; i.e. visible to the naked eye), as well as pellets (raw plastic material), which usually are not targeted during this type of survey 27 .
A total of 162,320 items were collected. After correction for the distance and excluding one beach (see "Methods"), this represents an average of 1197.6 ± 2978.0 items per survey (or per 100 m of beach), 526.9 ± 794.2 items if we exclude the plastic fragments of less than 2.5 cm. These results fall in the same order of magnitude as others studies carried out on Mediterranean Sea beaches 11,[29][30][31][32][33][34] . The vast majority of the items collected belong to the artificial polymer materials (94.2%; Supplementary Fig. S2), i.e. items partly or completely made of plastic. The average number of items collected per survey depends on the type of beach considered: the L beach has an average of 2157.2 ± 4680.2 items/100 m (n = 49), followed by the T beach 899.9 ± 858.8 items/100 m (n = 49) and the R beach 425.4 ± 1319.9 items/100 m (n = 42) (Fig. 2). Interestingly, when only the number of items collected is considered, no clear seasonal variation can be observed, no matter the type of beach considered (Fig. 2). Moreover, the number of items collected per 100 m of beach is always higher during the low season compared to the high season (× 1.03 for the T beach , × 1.14 for the L beach and × 2.31 for the R beach ). These first results suggest that the seasonal increase of visitors on the beaches does not affect the amount of items of litter. This is counter intuitive as, for example, the amount of municipal solid waste generated during the high season on the island of Menorca (west Mediterranean Sea, Spain) is directly related to the number of tourists 35 .
For a given beach, by taking into account the time elapsed between the monitoring and the last cleaning performed by local authorities for the collected items larger than 5 mm, the time elapsed between two consecutive monitoring for the items smaller than 5 mm (i.e. MPs and pellets), the surface monitored and the number of items collected, it is possible to estimate the accumulation rates (AR) of the marine litter (see "Methods"). During the high touristic season, the monitoring was done on average 6.2 days after the last cleaning for the T beach , 13.2 days for the L beach and 27.8 days for the R beach , while during the low season, it was done respectively on average 61.2 days, 79.0 days and 105.3 days after. For some beaches, no cleaning activity was performed by the local authorities. In this case, we used the time elapsed between each monitoring to calculate the AR.
When the ARs are considered, a completely different picture emerges: the AR calculated for the three types of beach show an increase during the high touristic season (Fig. 3a). The T beach has the highest accumulation with an average of 329.6 ± 444.9 items/1000 m 2 /day (n = 33) during the high season, followed by the L beach , 177.2 ± 265.3 items/1000 m 2 /day (n = 33) and the R beach , 13.7 ± 32.2 items/1000 m 2 /day (n = 28). This represents an increase of the AR of + 471.9% for the T beach and + 346.0% for the L beach compared to the low season, when the AR are respectively of 69.8 ± 110.1 items/1000 m 2 /day (n = 16) and 51.2 ± 101.2 items/1000 m 2 /day (n = 16). For the R beach , the high season increase is concealed by the high AR recorded in October-November (Fig. 3a). The ARs for this specific period fall in the same range as for the T beach and the L beach . This could suggest that the impact of the visitors on the R beach during the high season is lower than the "natural" beaching of the litter during the low season, which is usually characterized by worse weather conditions. This could possibly result in an increased wash up of the marine litter onto the beaches, even if very little is known about the processes that govern its beaching 36 . Then for the R beach , although a seasonal pattern is perceptible, the AR during the high season [13.7 ± 32.2 items/1000 m 2 /day (n = 28)] are on average 32.7% lower compared to the low season [20.3 ± 59.7 items/1000 m 2 /day (n = 14)].
To our knowledge, very few studies provide the necessary information to estimate the AR of marine litter on beaches [37][38][39][40][41] . Three of them were carried out on islands with very low anthropogenic pressure compared to our study: on a remote Aleutian island 37 , a remote island of the Hawaiian archipelago 40 and a remote sub-Antarctic island 38 . In each case, the estimated ARs are up to 2-4 orders of magnitude lower than on the Mediterranean islands beaches, and up to 1-2 orders of magnitude lower if we consider beaches from the Mediterranean Sea 34 . For the study performed on a beach of eastern Australia 39 , the estimated ARs fall in the same range, highlighting the relationship between the number of items collected and the time elapsed since the last cleaning/monitoring. www.nature.com/scientificreports/ Most frequent items and MePs-MPs possible source. Once corrected for distance, in order to make all the selected sites comparable (see "Methods"), the five most frequent items collected, were the MePs (35.3%), cigarette butts (12.4%), the MPs (11.2%), pellets (9.5%) and macroplastics (> 2.5 cm; 8.3%). The possible sources were classified according to the eight major categories described for the Adriatic and Ionian seas 27 . When we look at the items from the "shoreline, including poor waste management, tourism and recreational activities" (ST category, i.e. the items most likely left on the beaches by the visitors), the five most frequent for the surveys of 2017 are cigarette butts (12.4%), plastic caps and lids (3.5%), cutlery, trays and straws (1.6%), crisp, sweet packets and lolly sticks (1.4%) and metal bottle caps (0.7%); in agreement with previous studies 29,30 . During the high season, these five items present as well the highest AR of the ST category no matter the type of beach considered, with the exception that for the R beach , the cutlery, trays and straws are replaced with metal drink cans ( Supplementary Fig. S3). Unsurprisingly, cigarette butts have the highest AR with values of 173.5, 41.3 and 1.6 items/1000 m 2 /day measured respectively at the T beach , the L beach and the R beach during the high season.
Since the origin of small fragments/pieces of plastics cannot be clearly defined, we attributed the pellets to the "shipping" category, meaning that they are lost at sea during transportation. The MePs and the MPs are usually assigned to the "non-sourced" category. However, we observed a strong correlation between the AR of the MePs and the AR of the items from the ST category (R 2 = 0.72, n = 116, p < 0.001), as well as between the AR of the MPs and the AR of the MePS (R 2 = 0.48, n = 87, p < 0.001) (Fig. 3b), suggesting that the accumulation of small plastic fragments increases with the accumulation of the items of the ST category on the monitored beaches. From this observation, we hypothesize that a non-negligible part of both MePs and MPs could be produced directly on General map of the Mediterranean Sea. Top: the 8 islands selected for this study. The graphic next to each island shows the normalized monthly number of tourists (Y axis, arbitrary scale) welcome in 2016 or 2017 (X-axis, from January to December) (Supplementary Table S1). Yellow bars: high touristic season (monthly number of tourists above the annual average), blue bars: low touristic season (monthly number of tourists below the annual average). Red dots: marine litter surveys. Below: map of each island with the selected beaches (T beach : yellow dots, L beach : blue dots and R beach : green dots). The general map of the Mediterranean Sea was generated with Ocean Data View (Schlitzer, R., Ocean Data View, https ://www.awi-breme rhave n.de/GEO/ODV, 2001). www.nature.com/scientificreports/ the beaches by the fragmentation of larger plastic items 42 . During the summer in particular, plastic debris on beaches will undergo thermo-oxidative degradation 43 by solar irradiance, and mechanical degradation 43,44 by friction with the sand, accelerated by the high volume of visitors (our hypothesis). Taken together, the AR of the items from the ST category, the MePs and the MPs show a clear seasonal pattern (Fig. 3a). This seasonal variation mimics the average number of tourists welcomed in the participating islands for the T beach and the L beach (Fig. 3a). During each month of the high season, the items from the ST category represent on average 65.7% ± 2.8% (T beach ), 39.8% ± 17.9% (L beach ) and 35.5% ± 28.7% (R beach ) of the items accumulating on the beaches, falling in the same range reported by previous studies 29,34 ; and they increase respectively up to 79.2% ± 4.5%, 77.9% ± 9.4% and 58.8% ± 19.1% if the MePs and the MPs are considered. The ARs of the other category items remain quite stable over the year (Fig. 3a). The highest AR for the MePs and MPs observed in the L beach could be explained by the longer time elapsed between the monitoring and the last cleaning, allowing more time for larger plastic debris to fragment. Accumulation index. We found that the use of a marine litter accumulation index (AI) is critical for assessing the impact of tourism on coastal sites and specifically on beaches. This AI, based on the AR, is divided into seven categories from extremely low to extremely high (Table 1). Contrary to the existing clean coast index 45 , which provides a snap shot of the cleanliness of a beach, the AI enables the assessment of marine litter dynamics. The AI can be applied at different time scales (month or season) and for general or specific items of marine litter. Here we present the marine litter AI, for each Mediterranean island and for each type of beach monitored in this study (Fig. 4). In all cases, the AI is always higher during the high season compared to the low season, no matter the type of beach considered. The AI can be used as an indicator for municipalities to identify, address and mitigate the seasonal variation of waste generation on beaches resulting from tourism, by highlighting for example the most frequent items abandoned and to raise awareness in order to change visitors' behavior.
Positive impact of awareness campaigns on marine litter AR. Following the results obtained in 2017, pilot actions were conducted during the 2019 high season on 11 beaches previously monitored: both the T beach and L beach of Sicily, Malta, Crete, Rhodes and Cyprus; and the T beach of Mallorca. Due to the very low AR measured during the 2017 high season in the R beach , no pilot actions were implemented there. The pilot actions consisted mainly in awareness campaigns 46 through the distribution of flyers and the installation of posters at the main entrances of the beaches showing the results obtained in 2017 and emphasizing the potential role of the visitors on beach litter generation (11 beaches), the availability of ashtrays (6 beaches) and the intervention of representatives informing the visitors of the marine litter issue and its prevention (9 beaches). Moreover, when nonexistent, new trash bins were installed on the beaches and the existent ones, when necessary, were adapted to collect both mixed and recycled waste, with bilingual indications (national language and English) (5 beaches).
The surveys were conducted in August or September 2019, between two weeks and two months after the beginning of the implementation of the pilot activities and on average 6.7 days (T beach ) and 35.1 days (L beach ) after the last cleaning of the beaches. The surveys were carried out on the exact same portions of beach as for 2017 and following the exact same procedure. However, it has been only possible to conduct one survey per site. This constraint made impossible to estimate the AR for the MPs and pellets, the smallest items, after the implementation of the pilot actions. With the exception of the L beach of Sicily which shows a more than sixfold increase of the AR of the items from the ST category, the 10 other sites show a decrease of their AR ranging from − 20.8% (T beach of Cyprus) to − 98.8% (L beach of Rhodes) (Supplementary Fig. S4).  (Fig. 5). For the T beach tested, the decrease of the AR of the ST category items is concomitant to a decrease of the AR of both MePs (− 31.3%) and remaining litter (− 43.6%, MPs and pellets being excluded), while for the L beach tested it is concomitant to an increase of the AR of both the MePs (+ 26.0%) and remaining litter (+ 106.7%, MPs and pellets being excluded) (Fig. 5). Regarding the the MePs, the increase of  Table S1). The decrease observed during August for the AR of the T beach and the L beach as well as for the average number of tourists is related to the absence of monitoring performed on the island of Mallorca, which is the most visited island and one of the most affected by marine litter. (B) Comparison between the AR of the MePs and the AR of the items from ST category (yellow); and between the AR of the MPs and the AR of the MePs (grey). The solid yellow and grey lines represent the power regression for both comparisons (both X and Y axes are on a logarithmic scale). www.nature.com/scientificreports/ the AR on the L beach could again be explained by a much longer time elapsed between the last cleaning activities and the surveys. However, the difference regarding the other remaining items (i.e. not from the ST category or the MePs), cannot be explained by our results. Although it is possible to imagine that local/regional differences about the beaching of the litter, especially for islands from distinct sub-basins of the Mediterranean Sea, could play a significant role. According to the results from 2017, cigarette butts are the most common item left on the beaches by the visitors. Ashtrays were then made available for the visitors on the 2 selected beaches of Malta, Crete and Cyprus. The accumulation rates of cigarette butts for these 6 beaches decreased on average by 54.5% (± 12.0%) after the implementation of the pilot actions. For the beaches where no ashtrays were made available (T beach of Mallorca, Sicily and Rhodes; L beach of Rhodes), we observed a decrease of 57.8% (± 28.2%) between 2017 and 2019. No significant difference between the two groups of beaches is observed (p-value > 0.05). We do not interpret this as an ineffectiveness of the ashtrays to reduce the number of cigarette butts thrown on the beach, but rather as an  www.nature.com/scientificreports/ increase of the proportion of visitors who possibly refrain from smoking on the beach or who properly disposed of the cigarette butts whether ashtrays are available or not. Interestingly, the only beach which shows an increase of the AR of the cigarette butts in particular (8×) or of the items of the ST category in general (6×) was the L beach of Sicily where no representative was present to inform the visitors and no ashtrays were made available.
Regarding the four other most abundant items, they all show a decrease of their AR on the tested T beach : − 51.8% for the plastic caps and lids, − 74.0% for the cutlery, trays and straws, − 59.5% for the crisp, sweet packets and lolly sticks, and − 40.6% for the metal bottle caps. For the L beach considered (excluding Sicily), with an exception for the crisp, sweet packets and lolly sticks which shows an increase of + 153.8% (2 items collected in 2017 vs. 59 in 2019 for these 3 beaches), the other items present a similar decrease of their AR by − 50.7% (plastic caps and lids), − 65.0% (cutlery, trays and straws) and − 42.4% (metal bottle caps) (Supplementary Fig. S5).
The recreational use of beaches as a main source for marine litter. Massive tourism in Mediterranean islands and specifically in their beaches is a main driver of marine litter generation. Popular beaches (T beach and L beach ) are clearly the most impacted sites of this study. Every day, during the high touristic season peak (July-August), on every 100 m of beach, visitors will leave on average 844.5 (T beach ) and 295.2 (L beach ) items from the ST category. This represent respectively 71.0% and 31.1% of the total amount of items composing the marine litter during this specific period. If the MePs and the MPs are included, resulting from the fragmentation of larger plastic items as we hypothesized, then visitors could generate in total 1028.4 items (T beach ) and 798.3 items (L beach ), representing this time respectively up to 86.4% and 84.0% of the total amount of items composing the marine litter during this specific period. Once again it is worth noting that on average 2.7 times more MePs and MPs are generated on the L beach compared to the T beach during the high season peak. To our point of view it is clear that the frequency of the cleaning of the beaches by the local authorities play an important role regarding the prevention of the generation of MPs and MePs as the T beach are cleaned on average 2.1 times as often as the L beach . For the R beach , tourism could be responsible of the accumulation of 57.7 items per 100 m of beach per day, meaning that the remote sites are on average 14-20 times less impacted than the most popular beaches.
Even if the T beach and L beach are frequently cleaned during the high season, the portion of the marine litter accumulating on these beaches and reaching the sea, although hardly assessable, could represent a tremendous daily load as Mediterranean islands account for about 19 000 km of coastline 47 . At the region scale, we estimated that during the high season peak (July-August), visitors could be responsible for the accumulation of about 40.6 10 6 ± 11.5 10 6 items/day on the beaches of the Mediterranean islands ("Methods"). As a matter of fact, if all these items were to reach the marine environment, this would represent the equivalent of a daily accumulation of 2500 ± 709 items/km along the Mediterranean islands coastline. Although hypothetical, this value is quite realistic when compared to concentrations of floating plastic in the nearshore water strip (i.e. < 1 km off the coasts) 48 ranging from 28,000 to 578,000 items/km 2 .
Awareness campaigns, a promising tool to reduce beach littering. Our results show that awareness campaigns could be an efficient tool to reduce the amount of litter resulting from the recreational use of the Mediterranean island beaches. Indeed, we observed an average decrease of 52.5% ± 20.8% of the AR of the items from the ST category after the implementation of the pilot actions. These very encouraging results probably benefit from the growing attention of the public to the plastic pollution in the oceans or the measures adopted by the European Commission to reduce marine litter (e.g. the single-use plastic directive). However, this reduction of the marine litter has a price: to achieve these results, the average cost of the pilot actions, extrapolated to the whole high season (May-September) would be of 111.6 k€ per km of beach 49 . This cost could be restrictive for local authorities to implement such actions.
The number of visitors in the Mediterranean region is projected to increase rapidly with more than 500 million expected tourists by 2030 50 . There is an urgent need for a tourism lifestyle change and environmental education of key actors acting on the knowledge gaps of policy-making processes and on the investment in sustainable tourism. Proper monitoring of tourism sustainability is a strategic driver for the economic development and sustainability of islands that depends on ecosystem services already threatened by climate change.

Methods
Sampling strategy. In order to assess the seasonal variation of marine litter as an effect of tourism on Mediterranean islands, 24 beaches (3 per island) have been monitored for marine litter during both the low and high seasons. For each selected site, a fixed 100 m portion of beach (i.e. sampling unit) was defined, as recommended in most beach litter monitoring methodologies 27,28,51,52 , and going from the waterline to the back of the beach as the visitors can occupy any point within this zone. Moreover, the sampling unit should be connected, when possible, to at least one access to the beach around which the density of visitors is usually the highest, especially for the T beach and L beach . As on some islands the beaches can be very small (< 200 m) and narrow (< 10 m), the monitoring of more than one sampling unit 27,51 could not be implemented and the monitoring of few (very short) random transects 52 within one sampling unit possibly not representative. By taking into account all these points, we decided to use the OSPAR methodology 28 for the 24 selected beaches and adapt it in order to include the MePS, the MPs and the pellets.
Once defined, the fixed 100 m portions have been periodically monitored for marine litter: once a month during the high touristic season (from May to September) and one time before (February-April) and after (October-November) the high season. Although, the same fixed portion was always monitored for each beach, it appears that for some of them, the surface monitored slightly changed from one survey to another, depending on various factors such as the accuracy of the GPS, the presence of vehicles or the crowd on the beach. The highest variations were observed in August and September for the T beach of Malta where only one third of the www.nature.com/scientificreports/ defined surface was monitored. During the monitoring, all the items with an anthropogenic origin found on the sand were collected, counted and characterized 28 . During the surveys, the small plastic pieces including the mesoplastics (MePs, 0.5-2.5 cm), the large microplastics (MPs, 0.1-0.5 cm) and the pellets (raw plastic material) were collected at the same time as the larger marine litter items and on the same portion of beach. The surveyors had the instruction to only collect the small items laying directly on the surface of the sand (i.e. plastic debris/pellet buried below the surface of the sand were not monitored). Moreover, during a monitoring, there was one surveyor assigned to the collection of these small plastic items who was helped by the others surveyors once the bigger marine litter items were collected. The small pieces of plastic were then shipped to our institute, sorted according to their size and assigned to one of the following categories: macroplastics (> 2.5 cm), MePS (0.5-2.5 cm), MPs (< 0.5 cm) or pellets. Note that for the pellets, all the particles recognized as such were counted as pellets, no matter their size.
For each island, the 3 different beaches were selected as follows in order to include (1) an impact site (T beach ) where tourists represent most of the visitors to the beach, (2) a control site (L beach ) where locals represent most of the visitors and (3) a remote/preserved site (R beach ) where the frequency of both tourists and locals is low. T beach are characterized by a high volume of visitors, mainly tourists, especially during the high touristic season. In the area situated directly behind the beach high levels of infrastructure dedicated to tourism and recreational activities, such as hotels, restaurants, bars, souvenir shops, etc., are found. L beach are characterized by high volume of visitors, mainly locals, especially during the high touristic season. However, in the area situated directly behind the beach low level of infrastructures dedicated to tourism and recreational activities, such as hotels, restaurants, bars, souvenir shops, etc., are found. R beach are assumed to be the less impacted type of beach. They are characterized by low volume of visitors, even during the high touristic season. In the area situated directly behind the beach no infrastructure dedicated to tourism and recreational activities are found.
Regarding this last type of beach, although 8 beaches were monitored in 2017, we had to remove the results obtained from Malta. The volume of visitors, although reduced, cannot be considered comparable to other remote beaches of this study. It is situated in the municipality of Marsaxlokk and a harbor is bordering the beach to the west. Moreover the total number of items collected on this beach (4249) represents more than 20% of the total number of items collected on all the R beach (20,738), and once calculated the accumulation rates of marine litter for this specific beach appears to be 60 times higher (956.6 items/1000 m 2 /day) than the average accumulation rates calculated for the others R beach (15.9 items/1000 m 2 /day), making the R beach of Malta not suitable for the analysis.
Data compilation and analysis. For each of the 24 selected beaches a fixed 100 m portion of beach was defined. The coordinates of the starting point and the ending point of the portion were recorded and the distance between the two points was calculated with the following equation: where D is the calculated distance in m, Lat S is the starting latitude (in decimal degree), Long S the starting longitude (in decimal degree), Lat E the ending latitude (in decimal degree) and Long E the ending longitude (in decimal degree). The functions cos −1 returns the arccosine of an angle, cos returns the cosine of an angle, sin returns the sine of an angle and rad converts degrees into radians. When D is not equal to 100 m (most of the cases) the number of items collected on the portion were corrected for the distance in order to normalize all the data for all the beaches. The following equation was used to correct the data: where NI cor is the number of items corrected for distance, NI col the number of items collected on the fixed portion of beach and D the calculated distance from Eq. (1). Note that the number of items corrected for distance is only used to show the number of items per 100 m of beach: it is not used for the calculation of the accumulation rates or the accumulation index.
Finally, for each beach, the surface of the 100 m fixed portion was estimated by using the "Polygon" tool from Google Earth. For each beach, the considered surface was measured from the water line to the back of the beach, between the starting and the ending points of the portion.
The accumulation rates of the marine litter is used here to estimate the accumulation of marine litter or of a given item per unit of surface and per unit of time. In this study the accumulation rates are given in number of items/1000 m 2 /day. It is calculated as follows: where AR is the accumulation rate in number of items/1000 m 2 /day, NI col the number of items collected on the fixed portion of beach (not corrected for the distance), S the surface of the fixed portion of beach in m 2 and T the time elapsed between the survey and the last cleaning activity in days. The time elapsed since the last cleaning corresponds to the number of days elapsed since the last cleaning performed by the local authorities (or the previous survey when no cleaning activities were conducted by the local authorities). The AR of the MPs and pellets were always calculated by taking into account the time elapsed between each survey and not the time elapsed since the last cleaning, as due to their small size we assumed these particles cannot be removed even with mechanical cleaning. Moreover, for the very first survey performed on each beach, the AR of the MPs and the pellets were arbitrarily calculated by using the number of days elapsed between the survey and the 1st of January 2017 as it was not possible to estimate since when these particles were accumulating on the beaches. Then the average number of days used for the calculation was 83.6, 83.6 and 87.3 respectively for the T beach , the L beach and the R beach .