High levels of microplastics and microrubber pollution in a remote, protected Mediterranean Cladocora caespitosa coral bed

High le els o mic oplas ics and mic o ubbe pollu ion in a emo e, p o ec ed Medi e anean Cladoco a caespi osa co al bed La s Reuning a,* , La s Hildeb and b , Diego K. Ke s ing c , Daniel P ¨ o ock b a Ins i u e o Geosciences, CAU Kiel Uni e si y, Ludewig-Meyn-S . 10, 24118 Kiel, Ge many b Depa men o Ino ganic En i onmen al Chemis y, Helmhol z-Zen um He eon, Max-Planck-S aße 1, 21502 Gees hach , Ge many c Global Change, Conse a ion and Gene ics o Ma ine Species, Ins i u o de Acuicul u a To e de la Sal, Consejo Supe io de In es igaciones Cien í icas (CSIC), To e de la Sal S/N, 12595 Ribe a de Cabanes, Spain ABSTRACT Co al ee s a e inc easingly h ea ened by an h opogenic s esso s, including plas ic pollu ion. This s udy in es iga es he abundance and possible ecological impac o mic oplas ics (MPs) and mic o ubbe pollu ion in sedimen s om a Cladoco a caespi osa co al bed in he no h-wes e n Medi e anean. Despi e being loca ed in a emo e ma ine p o ec ed a ea wi h no local plas ic pollu ion sou ces, ou esul s indica e excep ionally high MP concen a ions (mean: 1514 pa icles/kg d y weigh ), a ibu ed o long-dis ance anspo o plas ics by he No he n Cu en . Lase Di ec s In a ed (LDIR) Chemical Imaging and ATR-FTIR spec oscopy we e used o cha ac e ize he MPs in e ms o size, shape and polyme ypes. Mos MPs a e agmen s (96 %), while ibe s con ibu e only 4 %. The mos abundan polyme s we e polye hylene (PE, 28 %), polye hylene e eph hala e (PET, 25 %), and polys y ene (PS, 19 %), wi h signi ican con ibu ions om polyu e hane (PU) and mic o- ubbe . Pa icle size analysis showed ha 92 % o MPs we e smalle han 250 μ m, wi h a median pa icle size a ying by polyme ype. No ably, polyme s wi h he e oa oms in hei main chain, such as PET and polyu e hane, exhibi ed signi ican ly smalle median sizes compa ed o polyole ins, possibly sugges ing di e en deg ada ion pa hways. The high MP concen a ions measu ed in sedimen s wi hin co al colonies sugges s ha MPs could ha e ad e se e ec s on he e o ophic eeding in C. caespi osa, a c i ical ene gy sou ce du ing s ess e en s. This s udy unde sco es he u gen need o a ge ed esea ch on MP e ec s on he esilience o C. caespi osa and o inc eased global and egional e o s o cu b plas ic pollu ion mi iga ion in o de o conse e co al popula ions in he Medi e anean. 1. In oduc ion Co als o m s uc u ally complex habi a s ha os e biodi e si y and p o ide impo an ecosys em se ices essen ial o human socie y (Hughes e al., 2002; Pendle on e al., 2016). The mone a y alue o hese ecosys em se ices is ac ually highe han ha o any o he biome on ea h (G oo e al., 2012). I is he e o e ala ming ha co al habi a s a e in decline. The global co e age o opical co al ee s and hei abili y o p o ide ecosys em se ices has declined by hal since he 1950s (Eddy e al., 2021) while co al biocons uc ions in mid-la i ude and empe a e seas, such as he Medi e anean, ha e been con inu- ously declining in ecen decades (Ke s ing e al., 2013; Ke s ing e al., 2022). In ac , hese ecosys ems a e globally he mos a ec ed by an h opogenic ocean wa ming (Bindo e al., 2019; Ga abou e al., 2009; Ga abou e al., 2022; Hughes e al., 2017). O he s ess ac o s such as chemical pollu ion a e known o in e ac syne gis ically wi h clima e change and exace ba e co al loss (Dono an e al., 2021). One s ess ac o ha has gained inc easing a en ion in ecen yea s is mic oplas ics (MPs) pollu ion (Hall e al., 2015; Pan os, 2022; Reiche e al., 2018; Saliu e al., 2019). Plas ic pa icles <5 mm in size a e called MPs (A hu e al., 2009). P ima y MPs a e in en ionally p oduced small plas ic pa icles, ypically o use in manu ac u ing o consume p oduc s. Seconda y MPs, on he o he hand, a e c ea ed when la ge plas ic i ems b eak down in o smalle pa icles due o na u al wea he ing p ocesses. Seconda y MPs can come om a a ie y o sou ces, including disca ded plas ic bags, bo les, and packaging, as well as ishing ne s and o he ma ine deb is. The uncon olled bu ning o was e on beaches makes he plas ic e en mo e b i le and suscep ible o he o ma ion o MPs (U ami e al., 2023). In labo a o y expe imen s, exposu e o co als o mic oplas ics has led o ad e se e ec s such as bleaching, nec osis and educ ion in g ow h a es (Hankins e al., 2021). Long- e m exposu e expe imen s using ealis ic mic oplas ic concen a ions con i med i s ha mic oplas ic pollu ion can ha e species speci ic nega i e impac s on ee co als (Reiche e al., 2019) and second, ha common ee -building co als canno adop o long- e m mic oplas ic exposu e (Rades e al., 2022). In addi ion, MPs can p omo e pa hogen ansmission, u he inc easing he suscep ibili y o ee -building co als o disease (Ki s ein e al., 2016). MPs can also ac as ec o s o hea y me als, ei he as addi i es in he plas ic o abso bed o i s su ace (Hildeb and e al., * Co esponding au ho . E-mail add ess: [email p o ec ed] (L. Reuning). Con en s lis s a ailable a ScienceDi ec Ma ine Pollu ion Bulle in jou nal homepage: www.else ie .com/loca e/ma polbul h ps://doi.o g/10.1016/j.ma polbul.2025.118070 Recei ed 10 Ma ch 2025; Recei ed in e ised o m 28 Ap il 2025; Accep ed 28 Ap il 2025 Ma ine Pollu ion Bulle in 217 (2025) 118070 A ailable online 5 May 2025 0025-326X/© 2025 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY license ( h p://c ea i ecommons.o g/licenses/by/4.0/ ). 2021; Pa e son e al., 2020). Recen s udies on he accele a ed wea h- e ing o di e en consume plas ics e ealed se e al 1000 chemical ea u es du ing non- a ge chemical analysis o igina ing om he UV deg ada ion o he di e en o ganic addi i es p esen in he es ed ma- e ials (Menge e al., 2024). Recen ield s udies ha e u he shown ha mic oplas ic concen- a ions in he coas al wa e s can be high enough o be ecologically ele an o co al ee heal h (Tang e al., 2021). They obse ed ha mic oplas ic concen a ions in ee co als co ela e nega i ely o hei symbion densi y, indica ing a nega i e e ec on he co al- Symbiodiniaceae symbiosis. In he long e m, MPs he e o e pose a h ea o ee co als and hei abili y o unc ion as amewo k builde s in co al ee sys ems. The Uni ed Na ions En i onmen P og am, he e o e conside s unde s and- ing plas ic pollu ion a ound co al ee s a key knowledge gap in ee ecosys em esea ch (Swee e al., 2019). This has spa ked a g owing in e es in he ole o plas ic, and especially mic oplas ic pollu ion, as a s esso o co al ee s (Biswas e al., 2024; Huang e al., 2021; John e al., 2022; U ami e al., 2021; U ami e al., 2023). Co al ee sys ems a e an iconic coas al ecosys em o he opics, bu also occu a highe la i udes. Cladoco a caespi osa is he only scle - ac inian, zooxan hella e co al wi h ee -building capaci y in he empe a e Medi e anean Sea (Mo i e al., 1994; Pei ano e al., 2001). This co al plays a key ole in o ming he ben hic habi a in he coas al zone o he Medi e anean and Ad ia ic Sea. Pei ano e al. (1998) desc ibed wo ypes o C. caespi osa colony dis ibu ions: beds and banks. Beds a e composed o a la ge numbe o dis inc subsphe ical colonies (e. g., Gul o T ies e, Schille (1993)), while banks a e made up o la ge colonies, eaching se e al decime e s in heigh and co e ing se e al squa e me e s in su ace a ea (e.g. Mjle Na ional Pa k, K uˇ zi´ c and Benko i´ c (2008)). Mixed dis ibu ions o beds and banks can also occu (e.g. Columb e es Islands, Ke s ing and Lina es (2012)). In addi ion, C. caespi osa can also occu as ee-li ing co al nodules o co alli hs, usually smalle han 10 cm in diame e (Ke s ing e al., 2017a, 2017b). Mos ee -building co als acqui e hei ene gy mainly h ough pho osyn hesis. This p ocess is acili a ed by symbio ic algae known as zooxan hellae, which con e sunligh in o nu ien s o main ain he co als me abolism. He e o ophic eeding on e.g. plank onic zooplank on is used as a supplemen a y ene gy sou ce by many opical co al species (Houlb ` eque and Fe ie -Pag` es, 2009), pa icula ly du ing s ess e en s when pho osyn hesis may be comp omised (Fe ie -Pag` es e al., 2010; G o oli e al., 2006). The abili y o use bo h eeding s a- egies is called he e o ophic plas ici y. C. caespi osa is pa icula ly dependen on he e o ophic plas ici y, as i co e s a la ge pa o i s ene gy demands in win e h ough he e o ophic eeding (Fe ie -Pag` es e al., 2011). Simila o i s opical coun e pa s, wa ming-induced mo ali y om ma ine hea wa es is he main h ea o C. caespi osa (Ke s ing e al., 2013). Se e al opical co al ee species espond o he mal s ess by inc easing hei ood in ake, which can make hem mo e esilien o bleaching (G o oli e al., 2006). This has also been hypo hesized o C. caespi osa (Quin ano Fe nandez e al., 2024). On he o he hand, a highe eeding a e could also inc ease he isk o unwan ed in e ac ions wi h MPs. Co als can ap and inges MPs due o i s simila i y in size o plank on (Hall e al., 2015; Reiche e al., 2018). A labo a o y s udy sugges s ha he e o ophic eeding is a majo ac o con ibu ing o he impac o MPs, wi h high eac ion and inges ion a es in highly he - e o ophic species (Reiche e al., 2024a). In ac , he e o ophic eeding has been sugges ed as he pa hway o inco po a ion o o he kind o an h opogenic pa icles ( ly ash) in C. caespi osa skele ons (Robe s e al., 2024). T emblay e al. (2011) epo ed ha C. caespi osa has a g ea e capaci y o he e o ophy han o he scle ac inian symbio ic co als and Robe s e al. (2024) sugges ed ha hese ai s may a o he up ake o mic o-pa icles by C. caespi osa. Inges ing MPs ins ead o ood could impac co al esilience by educing he e ec i eness o eeding as an al e na i e ene gy sou ce du ing hea -induced bleaching. Addi ionally, plas ic associa ed con aminan s (e.g. ph halic acid es e s) wi h po en ial ad e se e ec s on co al heal h (Saliu e al., 2019), ha e been de ec ed in Medi e anean an hozoans, including he highly he - e o ophic C. caespi osa (Gobba o e al., 2024). Medi e anean co al biocons uc ions o med by C. caespi osa he e o e likely ace mul iple s esso s no only om an h opogenic wa ming bu also plas ic pollu ion. The no h-wes e n Medi e anean is a ho spo o ma ine li e (So o- Na a o e al., 2020). The concen a ion o loa ing plas ic deb is in he a ea is one o he highes in he wo ld i aling he amous plas ic ga bage pa ches in he sub opical gy es (C´ oza e al., 2015). This egion ha bo s one o he la ges C. caespi osa biocons uc ions in he Medi- e anean Sea, loca ed in he Columb e es Islands Ma ine Rese e. In his si e, C. caespi osa o ms a mixed dis ibu ion o beds and banks wi h a o al co al co e o 2900 m 2 (Ke s ing and Lina es, 2012). This emo e a chipelago is conside ed a global change sen inel si e, hos ing decades long moni o ing on co al heal h and empe a u e (Ke s ing e al., 2013; Ke s ing and Lina es, 2019). We s udied he sedimen s om his C. caespi osa popula ion o analyze how s ongly hei habi a is a ec ed by mic oplas ic pollu ion. The samples we e aken om he sea loo adjacen o C. caespi osa colonies and om sedimen s apped wi hin he colonies. In pa icula , he s udy aims no only o analyze he concen- a ion o mic oplas ics in he sedimen , bu also o cha ac e ize he size and polyme ype o all mic oplas ics ound in he samples. The ela i e abundance o he polyme ypes and hei concen a ion in he sedimen a e used o calcula e pollu ion isk indices ha help o e alua e hei en i onmen al impac . In he s udy, he pa icle-size dis ibu ion o he mic oplas ics is examined in pa icula , as i has been shown ha he size o he mic oplas ics can in luence hei in e ac ion wi h he co als (Hankins e al., 2022; Hankins e al., 2021). In combina ion hese da a p o ide impo an insigh s in o he ole o mic oplas ics as a po en ial s esso o one o he la ges biocons uc ions o he scle ac inian symbio ic co al C. caespi osa. 2. Ma e ials and me hods 2.1. S udy si e Illa G ossa (39◦53.825 ′ N, 0◦41.214 ′ E), he la ges o he Columb e es Islands (0.14 km 2 ), is a subme ged Qua e na y olcanic calde a (Mu˜ noz e al., 2005). I is loca ed on he Spanish ou e Medi e anean shel , app oxima ely 55 km o he Spanish coas (Fig. 1). The C-shaped island is open o he no heas in he main di ec ion o win e s o m wa es (Ke s ing and Lina es, 2012) and he No he n Cu en (Ou mie es e al., 2023), which lows pa allel o he slope along he con inen al ma gin o he Ibe ian Peninsula (Fig. 1). The e a e no beaches; he ocky slopes o he isle d op s eeply in o he bay o a minimum wa e dep h o 5 m. The a e age dep h o he bay is 15 m, wi h a maximum dep h a 30 m. The co al colonies o m beds and banks in he ocky bo oms o his semi- enclosed bay (Fig. 2), wi h highes co al co e in a dep h be ween 10 and 20 m. The co al colony size a ies be ween 5 and 150 cm, wi h a mean colony age o ~50 yea s, based on a e age g ow h a es (Ke s ing and Lina es, 2012). Co al co e exceeds 10 % in he no hwes e n and sou heas e n pa s o he bay (Fig. 2, Fig. S1), whe e he co als a e ela i e p o ec ed om he s ong NE s o ms in au umn and win e . These wo a eas a e sepa a ed by a NE–SW unning cen al channel whe e co al co e is gene ally below 2 % (Fig. 2). These cen al and deepe pa s o he bay a e co e ed wi h sand and g a el om alkaline olcanic ock and bioclas s (Ke s ing and Lina es, 2012). The s udied bay is pa o he Columb e es Island Ma ine Rese e, a 5400-ha ma ine p o ec ed a ea which was es ablished in 1990 o p o ec he a ea om di ec an h opogenic in luences. The only ec ea ional ac i i y pe mi ed in Illa G ossa Bay is scuba di ing, bu i equi es au ho iza ion om he ma ine ese e managemen au ho i ies. The s udy a ea he e o e is ideal o es o he i s ime he le el o mic oplas ic pollu ion on a emo e, p o ec ed Medi e anean co al ecosys em. L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 2 2.2. Con amina ion mi iga ion A s ic p o ocol was ollowed in o de o p e en con amina ion om he sampling equipmen , labo a o y equipmen , eagen s, clo hing and ai bo ne sou ces. This included he igo ous use o lamina low benches class II, il a ion o all eagen s, he use o cleaned me al o glass labo a o y equipmen and he conduc ion o me hod blanks. Howe e , no MPs in he in es iga ed size ange we e ound in he blank es ha was pe o med o e alua e he backg ound alues o MPs. A mo e de ailed desc ip ion o he conduc ed con amina ion mi iga ion p ocedu es can be ound in Hildeb and e al. (2022b). 2.3. Sample collec ion Fi e sedimen samples we e collec ed by scuba di ing in he bay o Illa G ossa be ween 2017 and Oc obe 2022. Two sedimen samples we e eco e ed om ~10 cm high, li ing C. caespi osa colonies om a wa e dep h o 15 m (Fig. 2, Table 1). Based on an a e age g ow h a e o 3.2 mm/yea (Ve go i e al., 2025), he co al colonies ep esen a g ow h his o y o ~30 yea s s a ing in he mid-1980s.The ela i ely open, phaceloid g ow h o m o C. caespi osa colonies (Fig. 2) is a o - able o he accumula ion o sedimen be ween he indi idual co alli es. The sample size was limi ed by he amoun o sedimen con ained in he co al colonies (11 and 61 g d y weigh (dw)). Th ee addi ional sedimen samples o abou 200 g dw each we e collec ed om he sea loo (0–5 cm dep h in sedimen ) in wa e dep h be ween 17 and 21 m (Table 1). Two o hese samples we e aken adjacen o C. caespi osa colonies in a eas wi h a o al co al co e be ween 10 and 11 %, while one sample was aken in he a ea o spa se co al co e (1 o 2 %) in he NE–SW cen al channel o he bay (Fig. 2, Table 1). All samples we e aken o he labo a o y a CAU Kiel and eeze-d ied o cons an weigh . 2.4. G ain-size analysis D ied samples we e sie ed by hand h ough 1 mm, 2 mm, 4 mm and 8 mm sie es (Re sch GmbH, Ge many) a CAU Kiel. All indi idual g ain- size ac ions we e weighed. Mean g ain size and so ing we e calcula ed using he so wa e G adis a (Blo and Pye, 2001). The so ing o samples was quan i ied as g aphic s anda d de ia ion using he equa ion o Folk and Wa d (1957). A e g ain-size analysis he samples we e spli in wo ac ions (<1 mm and >1 mm) o u he p ocessing. 2.5. S a is ical analysis Two non-pa ame ic s a is ical es s we e used o analyze he pa icle-size dis ibu ions o all polyme ypes. The Mann-Whi ney pai wise U es was used o es i he median pa icle size o di e en polyme s is he same. The Kolmogo o -Smi no wo-sample es was used o es i he pa icles om di e en polyme s show he same size dis ibu ion. A Mon e Ca lo Simula ion was used o e alua e he likely Fig. 1. Loca ion o Illa G ossa (yellow s a ), he la ges o he Columb e es Islands in he no h-wes e n Medi e anean Sea. Illa G ossa is loca ed in he pa hway o he No he n Cu en , he mos impo an bounda y cu en o he no h-wes e n Medi e anean. I lows pa allel o he slope along he con inen al ma gin o he Ibe ian Peninsula. No hwa d lowing, seconda y cu en s connec he Balea ic Sea wi h he Alge ian Subbasin (Amo es e al., 2013). In e annual a ia ions in he egional oceanog aphy a e caused by mesoscale ea u es (dashed line), such as eddies (Ou mie es e al., 2023). Base map c ea ed wi h GeoMapApp (www.geom apapp.o g)/CC BY using he Global Mul i-Resolu ion Topog aphy (GMRT) syn hesis (Ryan e al., 2009). The inse shows a sa elli e image o Illa G ossa, a C-sha- ped d owned olcanic calde a open o he no heas in he main di ec ion o win e s o m wa es and he No he n Cu en . Sa elli e image om Google Ea h©. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he web e sion o his a icle.) L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 3 ange o haza d sco es o he polyme g oup PU/ac yla e/ a nish (see below). The PAST 4.0 so wa e (Hamme e al., 2001) was used o all s a is ical analyses, wi h he excep ion o he Mon e Ca lo Simula ion, which was ca ied ou in Mic oso Excel 2019. 2.6. P ocessing o >1 mm ac ion The >1 mm ac ion was checked o MPs unde a binocula mic o- scope (Zeiss, Disco e y.V8, Ge many). La ge MPs we e iden i ied ollowing p o ocols de eloped by Lushe e al. (2017). Unna u al colo s and/o shininess and unna u al o ms/s uc u es we e used as indica o s o po en ial MPs. Pa icles wi h po en ially cellula o o ganic s uc u es and anslucen ibe s we e ejec ed as MPs. T anslucen ibe s and i- be s ha we e no cha ac e ized by h ee-dimensional bending and uni o m hickness we e also ejec ed (Ma in e al., 2017). All suspec ed MPs in he >1 mm ac ion we e pho og aphed and hei size measu ed on he digi al pho og aphs using he so wa e Fiji (Schindelin e al., 2012). Fig. 2. Map and pho os o C. caespi osa co al banks in Illa G ossa Bay. A) Map o co al co e (%) in Illa G ossa Bay (modi ied om Ke s ing and Lina es (2012), no e ha he isolines in his map e lec s he co al co e alues, no he sea- loo mo phology). The inse a he op igh shows an enla gemen o he sampling a ea nea he la ges co al banks in he sou h-wes o he bay. The sampling si es a e ma ked as yellow s a s. B) La ge (~ 100 cm in diame e ) C. caespi osa colony nea sampling si es IG4 and 5. C) The sea loo a a wa e dep h o ~17 m nea sampling si e IG3 consis s mainly o sandy g a el om alkaline olcanic ock (black) and bioclas s (whi e). The wid h o he ield o iew in he o eg ound o he image is ~20 cm. D) Close-up o a C. caespi osa colony showing a ela i ely open, phaceloid g ow h o m wi h sedimen apped be ween he indi idual co alli es (blue a ow). The ypical diame e o C. caespi osa co alli es in his image is 4–5 mm. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he web e sion o his a icle.) Table 1 Sampling da e, loca ion, en i onmen al pa ame e s and sedimen ex u e a sampling si es. Sample Da e Coo dina es Wa e dep h (m) Co al co e Sedimen ex u e So ing G ain size (mm) IG1 Oc . 2022 39.89635◦; 0.68643◦21 1–2 % Sandy G a el poo 1.5 IG2 Oc . 2022 39.89585◦; 0.68668◦17 10–11 % G a elly sand mode a e 1.3 IG3 Ap . 2019 39.89579◦; 0.68674◦17 10–11 % G a elly sand poo 1.2 IG4 Aug. 2017 39.89574◦; 0.68667◦15 co al colony G a elly sand poo 0.9 IG5 Aug. 2017 39.89572◦; 0.68671◦15 co al colony Sandy G a el poo 1.7 L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 4 2.7. Polyme cha ac e iza ion (>1 mm ac ion) Po en ial la ge MPs (>1 mm) we e analyzed a he Helmhol z- Zen um He eon by a enua ed o al e lec ance - Fou ie ans o m in a ed (ATR-FTIR) spec oscopy (Alpha I, diamond ATR c ys al, B uke Dal onik GmbH, B emen, Ge many). Measu emen s we e pe - o med h ee imes wi h 32 scans and a esolu ion o 4 cm −1 (wa e- numbe ange: 4000 cm −1 –400 cm −1 ). Spec a we e compa ed wi h e e ence spec a om he siMPle lib a y de eloped by P impke e al. (2020b, 2020a). Spec al assignmen s ( ec o -no malized i s de- i a i es) wi h a hi quali y index (HQI) ≥700/1000 we e accep ed. 2.8. P ocessing o <1 mm ac ion MPs we e p ocessed a CAU Kiel by ca bona e dissolu ion ollowed by densi y sepa a ion and o ganic ma e diges ion. To p e en CO 2 ou gassing in he acidic lo a ion media, he ca bona e clas s in he <1 mm ac ion we e dissol ed using hyd ochlo ic acid (10 % / ) un il he eac ion s opped. Subsequen ly, up o 30 g o d ied sedimen was mixed wi h 80 ml CaCl 2 (densi y: 1.5 g/cm 3 ) sal solu ion in a beake . CaCl 2 - solu ion has he ad an age o being non- oxic and ela i ely cheap. Sedimen and solu ion we e s i ed h ee imes o 10 min and a e - wa ds le o se le o 1 h un il he supe na an was clea . Floa ing solids we e sepa a ed ia o e low om he supe na an , collec ed and ans e ed o a cellulose ni a e il e wi h a po e size o 0.45 μ m. These sepa a ion s eps we e epea ed o 30 g ba ches o sedimen un il he en i e sample was p ocessed. The eco e y a e o his me hod was moni o ed in he labo a o y a CAU Kiel using a i icially spiked sedi- men samples. Pa icles o common polyme ypes (PP, HDPE, LDPE, PS, PVC) co e ing a la ge densi y (0.9–1.38 g/cm 3 ) and size (100 μ m o 1 mm) ange we e mixed wi h p e-cleaned sand. Using he me hod desc ibed abo e, he eco e y a e o MPs was 97 ±5 %. A Helmhol z- Zen um He eon, wo samples we e subjec ed o an addi ional ea men wi h H 2 O 2 solu ion (30 %, / ) a 40 ◦C o 24 h, o oxidize excess o ganic ma e . The suspensions (V (50 % e hanol) =80–100 ml) we e p e-concen a ed (T hea e co e =40 ◦C, T hea e base =40 ◦C) o a inal olume o 1 ml using a Synco e-Plus® au oma ed e apo a ion sys em in conjunc ion wi h a EasyFill Rack R-12 Poly ap, a Vacuum Co e R-12 wi h PTFE sealing disks, a se o 12 g adua ed glass ubes o EasyFill ack R-12 Analys , esidual olume 1.0 ml, and a Flushback Module R-12 (Büchi Labo echnik AG, Flawil, Schwi ze land). Subsequen ly, all pa - icles we e ans e ed on o Mi IR slides (Ke ley Technologies, USA) o he inal lase di ec in a ed analysis using a glass pipe e. L´ opez- Rosales e al. (2022) achie ed high eco e ies (88 %) and high p ecision (RSD =4 %) o mic oplas ic pa icle ans e using he Synco e® au oma ic e apo a ion sys em. 2.9. Polyme cha ac e iza ion (<1 mm ac ion) The <1 mm sample ac ion was analyzed a he Helmhol z-Zen um He eon using he Agilen 8700 Lase Di ec s In a ed (LDIR) Chemical Imaging sys em (Agilen Technologies) in ans lec ion mode. The in- s umen ’s unc ional p inciples a e desc ibed in mo e de ail in p e ious publica ions (Da Cos a Filho e al., 2020; Dong e al., 2022; Hildeb and e al., 2020; Sci cle e al., 2020). The pa icle analysis wo k low o he Agilen Cla i y so wa e ( e sion 1.1.2) was used o he au oma ed analysis o he en i e sample se . He eby, he sensi i i y was se o he maximum (6/6). The pa icle analysis wo k low includes a comple e analysis o he size and shape o all pa icles. Fibe s we e dis inguished om agmen s based on hei elonga ion ac o (aspec a io) o >3 (Hildeb and e al., 2020). In he ollowing, we use he e m pa icles o e e o bo h shape classes, agmen s and ibe s. Spec a we e acqui ed wi h a spec al esolu ion o 8 cm −1 . The pa icle size ange o he LDIR imaging sys em was se o 50 μ m–5000 μ m. The au oma ic wo k low o he LDIR echnically enables MP de ec ion down o 10 μ m. Howe e , he p ac ically achie able size de ec ion limi highly depends on he analyzed ma ix and he le el o cleanliness o he sample. The used spec al lib a y (Mic oplas ic s a e 1.0, Agilen Technol- ogies) was expanded by spec a o in-house e e ence MPs and o ele- an en i onmen al pa icles (Hildeb and e al., 2022a). The au oma ed wo k low o he Cla i y so wa e acqui ed IR spec a o all pa icles. The hi quali y h esholds o a posi i e assignmen we e adap ed acco ding o he p ese alues. MP iden i ica ions we e ei he accep ed, manually assigned o ano he polyme class, o no accep ed. Only spec a in conjunc ion wi h high hi quali y alues ela ed o e e ence spec a (>0.80) we e conside ed o he inal s a is ics wi hou u he manual con i ma ion. In o de o p e en any o e es ima ion o he MP concen a ions, all analyses we e ho oughly e-e alua ed manually in ans lec ion mode and, i necessa y, also by he LDIR’s μ -ATR unc ion. I unambiguous con i ma ion o he assignmen was no possible, he espec i e pa icles we e assigned o na u al ma e ial classes o ma ked as “unknown”. Pa icles wi h hi quali ies <0.60 we e au oma ically classi ied as “unknown”. Ac yla es, polyu e hane (PU) and a nish a e polyme s ha a e di icul o dis inguish using FTIR spec oscopy (P impke e al., 2018) and LDIR and we e he e o e analyzed as one class o polyme s (Hildeb and e al., 2022b) e e ed o as PU in he ollowing. The wa enumbe ange o he LDIR (1800 cm −1 –975 cm −1 ) hampe s he accu a e di e en ia ion be ween na u al polyamide (PA) and syn he ic PA in en i onmen al samples (Hildeb and e al., 2022b). Pa icles assigned o PA by he au oma ed wo k low o he Cla i y so wa e we e he e o e excluded om u he analysis. The ew poly e a luo oe hylene (PTFE) MPs ha we e de ec- ed in he samples we e also disca ded om u he analysis, since labo a o y equipmen con aining PTFE we e used du ing sample p o- cessing. Syn he ic ubbe di e s om con en ional plas ics in ha i is an elas ome o med by he connec ion o polyme chains wi h sul u b idges. The majo i y o ubbe elas ome s in use a e a blend o na u al and syn he ic ubbe . Mic o ubbe he e o e is no iden ical o MPs in a s ic sense, bu was included in he analysis since i is an impo an pollu an wi h a simila e ec on co als (Reiche e al., 2024b). 2.10. Calcula ion o isk indices The heal h haza d o di e en polyme s a ies widely (Li hne e al., 2011). The en i onmen al isk o MPs pollu ion he e o e depends no only on he concen a ion o MPs, bu also on he ela i e abundance o di e en polyme s (Xu e al., 2018). Mos en i onmen al isk assess- men s he e o e a e based on hese wo pa ame e s, e en i he calcu- la ion o en i onmen isks di e s in de ail in he indi idual s udies. We calcula e he Con amina ion Fac o (CF) o compa e he ela i e abun- dance o MPs be ween di e en sampling s a ions and wi h he li e a u e (Tomlinson e al., 1980; Xu e al., 2018): CFi=Ci/Co(1) Whe e i deno es he sampling si es and CF i is he Con amina ion Fac o o he i h sampling si e. C i ep esen s he abundances o MPs a he i h si e and C 0 deno es he baseline concen a ion o MPs in sedimen s. The lowes MPs concen a ion ound in his s udy was chosen as baseline concen a ion, since i is nea ly iden ical o he lowes concen a ion epo ed o sedimen s om he Spanish Shel (Filguei as e al., 2019). A Polyme ic Risk Index (H) exp esses he combined en i onmen al haza d o all polyme s p esen in a sample and was adop ed om p e- ious s udies (Rakib e al., 2022; Xu e al., 2018): Hi=∑(Pji/Ci*Sj)(2) Whe e H i is he Polyme ic Risk Index a sampling si e i. Pji is he abundance o he speci ic polyme j a he sampling si e i, C i is he o al abundance o MPs a he sampling si e i and S j is he haza d sco e o he speci ic polyme j. The alues o all polyme ypes p esen a he si e a e summed up o calcula e H i . The haza d sco es o he polyme s a e based L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 5 on Li hne e al. (2011), wi h PP =1, PET =4, PE =11 and PS =30. Ac yla es/polyu e hane/ a nish is a g oup o polyme s which a e di icul o dis inguish using FTIR o LDIR (P impke e al., 2018; Hil- deb and e al., 2020). The polyme s wi hin hese g oup ha e di e en haza d sco es, bu include many polyme s ha a e ca ego ized as pa icula ly haza dous by Li hne e al. (2011). The haza d sco es o nine di e en polyme s om his g oup a e epo ed by Li hne e al. (2011) and ange om 230 o 13,844. A Mon e Ca lo Simula ion was employed o es ima e he po en ial dis ibu ion o o al haza d sco es o a sample composed o hese nine di e en polyme s. We simula ed hei unknown ela i e abundances using 1000 i e a ions o ep esen possible composi ions. Fo each simula ed composi ion, a weigh ed a e age haza d sco e was compu ed. The esul ing dis ibu ion o haza d sco es was used o assess he likely ange o haza d sco es o his poly- me g oup s a is ically. The a e age haza d sco e o he dis ibu ion is 7197, while 5386 ep esen s he 5 h pe cen ile and 8912 he 95 h pe cen ile o he dis ibu ion. These alues can be used as inpu o he haza d sco e o he ac yla es/polyu e hane/ a nish g oup in he calcula ion o he Polyme Risk and he Pollu ion Risk indices. This e- sul s in a ange o likely Polyme Risk and Pollu ion Risk indices o each sample con aining polyme s o he ac yla es/polyu e hane/ a nish g oup. The polyme PVDC is no di ec ly co e ed in Li hne e al. (2011). Howe e , inylidene chlo ide wi h a haza d co e o 111 (Li hne e al., 2011) is he only monome used in he p oduc ion o PVDC. A haza d co e o 111 is he e o e also assigned o PVDC. No haza d sco e can be gi en o ubbe , which is he e o e no conside ed o he calcula ion o he Polyme ic Risk Index. Fo a gi en sampling si e (i) he Polyme ic Risk Index (H i ) is com- bined wi h he Con amina ion Fac o (CF i ) o calcula e he Pollu ion Risk Index (PRI i ) ollowing he equa ion: PRIi=Hi*CFi(3) The Polyme ic Risk Index (H) and he Pollu ion Risk Index (PRI) o he en i e s udy a ea a e exp essed as geome ic means. Risk ca ego ies o each o he indices (CF, H and PRI) we e de ined om low (I) o e y high (V) and epo ed in Table S1. 3. Resul s The mos impo an sedimen a y componen s a e olcanic- ock agmen s, mine als and bioclas s. The mos equen ly iden i ied bio- clas s o igina e om echinode ms, bi al es, co als and ba nacles. Bio- clas s om calca eous ed algae, gas opods and polyplacopho ans a e less abundan . The mos impo an in o ma ion on he ex u e o he sedimen samples is summa ized in Table 1. Mean g ain sizes a y be- ween coa se sand (0.9 mm) and e y coa se sand (1.2 o 1.7 mm), while he sedimen ex u e can be classi ied as sandy g a el (IG1 and 5) o g a elly sand (IG2 o 4). The so ing o he sedimen is gene ally poo , excep o sample IG2 which shows a sligh ly be e mode a e so ing (Table 1). The g ain-size dis ibu ions o all samples show a sligh ly nega i e skewness, i.e. an asymme ic g ain size dis ibu ion owa ds coa se g ain sizes. The g ain-size dis ibu ion is unimodal in all samples excep IG4, which is bimodal. The bimodal dis ibu ion in his sample om a co al colony is due o he con ibu ion o ela i ely la ge co al agmen s. This leads o a second peak a la ge g ain sizes in he dis- ibu ion. Rela i ely la ge co al agmen s also con ibu e o he ela- i ely la ge a e age g ain size o he sedimen s in he o he co al colony sample (IG5). O he wise he g ain size ends o inc ease wi h wa e dep h and dec ease wi h co al co e (Table 1). The au oma ed pa icle analysis wo k low o he Agilen Cla i y so wa e iden i ied a o al o 3157 pa icles om all sampling si es, which we e analyzed using LDIR. Two hund ed h ee (203) o hese pa icles we e cha ac e ized as MPs, including ubbe . The majo i y o he MPs a e agmen s (96 %), while ibe s con ibu e only 4 % (Table S2). The concen a ion o MPs a ies s ongly be ween 41 and 6345 pa icles/kg dw (Table 2), wi h a mean o 1514 pa icles/kg dw and an in e qua ile ange o 530 pa icles/kg dw. The lowes alue was ound a si e IG1 loca ed in he cen al channel wi hou signi ican co al co e , while he highes concen a ion was ound in sedimen s apped in one o he co al colonies (IG5, Table 2, Fig. 2). All, excep wo MPs, a e smalle han 500 μ m (Fig. 3) and 92 % o all MPs a e smalle han 250 μ m. The wo la ge ou lie s a e 736 and 2009 μ m in size. The mean pa icle size o all MPs excluding he ou lie s is 123 ±88 μ m (Table 2). The size dis ibu ion o MPs shows a gene al inc ease in abundance owa ds he de ec ion limi a 50 μ m. The e is no clea ela ionship be ween he mean g ain size o sedimen in a sample and he size o MPs. This lack o co ela ion appea s o be p ima ily due o he abundance o ela i ely la ge co al bioclas s in he co al colony samples. Ins ead he e is a gene al endency o he size o MPs o dec ease wi h he densi y o he co al canopy (Tables 1 and 2). PE and PET a e he mos common polyme s in he s udy a ea and oge he accoun o >50 % o all MPs (Table 2, Fig. 3). The ibe s in pa icula , which only accoun o a ela i ely small p opo ion o all pa icles (4 %), consis mainly o PET (75 %). PP, ubbe and PS each con ibu e be ween 10 and 20 % o all polyme s, while PU and PVDC a e less abundan (<5 % each). The pa icle size o MPs a ies s ongly wi h polyme ype (Fig. 3), e. g. PE and PS a e he only polyme ypes p esen in he size ange >250 μ m. In con as , all PU and PVDC pa icles and >90 % o all PET pa icles a e smalle han 100 μ m. This esul s in median size alues o <100 μ m o PP, PU, ubbe , PET and PVDC, while PE and PS ha e median size alues o 141 and 134 μ m espec i ely (Table 3). The Mann-Whi ney U es con i ms ha he median size alues o mos polyme ypes a e di e en om each o he . This is suppo ed by he Kolmogo o –Smi no es , which shows ha he g ain size dis ibu ions o he polyme ypes, which ew excep ions, a e di e en om each o he (Table 3). The CF alues show a high a iabili y (Table 4) and ange om 1 o 115.5, indica ing mode a e con amina ion in wo samples (IG1 and 2) and e y high con amina ion in h ee samples (IG3–5). In a e age he con amina ion in he s udy a ea is e y high, as indica ed by an a e age CF o 28.8 (Table 4, Fig. 4). The Polyme ic Risk Index o he indi idual samples, a ies be ween low and e y high isk. This a iabili y p i- ma ily e lec s he ela i e abundance o PU (Table 2), he polyme g oup wi h by a he highes haza d sco e. The Polyme ic Risk Index (H) o he en i e s udy a ea gene ally indica es a conside able isk (Table 4). A mode a e isk only applies unde he assump ion o a low haza d sco e o PU (5 h pe cen ile o he Mon e Ca lo Simula ion). The Pollu ion Risk Index (PRI) indica es a high isk o he s udy a ea, while he isk o indi idual samples anges om low o e y high (Table 4, Fig. 4). Wi h he excep ion o sample IG5, he Pollu ion Risk Index is mainly de e mined by he Polyme ic Risk Index (H) a he han he concen a ion ac o (CF). This is because he haza d sco es o he polyme s a y mo e s ongly compa ed o he MP exposu e le el exp essed as concen a ion ac o (Table 4). Only he Pollu ion Risk Ca ego y o sample IG1 is in luenced by he chosen haza d sco e o PU, indica ing high isks o high and a e age haza d sco es bu only conside able isk o low haza d sco es (5 h pe cen ile o he Mon e Ca lo Simula ion). O e all, his shows ha he isk assessmen does no depend signi ican ly on he exac haza d sco e selec ed o PU. 4. Discussion The concen a ion o loa ing plas ic deb is in he no h-wes e n Medi e anean is one o he highes in he wo ld and simila o he plas ic ga bage pa ches in he sub opical gy es (C´ oza e al., 2015). Fagiano e al. (2023) ecen ly con i med ha his a ea, including he Columb e es Islands, ecei es high amoun s o loa ing plas ics. The e- o e, i is no su p ising ha he absolu e concen a ion o MPs in he sedimen s o Illa G ossa is e y high (Tables 2 and 4). Howe e , he a e age pollu ion exp essed as Con amina ion Fac o (CF) is also much L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 6 highe compa ed o nea ly all o he shel sedimen s in he wes e n Medi e anean (Fig. 4). This e en holds i he e y highly con amina ed sample (IG5) is no conside ed in he calcula ion o he a e age. 4.1. Co al habi a s as plas ic sinks The highe MPs concen a ion in he bay o Illa G ossa compa ed o o he egions o he wes e n Medi e anean could be pa ially ela ed o he p esence o he C. caespi osa co al colonies. Sedimen s in co al ee en i onmen s ha e p e iously been iden i ied as mic oplas ic sinks (Huang e al., 2021; Lin e al., 2024; U ami e al., 2021). Habi a - o ming species such as seag ass, mac oalgae and co als, a e known o ap MPs in o sedimen s (Feng e al., 2020; Mend ik e al., 2024; Sanchez-Vidal e al., 2021; Smi e al., 2021). These habi a o ming species e ec i ely educe he low eloci y and u bulen kine ic ene gy on he sea loo h ough ene gy dissipa ion (Hend iks e al., 2010) and hus p omo e he se ling and e en ion o MPs on he sedimen (Smi e al., 2021; Yen e al., 2024). Seag ass is absen om Illa G ossa Bay (Templado and Cal o, 2002), bu mac oalgae and co als a e abundan (Ke s ing and Lina es, 2012; Ke s ing e al., 2014; Pons-Fi a e al., 2019) and likely con ibu e o apping o MPs. Howe e , he e ec o mac oalgae is likely limi ed o hei g ow h pe iod in sp ing and summe (Pinedo e al., 2015), whe eas co als a e p esen all yea a ound. Di e en apping mechanisms o MPs in co al canopies ha e been iden i ied (Mend ik e al., 2024): co als can ac as an obs acle o luid low, leading o deposi ion o MPs on he up-cu en side o colonies. MPs can ge apped wi hin he co al colonies hemsel es, o accumula e in he wake zone on he down-cu en side o co als. E en spa se co al canopies lead o s ongly enhanced mic oplas ic apping compa ed o Table 2 MPs and mic o ubbe concen a ion, size in o ma ion and ela i e abundance o di e en polyme s. Rela i e abundance o polyme s (%) Si e MPs pa icles/kg dw MP + ubbe pa icles/kg dw MPs + ub. Mean size ( μ m) >250 μ m (%) PE PP PU Rubbe PS PET PVDC IG1 41 41 172 20 50 10 10 0 30 0 0 IG2 49 49 116*20 20 0 20 0 20 20 20 IG3 549 579 152*15 51 0 0 5 40 4 0 IG4 522 555 109 3 29 59 0 6 6 0 0 IG5 4736 6345 83 0 0 3 7 25 0 65 0 S udy a ea 1179 1514 123 8 28 11 4 12 19 25 1 * Mean was calcula ed wi hou he wo ou lie s in samples IG2 (2009 μ m) and IG3 (736 μ m). Fig. 3. Rela i e abundance o di e en polyme ypes in he s udy a ea o he en i e size spec um (A) and only MPs la ge han 250 μ m (B). (C) shows he pa icle size dis ibu ion o all polyme ypes combined ( iolin plo ) and o each indi idual polyme ype (ji e plo ). No included in (C) a e wo ou lie s (PS, 736 μ m and PE, 2009 μ m). (D) Pa icle size dis ibu ion plo o all polyme ypes excep PVDC, wi h he 50 % line indica ing he median pa icle size. Table 3 Median pa icle sizes o di e en polyme s (A) and s a is ical pa ame e s o he equali y o dis ibu ions (B and C). B: The p- alues o he wo- ailed Mann- Whi ney es indica es whe he he median alues o wo polyme ypes a e equal. The medians can be ega ded as s a is ically di e en , when he p- alue is <0.05. C: The Kolmogo o -Smi no es indica es whe he wo dis ibu ions a e equal. The D- alues a e a measu e o he maximum absolu e di e ence be ween he wo es ed dis ibu ions. The di e ence can be ega ded as s a is ically dis inc , when he p- alue is <0.05. (a) PE PP PU Rubbe PS PET PVDC Median size ( μ m) 141 75 65 89 134 61 81 (b) Mann-Whi ney pai wise es (p alue) PE PP PU Rubbe PS PET PE 0.0006 0.0002 0.0144 0.4348 3.69E-12 PP 0.0675 0.2639 0.0020 0.0024 PU 0.0109 8.37E-05 0.7149 Rubbe 0.0367 3.39E-05 PS 3.95E-11 PET (c) Two-sample Kolmogo o –Smi no es (D alue) PE PP PU Rubbe PS PET PE 0.40528 0.82143 0.32738 0.16892 0.68768 PP 0.47826 0.24457 0.44066 0.44075 PU 0.625 0.83784 0.23529 Rubbe 0.29842 0.55882 PS 0.68362 PET L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 7 Table 4 Con amina ion, Polyme ic and Pollu ion Risk Indices and isk ca ego ies o indi idual samples and he en i e s udy a ea. A ange o alues o H and PRI we e calcula e o samples con aining PU. The a e age alue and uppe and lowe bounds o hese isk indices we e calcula ed using a Mon e Ca lo Simula ion o he likely haza d sco e dis ibu ion o PU. H and PRI o he s udy a ea a e exp essed as geome ic mean. The isk ca ego ies a e low (I), mode a e (II), conside able (III), high (IV) and e y high (V). Fo de ails on he associa ion o isk indices wi h ca ego ies see Table S1. Si e Con amina ion Fac o (CF) Con amina ion Risk Ca ego y Polyme ic Risk Index (H) Polyme Risk Ca ego y Pollu ion Risk Index (PRI) Pollu ion Risk Ca ego y Range Range Range Range low a e age high low a e age high low a e age high low a e age high IG1 1.0 II 544 725 897 III III III 544 725 897 III IV IV IG2 1.2 II 1106 1469 1812 IV IV IV 1322 1755 2165 V V V IG3 13.4 V 8 I 107 I IG4 12.7 V 4 I 51 I IG5 115.5 V 485 648 802 III III III 55,998 74,825 92,665 V V V S udy a ea (mean) 28.8 V 99 117 133 II III III 739 878 997 IV IV IV Fig. 4. (A) Map o co al co e (%) (modi ied om Ke s ing and Lina es (2012)) in Illa G ossa Bay wi h he ca ego y o he Pollu an Risk Index (PRI) o indi idual samples and he s udy a ea. The PRI in he igu e was calcula ed assuming an a e age haza d sco e o PU. De ails on he ela ionship be ween isk ca ego ies and pollu ion indices can be ound in Table 4 and Table S1. (B) Map o he wes e n Medi e anean showing he Con amina ion Fac o (CF) isk ca ego y o sub idal shel sedimen s. The CF was calcula ed om MP concen a ions om his s udy ( ed box) and 1: Filguei as e al., 2019; 2: Aloma e al., 2016 and 3: Fagiano e al., 2023. (Fo in e p e a ion o he e e ences o colou in his igu e legend, he eade is e e ed o he web e sion o his a icle.) L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 8 ba e sea loo s, bu he e ec inc eases u he wi h co al canopy densi y (Mend ik e al., 2024). This is b oadly consis en wi h he esul s om his s udy. Ve y high MP con amina ion in he s udy a ea is limi ed o samples (IG3–5) om co al colonies and wi h high co al co e (>10 % co al). The sample om he cen al a ea (IG1) ha is nea ly de oid o co als in con as shows he lowes MPs concen a ion (Fig. 2; Tables 1 and 2.). Hyd odynamic s udies sugges ha he co al canopy aps smalle MPs mo e e icien ly han la ge MPs (Smi e al., 2021). This is e lec ed in he smalle g ain size o MPs in co al colonies and a eas wi h highe co al canopy densi y compa ed o he sample wi h <2 % co al co e (Fig. 2; Tables 1 and 2.). The concen a ion and pa icle size o MPs in he analyzed samples he e o e appea o be consis en wi h hyd odynamic conside a ions, al hough he ela i ely small numbe o samples p ecludes a mo e igo ous analysis. The di e ences in sampling yea s (Table 1) and hus he age o he sedimen canno explain he di e ences in he MPs concen a ion be- ween he si es. Samples IG4 and 5 we e aken wi hin co al colonies. Based on he g ow h a es o he co als, hese sedimen s ha e likely been deposi ed since he mid-1980s (see sec ion “Me hod”) and in a e age a e he e o e likely olde compa ed o he sedimen s om he sea loo (IG1 o 3). Due o he inc ease in MP con amina ion o e he las decades, one could assume ha olde sedimen s migh con ain lowe MP concen a- ions. This does no seem o be he case. In ac , he “old” sedimen s wi hin he co al colonies ha e he highes MP concen a ions (Table 2), sugges ing ha he apping po en ial o co als seems o cancel ou he e ec o sedimen age. In addi ion o he hyd odynamic e ec s o he co al canopy, i is assumed ha he speci ic p ope ies o co al ee sedimen s a o he apping o MPs (U ami e al., 2021): he biogenic g ains in ee en i- onmen s a e o en po ous and i egula ly shaped which can a o en anglemen wi h MPs, he eby apping hem in he sedimen . MP can also be p o ec ed agains esuspension and en ainmen by he o en compa a i ely la ge biogenic g ains in hese en i onmen s. This could also be he case in ou s udy wi h MPs o conside ably smalle size compa ed o he mean sedimen g ain sizes (Tables 1 and 2). 4.2. Oceanog aphic and opog aphic in luences In ou s udy a ea he con amina ion by MPs is ela i ely high e en compa ed o mos o he ee sys ems (Huang e al., 2021; Pa e son e al., 2022; P adhap e al., 2023; U ami e al., 2021). The high MPs measu ed can he e o e no solely be explained by he hyd odynamic e ec s o he co al canopy and he associa ed sedimen s. The high MP concen a ions (Table 2) a e mo e simila o he ones ound in ee sys ems close o highly popula ed a eas (Pa i e al., 2020; Zhou e al., 2023) and ou ism ho spo s (Lim e al., 2022; Lin e al., 2024). Howe e , pollu ion om local sou ces seems unlikely o ou s udy a ea, which is loca ed in a emo e Ma ine Rese e wi hou di ec land-based sou ces o plas ic con amina ion. I he e o e seems likely ha long dis ance anspo plays a ole o he accumula ion o MPs in Illa G ossa Bay, as i was p oposed o simila ly high MP concen a ions in sedimen s o he sou he n Xisha Islands in he Sou h China Sea (Lin e al., 2024). This is also in line wi h p e ious obse a ions (Aloma e al., 2016; Fagiano e al., 2022) and modelling esul s (Ha zonikolakis e al., 2022) om he wes e n Medi e anean, which show ha he concen a ions o MPs in Ma ine P o ec ed A eas (MPAs), whe e local sou ces a e sca ce o non- exis en , a e simila o o e en highe han in ou is ic ho spo s (Aloma e al., 2016; Fagiano e al., 2023). Modelling s udies sugges s ha he No he n Cu en (Ou mie es e al., 2023; So o-Na a o e al., 2020), he majo cu en o he no h- wes e n Medi e anean zone (Fig. 1), is he main ec o o MPs o he Balea ic Sea. The No he n Cu en o igina es in he Ligu ian Sea and lows sou hwes -wa ds pa allel o he con inen al slope in o he s udy a ea. Some o he mos impo an poin sou ces o plas ic ma ine li e in he Medi e anean, such as he ci y o Ba celona, he Rhone i e (Liuba se a e al., 2018), and he Eb o i e (Simon-S´ anchez e al., 2019) a e si ua ed along i s pa h. I he e o e can anspo loa ing MPs om he densely popula ed coas al o inland a eas o I aly, F ance and Spain in o he s udy a ea (Ou mie es e al., 2023). In addi ion, pa icle acking models show ha , a leas seasonally, pa icles om e en mo e dis an sou ces such as he Alge ian coas can each he Columb e es Islands h ough no hwa d di ec ed cu en s (Ke s ing e al., 2020). This is consis en wi h he obse a ion ha loa ing plas ics in he a ea o Columb e es a e cha ac e ized by ela i ely small i ems composed mainly o agmen s (Fig. S2), which was in e p e ed o indica e mo e aged plas ic om dis an sou ces (Fagiano e al., 2022). In addi ion, he C-shaped island o Illa G ossa is open owa ds he no hwes (Fig. 1), he main di ec ion o s o ms and wa es (Ke s ing and Lina es, 2012), and he e o e could ac as a ap o ma ine li e . Floa ing plas ics, b ough o he egion om dis an sou ces, could he e o e been apped in he inne bay o he island and e en ually e ained in i s sedimen . Simila shape dependen apping e ec s ha e been p e iously obse ed o conca e beaches ha ap ma ine deb is much mo e e icien ly han o he shaped sho elines (B ennan e al., 2018). The combina ion o his opog aphic ea u e wi h he apping po en ial o he co al canopy likely is he eason why he CF a Illa G ossa is highe compa ed o almos all o he si es in he wes e n Medi e anean (Fig. 4), despi e he ac ha he island is si ua ed in a Ma ine P o ec ed A ea. This con i ms p e ious s udies ha ha e shown ha Ma ine P o ec ed A eas can be ho spo s o plas ic pollu ion (Fagiano e al., 2023). Local p o ec ion alone is he e o e no e ec i e in educing he pollu ion load. 4.3. Polyme composi ion The polyme PE, he plas ic ype wi h he highes global p oduc ion a e (Plas ics Eu ope, 2022), is also he mos common MP ype ound in he s udy a ea (28 %). The abundance o PU also oughly co esponds o i s impo ance in global p oduc ion (~ 5 %). The low abundance o he polyme ype PVDC (1 %), which only occu s in one sample, is consis en wi h i s ela i ely low p oduc ions a es compa ed o o he plas ic ypes. To ou knowledge, i was p e iously epo ed in only one o he s udy on MPs in ee sys ems, whe e i was iden i ied as a copolyme (Ding e al., 2019). In con as , he polyme ypes PET (25 %) and PS (19 %) a e highly o e ep esen ed in ou s udy a ea compa ed o hei sha e o global p oduc ion, which is be ween 5 and 7 % each (Plas ics Eu ope, 2022). A s udy on loa ing plas ic li e (>355 μ m), om he coas al egions o se e al islands in he NW Medi e anean shows ha he polyme composi ion o he Columb e es Islands, including Illa G ossa, di e s om o he islands in he egion (Fagiano e al., 2022). The coas al wa- e s o he Columb e es Islands show highe abundances o PC and PS and he lowes abundance o PP compa ed o he o he islands in he a ea (Fagiano e al., 2022) and o he a eas o he cen al wes e n Medi e anean (Sua ia e al., 2016). This is consis en wi h ela i ely low PP (11 %) and high PS (19 %) concen a ions in he sedimen s o Illa G ossa Bay (Table 2). Fagiano e al. (2022) sugges ed ha hese high le els o PS in he egion could be ela ed o i s use o con aine s and packaging ma e ial in he ishing indus y. PET agmen s by a p e- domina e o e PET ibe s in ou s udy a ea, indica ing ha PET-MPs likely o igina e om he deg ada ion o la ge PET deb is such as bo - les (Ioakeimidis e al., 2016) a he han om ex iles (Oli ei a e al., 2023). Li le is known abou he abundance o mic o ubbe in he Medi- e anean Sea. To he bes o ou knowledge, mic o ubbe abundance has no been quan i ied o sub- idal sedimen s in he Medi e anean o da e. The analysis shows a ela i ely high con ibu ion (12 %) o ubbe o he o al MP concen a ion, wi h e y high concen a ions (25 %) in sedimen s om wi hin one o he co al colonies. P e ious s udies on mic oplas ic pollu ion in he Medi e anean Sea indica e ha ubbe con ibu es only e y small amoun ( ypically <1 %) o he o al loa ing plas ic load (Kedzie ski e al., 2022; Sua ia e al., 2016). L. Reuning e al. Ma ine Pollu ion Bulle in 217 (2025) 118070 9