Validation of a Lagrangian model for large-scale macroplastic tracer transport using mussel-peg in NW Spain (Ría de Arousa)

Valida ion o a Lag angian model o la ge-scale mac oplas ic ace anspo using mussel-peg in NW Spain (Ría de A ousa) Sa a Cloux a, ⁎,Sil ia Allen-Pe kins b ,Hilda de Pablo c ,Daniel Ga aboa-Paz a , Ped o Mon e o b ,Vicen e Pé ez Muñuzu i a a CRETUS, Nonlinea Physics G oup, Facul y o Physics, Uni e si y o San iago de Compos ela, Spain b INTECMAR, Ins . Tecnolóxico pa a o Con ol do Medio Ma iño de Galicia, Vilaga cía de A ousa, Spain c MARETEC, Ins i u o Supe io Técnico, Uni e sidade de Lisboa, Po ugal HIGHLIGHTS •The s udy a ea o ien a ion wi h espec o he in e idal mo emen plays a majo ole. •Wind e ec has a g ea influence on he dis ibu ion o ma ine deb is accumula- ion on he Ria de A ousa. •The beaching pa ame iza ion can ep e- sen di e en e en ion scena ios. GRAPHICAL ABSTRACT ABSTRACTARTICLE INFO A icle his o y: Recei ed 16 No embe 2021 Recei ed in e ised o m 18 Janua y 2022 Accep ed 18 Janua y 2022 A ailable online 29 Janua y 2022 Edi o : Ke in V. Thomas Ma ine deb is is a g owing p oblem in ecen yea s due o popula ion g ow h a ound he wo ld. The inco ec manage- men o plas ic was e causes hese bodies each he seas and oceans, becoming a wo ldwide p oblem. Once hey each he seas and oceans, hey begin a long pe iod o deg ada ion, mo ing om a mac o s a e (plas ics whose diame e is g ea e han 0.5 cm) o a mic o s a e (diame e less han 0.5 cm). The mic oplas ics sp ead h oughou he oceans, en- e ing he ood chain o ma ine species and, subsequen ly, o humans. The e o e, i is impo an o s op he p oblem while i emains a he mac oscale. In his wo k, a alida ion o a ecen ly de eloped Lag angian compu a ional model o ack he mo emen o mac o plas ics in seas and oceans is p esen ed. This alida ion is pe o med on a e- gional scale, in he Ría de A ousa, one o he mos impo an es ua ies o mussel cul i a ion in no hwes e n Spain. Du ing mussel cul i a ion in a s, a ype o floa ing plas ic s ick a e eleased, he mussel-pegs. The po en ial o his s udy is ha we can compa e he accumula ion esul s o he model wi h he accumula ion da a collec ed on he Ga- lician beaches. In a gene al amewo k, he influence o wind on he spa ial dis ibu ion o he accumula ions gi en by he model was obse ed. Fo he moni o ing da a, simila esul s we e ound o he accumula ion ends o e he en i e o al pe iod. Fo he mon hly ep esen a ion, some disc epancies we e obse ed. These di e ences can be a ibu ed o pa icula synop ic si ua ions, poo ep oduc ion o he coas line o o he e y o ien a ion o he s udy a ea wi h espec o he in e idal dynamics. Keywo ds: Mac oplas ics Mussel-pegs NW Spain Lag angian- acking Moni o ing Ría de A ousa Science o he To al En i onmen 822 (2022) 153338 ⁎Co esponding au ho . E-mail add esses: [email protected] (S. Cloux), scace es@in ecma .gal (S. Allen-Pe kins), hildadepablo@ ecnico.ulisboa.p (H. de Pablo), angeldaniel.ga aboa@ou look.es (D. Ga aboa-Paz), pmon e o@in ecma .gal (P. Mon e o), pe [email protected] (V. Pé ez Muñuzu i). h p://dx.doi.o g/10.1016/j.sci o en .2022.153338 0048-9697/© 2022 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). Con en s lis s a ailable a ScienceDi ec Science o he To al En i onmen jou nal homepage: www.else ie .com/loca e/sci o en 1. In oduc ion Conce n abou ma ine li e has inc eased in ecen decades (Rangel- Bui ago e al., 2020). This is pa icula ly ue o plas ics (also e e ed o as plas ic bodies o plas ic deb is), which a e he majo i y o li e consis ing o objec s ha ha e been manu ac u ed o used by people be o e being dumped in i e s, seas and beaches (Williams and Rangel-Bui ago, 2019). These plas ic bodies ep esen a huge p oblem o ou oceans and coas s (Amelia e al., 2021;K ause e al., 2020), ha ing a g ea impac no only on he en i onmen , bu also on he human and ma ine wildli e heal h (Ka balaei e al., 2018). As hey can emain floa ing in he ocean o yea s, hese plas ics a e b oken down in o small agmen s by a ious agen s (wind, salini y, sola adia ion, empe a u e and mechanical s ess due o cu en s and wa es mo ion) as epo ed Min e al. (2020) and Wayman and Niemann (2021). I is possible o di e en ia e be ween mac oplas ics (whose diame e is g ea e han 0.5 cm) and mic oplas ics (whose diame e is less han 0.5 cm). These small bodies en e he ophic chain o ma ine animals and hen, humans. The e ec s o e animal and human heal h a e i ual unknown (Akdogan and Gu en, 2019;Hwang e al., 2020). In addi ion, he e a e huge economic losses o ma ine- ela ed sec o s (Abalansa e al., 2020). The spa io- empo al moni o ing o mic oplas ics is di ficul (Li e al., 2020;Ca e e o e al., 2022). The e o e, p e en ion is he fi s ac ion equi ed, ocusing on he s udy o mac oplas ics in i e s and coas s in hei ini ial s a e (Rangel-Bui ago e al., 2020). Despi e being one o he main pollu an s in he ocean, he anspo o ma ine plas ics h ough ocean ci cula ion emains poo ly unde s ood (Chassigne e al., 2021). Se e al ac o s can influence he unde s anding and desc ip ion o he mo emen o his ma ine li e , as epo ed Van Sebille e al., 2020. No only in e ms o ocean p ope ies such as empe a- u e, salini y, o biodeg ada ion (Coope and Co co an, 2012;U banek e al., 2018); bu also depending on he cha ac e is ics o he plas ic body (such as i s weigh , densi y and deg ada ion). S udies ha e beendone ode- ec mic oplas ic acumula ions by using sa elli e images. Da aasu en e al. (2018) uses SAR sa eli e in o ma ion o de ec hese plas ic pa ches o e he oceans. Goddijn-Mu phy and Williamson (2019) uses he mal in a ed (TIR) sensing o y o con ol wa e pollu ion by plas ics. They applied a- dia i e ans e heo y o es ima e he su ace ac ion o wa e co e ed by plas ic li e . D i e s acking can also be use ul o de ec con e gence zones depending on he oceanic dynamic (Maximenko e al., 2012). An- o he way o s udy he e olu ion o plas ics is he de ec ion on beaches. This ype o s udy equi es ained pe sonnel as well as he de e mina ion o he ype o plas ic body o be s udied. Howe e , many inconsis encies can a ise in his ype o s udy, as he de ec ion o plas ic pa icles in sand may depend on mul iple ac o s such as pe sonal expe ience o isibili y (La e s e al., 2016). Mo eo e , in his kind o s udy he o igin o he plas- ics collec ed on he beach is unknown. I is he e o e c i ical o ocus on he anspo o mac oplas ics ha mo e on he su ace, be o e deg ading and sinking. Ac ually, he mos app op ia e way o s udy he influence o plas- ics is he use o Lag angian anspo ools. Lag angian me hodologies a e widely used in ma ine li e anspo s udies (Ca lson e al., 2017). The e a e many s udies using Lag angian models in di e en pa s o he wo ld o ack he mo emen o plas ic in he ocean and sea (Chassigne e al., 2021;Poli ikos e al., 2020;Tu ell, 2020;Khoi unnisa e al., 2020; Zambianchi e al., 2017;Sousa e al., 2021). These models a e based on su - ace cu en models o define he mo emen o plas ics, s udy hei o igin and hei a e, since hey can p o ide he posi ion (x,y,z) a each ime s ep. In his s udy, a Lag angian model was conside ed o simula e he anspo o he mussel-pegs along one o he mos impo an es ua ies o he Galician coas o he mussel a ming indus y. Galicia has a coas line o 1195 km (as epo ed he Ins i u o Nacional de Es adis ia, INE). Geog aphically, he Rías a e one o he peculia i ies o Ga- lician coas (see Fig. 1). They a e inden a ions in he coas whe e he sea flooded i e alleys by lowe ing he land le el ( ela i e ise in sea le el). These kind o es ua ies a e impo an o fishing, con ibu ing o he ac ha he Galician coas is one o he mos impo an fishing and sea ood a eas in he wo ld (Laba a e al., 2004). The es ua ies a e adi ionally di- ided in o Rías Al as and Rías Baixas, acco ding o hei posi ion wi h e- spec o Finis e e as he mos wes e n cape o Galicia. In Galicia, he eno mous impo ance o mollusc aquacul u e, mainly mussels, makes i he mos impo an economic ac i i y in he communi y (Laba a e al., 2004;Laba a and Fe nández-Rei iz, 2019). Mussels a e cul i a ed in a s (Fig. 2). These a e floa ing nu se ies consis ing o a ec angula wooden s uc u e o which he mussel opes a e a ached. The wooden s uc u es a e a ached o floa ing bodies ha hold hem on he su ace o he sea, and also fixed o he seafloo by a chain (see Fig. 2). They a e hus suspended cul u es g owing in he wa e column. The mussel spawn is w apped a ound he ope wi h he help o a fine biodeg adable ayon ne , gi ing he mussel enough ime o a ach i sel o he ope. A e a ew mon hs, due o he conside able inc ease in weigh o he mussels, i be- comes necessa y o di ide he opes, i.e. o make new opes o lowe mussel densi y. Then, e e y 30–40 cm o ope, wooden (in he pas ) o plas ic (mo e common nowadays) s icks a e placed be ween he s ands o ope o p e en clus e s o mussels om becoming de ached as shown in Fig. 2 b. These mussel-pegs p o ide a be e g ip on he ope, p e en de achmen and make i easie o fix he mussels. They a e plas ic pieces 225 mm long by 30 mm wide (see Fig. 2c). The wo k o s inging, un olding and ex ac ing he mussels means ha he mussel-pegs, ei he whole o b oken, a e eleased and acciden ally all in o he sea, becoming one o he mos common ypes o ma ine li e in he Galician coas . Because he la ge ex- ension o he c op, as well as he ele an quan i y o hese plas ics icks e- leased in o he wa e and hei cha ac e is ics ( hey ha e a known o igin, a e easy o cha ac e ize and o coun ), hey can be used as indica o s o ma- ine li e a eas o accumula ion. The s udy he e p esen ed is loca ed in he Ría de A ousa as pa o he Rías Baixas. This pa icula es ua y is he mos ac i e inden a ion o he mussel p oduc ion (Comeau e al., 2018). Since he e is no deg ada ion o he mussel-peg om he ime i is e- leased un il i eaches he coas , i can be conside ed as a mac oplas ic body. In addi ion, due o he collec ions o mussel peg on beaches, i is pos- sible o know how his ype o plas ic accumula es along he coas line o he A ousa es ua y. The e o e, he main objec i e o his s udy is o know whe he a Lag angian model can p o ide a good es ima ion on he accumu- la ion o ma ine li e along he coas line in a small egion wi h he dynam- ics o he Ría de A ousa. A simila s udy was ca ied by Decle ck e al. (2019) and Ca lson e al. (2017) by using d i e ajec o ies and floa ing mac o deb is espec i ely. In he p esen wo k, a new ool ecen ly de eloped by he Ins i u o Supe io Técnico de Lisboa and he Uni e si y o San iago de Compos ela, he MOHID-Wa e Lag angian model was used. The me hodology o which is desc ibed in he Me hods sec ion. Al hough he Illa de A ousa is pa o he Ría de A ousa, om he e on we will e e o he Ría de A ousa as he inland coas al egion, excluding he Illa de A ousa. The objec i e o his s udy is o de e mine whe he he mussel-peg is a good Lag angian indica o . Then, could he me hodology he e be used o alida e he ecen ly de eloped MOHID-Lag angian ool and o de e mine acc e ion zones? This s udy p o ides new insigh o Lag angian s udy in high esolu ion a eas whe e he o ien a ion and cha ac e iza ion o he s udy a ea plays an majo ole. 2. Me hods 2.1. S udy a ea The Ría de A ousa (Fig. 1b) is loca ed a he no hwes coas o he Ibe- ian Peninsula (Fig. 1a), being pa o he Rías Baixas o he Galician coas . I is he la ges es ua y on he sou h Galician coas , wi h a su ace a ea o 230 km 2 app oxima ely. I can be subdi ided in o wo zones: he inne zone (98 km 2 ), mo e exposed o flu ial p ocesses, shallowe and mo e iso- la ed habi a s, and he ou e zone (132 km 2 ), wi h g ea e in e ac ion wi h oceanic p ocesses (O o, 1975;Rosón e al., 1995;Ál a ez-Salgado e al., 1996), deepe and mo e exposed habi a s (Ou ei o e al., 2018). The a e - age sea le el oscilla ion is a ound 2.5 m, making i a meso- idal es ua y S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 2 (be ween 2 and 4 m idal ange). The a e age wa e heigh anges be ween 1 and 2.5 m o he sp ing and summe seasons, while o he au umn and win e mon hs, his alue oscilla es be ween 3 and 5 m (Conselle ia de Medio Ambien e, 2010). These a ia ions in sea le el heigh gi e ise o cy- clical mo emen s o incoming and ou going wa e , called idal cu en s, which can be o g ea in ensi y. The Ría de A ousa as a pa ially mixed es ua y (Rosón e al., 1995; Dye , 1973) and a e age he cu en s o e se e al idal cycles because he esidual ci cula ion ob ained is posi i e in wo laye s: an inflow cu - en o sal wa e h ough he bo om om he ocean and an ou flow cu - en o b ackish wa e h ough he su ace. In he Ría de A ousa he e a e also se e al sou ces o s a ifica ion: eshwa e inpu s, hea flow h ough he a mosphe e- ia in e ace and wa e exchanges be ween he shel and he es ua y. All o hese gi e ise o longi udinal densi y g adien s, which ac as d i e s o he posi i e ci cula ion desc ibed abo e. Winds also ha e a g ea influence on he a iabili y o he esid- ual ci cula ion bo h on he shel and in he es ua y, which can lead o h ee di e en scena ios: (i) ein o ced posi i e ci cula ion when winds om he no h domina e, (ii) a si ua ion o elaxa ion in ansi- ion pe iods wi h a e y slow posi i e ci cula ion main ained o days, and (iii) an in e se o nega i e ci cula ion a o ed by winds om he sou h. In addi ion, depending on i s ba hyme y, he Galician coas can be cha ac e ized in di e en ways. Al hough i is possible o dis in- guish be ween 11 ypes o pu e ypologies, he eno mous a ie y on he Galician coas and he changing hyd odynamic condi ions gi e ise o he appea ance o an ex ensi e se ies o in e media e ypologies. In pa - icula , he beaches o he Ría de A ousa a e small, o g ea ypological a ie y, and nume ous due o he in ense ac u ing o he e ain (Conselle ia de Medio Ambien e, 2010). In pa icula , his es ua y is domina ed by shel beaches, which means ha he e a e ocky in e idal zones ha influence he dynamics. Compu a ionally, depending on he g id mesh, ocky a eas may no be ep esen ed so Lag angian pa icles mo e eely wi hou beaching, o a leas be delayed by he p esence o obs acles. Fig. 3 shows he na iga ion cha s o he s udy a eas. As i can be seen, mos o hem ha e ocky in e idal zones jus in on o hecollec iona ea,bu o Ba añaandAduana egions.Fo hecase o Po co beach (Fig. 3 e. i) he collec ion a ea is loca ed on he le side o he bu e , in on o he Pun al de las Sinas. Compu a ionally, his e ec can only be pa ame e ized in he MOHID-Lag angian by he beaching ac o and he beaching le el. Fig. 1. Maps showing he Galician coas (solid black line on he le map) and he Ría de A ousa (blue squa e on he le map); and he spa ial dis ibu ion o he mussel a s (blue do s on he igh map). Fig. 2. Schema ic o a mussel cul i a ion a . (a) Shows he main componen s o he pan: (i) wooden s uc u e, (ii) floa s, (iii) cul i a ion opes and(i ) ancho ing.(b) Shows he a angemen o he mussel-peg on he ope and (c) he dimensions o he mussel-peg. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 3 Fig. 3. Na iga ion cha s o he Ría de A ousa (a). Going om sou h o no h and om wes o eas in each image, he shaded a eas co espond o: (b.i) Mosquei os, (b.ii) Fon án, (c.iii) Illa, (c.i ) Ninei iños, (d. ) Ba aña, (e. i) Po co, (e. ii) Aduana, ( . iii) S1,( .ix) S2, ( .x) S3, ( .xi) S4, ( .xii) S5, ( .xiii) S6, ( .xi ) S7, ( .x ) S8, ( .x i) S9, (g.x ii) Cas iñei as and (g.x iii) Ca ei o. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 4 2.2. Inpu da a. Hyd odynamic model The MOHID-Lag angian model (www.mohid.com) was employed, a e- cen Lag angian ool de eloped o he dynamic s udy o ma ine anspo . In pa icula , i was used i o ack o wa d in ime he anspo o ma ine deb is in he oceans and o es ima e i s accumula ion along he coas . I is a comple e high- esolu ion Lag angian acking model ha can unc ion as a lib a y o he MOHID Wa e modelling sys emo asa s and-alone p og am. Wo king independen ly, ou Lag angian model will be o ced wi h he cu - en alues ob ained using he MOHID-Wa e 3D hyd odynamic model (Ma ins e al., 2001) o ep oduce he cu en , salini y and empe a u e in he RV. This ope a ional sys em is daily execu ed by Me eoGalicia (www.me eogalicia.gal) o ob ain he ad ec i e con ibu ion. The ou pu esul s has a 300 m o ho izon al esolu ion and he e ical disc e iza ion consis s by wo domains wi h an in e ace a 8.68 m. While he uppe domain is composed by 11 sigma laye s, ha ing a high esolu ion in he op laye accoun ing o he wind influence, he bo om domain has 16 z- le el laye s. The WRF model (Wea he Resea ch and Fo ecas ing Model) p o ides he su ace bounda y condi ion, applied o he Rías Baixas a ea wi h 1.3 km ho izon al esolu ion. A de ailed examina ion o his modelop- e a ed by Me eoGalicia can be ound in Huhn e al. (2012) and Venâncio e al. (2019). Fu he mo e, he model also con ains he Ve dugo-Oi abén i e s discha ges h ough he SWAT model. Besides, ROMS coas al model is also used as bounda y condi ions. 2.3. Lag angian model To ack pa icles, MOHID-Lag angian module uses he concep o ace , whose posi ion (x,y,z) is well-known a each ime s ep. The mo e- men o he ace s depends on he eloci y om he hyd odynamic mod- ule, he wind om he su ace module and he andom eloci y om he wa es ac ion. The eloci y field is in e pola ed om he hyd odynamic eule ian eloci y field o ob ain a new g id wi h a esolu ion se by he use . To calcula e he posi ion o he pa icle, his eloci y field and a di u- si e e m a e in eg a ed as: dxi d ¼ ixi ðÞ, ðÞþDi(1) whe e i is he eloci y field a a ins an and x i posi ion, and D i is he di u- sion eloci y, calcula ed as de Pablo e al. (n.d.) desc ibed. The s eng h o his module lies in he pa ame e iza ion o beaching. I is possible o gi e a h eshold dep h alue abo e which he pa icle is s anded in he sand, as well as a p obabili y ha he pa icles will be s anded. These wo a gu- men s can accoun o di e en scena ios in which accumula ion can occu . Ano he obus ness o his model is ha i can be sepa a ed he accu- mula ions on he coas acco ding o he emission sou ces. This means ha he influence o each emission sou ce can be sepa a ely e alua ed. In his s udy, h ee di e en scena ios we e conside ed (see Table 1 in he Supplemen a y ma e ial).Fi s , he e ec o accumula ion along he en- i e coas is calcula ed o a ime pe iod o 1.5 yea s ( om Oc obe 1, 2018 o Ap il 1, 2020). Wi hin his pe iod, h ee di e en wind o ien a ion si u- a ions we e selec ed. Fo each, pe cen iles P25, P50, P75 and P95 we e cal- cula ed wi h espec o he maximum accumula ion o all segmen s. This makes i possible o dis inguish be ween high and low accumula ion zones and o show he spa ial dis ibu ion o ma ine li e wi hin he Ría de A ousa. The second scena io akes place in he Illa de A ousa om Oc o- be 1, 2018 o June 1, 2019; and he hi d scena io, om Oc obe 1, 2020 o Ap il 1, 2021, akes place along he no h o he Ría de A ousa sho e. These wo las ime pe iods ha e been chosen acco ding o he wo beach moni- o ing da a he e p esen ed. To es ima e he accumula ion a ios along he coas al line, he sho e o he Ría de A ousa is di ided in o 427 segmen s (see Fig. 4). These segmen s (also called bu e s) ha e 1000 m leng h by 100 m o wid h each. These dimensions has been se o ensu e he bes e- p oduc ion o he coas line, since a lowe esolu ion canno ensu e he co - ec defini ion along he coas . Because somes udy egions a e composed o di e en beaches wi hin one single sho e segmen , especially in he second case, he s udy egions we e joined by segmen s, he e e e enced as S1, S2, S3, S5, S6, S7, S8 and S9 (see Fig. 3 ). In he hi d scena io, nine beaches we e selec ed. A o al o 18 s udy a eas a e a ailable o compa ison o sim- ula ed and expe imen ally collec ed da a. Since he ex en o he bu e may be la ge han he ex en o he sho e whe e collec ions ake place, he es i- ma ed accumula ion will be g ea e han he ac ual accumula ion. The e- o e, he esul s ob ained om he simula ion a e co ec ed using an a ea ac o . This ac o is calcula ed as he a io o he moni o ed beach a ea o he compu a ional bu e a ea. The simula ion domain co e om 9.15°W o 8.15°W and om 42.2°N o 42.6°N wi h a 200 m spa ial esolu ion and 3 h empo al esolu ion. Ac- co ding o he mussel a dis ibu ion, he emission poin s we e es ablished a he same loca ion (see Fig. 1c). Fo he en i e simula ion pe iod, Lag ang- ian pa icles a e eleased om all emission poin s e e y hou . Mussel-pegs collec ed on he sampling beaches can come om o he es ua ies wi h shell- fish ac i i y. To conside his e ec , wo emission poin s we e also consid- e ed a he no he n and sou he n edges o he simula ion domain. Fo each polygon, emission poin s a e assumed e e y 140 m along he polygon geome y. In he case o ex e nal sou ces, we define a egula emission wi h an a ea o 9 km 2 . Because he di e si y o beaches in he Ría de A ousa, sim- ula ions ha e been ca ied ou a ying he p obabili y o s anding. This is in ended o simula e h ee possible e en ion scena ios. The 80% beaching ac o eflec s a eas o high e en ion, 50% beaching ac o eflec s medium e en ion and 20% beaching ac o eflec s a eas o low e en ion, mainly ocky ones. Then, ends we e calcula ed using ela i e accumula ion (% RA) o each beach. Fo each s udy a ea, he mean accumula ion e e y 4 h is ob ained. This is in ended o smoo h possible a ypical e ec s due o pa icula synop ic wea he condi ions. Fo each ime pe iod conside ed, he o al accumula ion (A TOT ) and he maximum accumula ion be ween he whole conside ed s udy a eas (A MAX ) a e ob ained. The ela i e pe cen - age o each egion is ob ained by di iding he accumula ion o ha a ea by he maximum accumula ion as, %RA ¼ATOT AMAX (2) 2.4. Obse a ional da a. Beaches collec ion The moni o ing da a p esen ed he e is pa o one o he s a egies o he CleanA lan ic p ojec . This p ojec aims o p o ec biodi e si y and ecosys- em se ices in he A lan ic A ea by imp o ing capaci ies o moni o ing, p e en ion and emo al o mac o ma ine li e . Thep ojec alsocon ibu es o aising awa eness and imp o ing ma ine li e managemen sys ems. As pa o his p ojec , a s udy has been conduc ed o analyze di e en s a e- gies o iden i ying ma ine deb is accumula ion a eas ha can be inco po- a ed in o moni o ing, cleanup and collec ion p o ocols. One o he s a egies es ed was he use o mussel a s as an indica o o plas ic deb is accumula ion. These indica o s, along wi h o he measu es and p oxies, (i.e., oceanog aphic cu en s, socioeconomic da a), can be used o de ec he p esenceo ma ineli e in each habi a andspa ial and empo al ends in accumula ion (GESAMP). The e o e, he sampling plan was designed wi h he in en ion o de e - mine he main egions o li e accumula ion along he island's coas line and whe he he e a e any specific loca ions ha s and ou om he es . The chosen beaches me he ollowing equi emen s (1) o be o med by sand o g a el and exposed o he open sea, (2) o be accessible o he sam- ple s h oughou he yea , as well as o acili a e he emo al o ma ine li - e , and (3) no o be subjec o o he li e emo al ac i i ies. A weekly equency was es ablished, and da a collec ion was always conduc ed one hou a e high ide o a oid sampling being a ec ed by he ising ide. A each collec ion, he numbe o mussel-pegs ound, bo h whole and b oken, was coun ed and emo ed om he beaches. The sampling o all selec ed beaches was ca ied ou in wo campaigns: A) Illa de A ousa: ca ied ou be- ween Oc obe 2018 and Ap il 2019. The poin s loca ed in A Illa de A ousa S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 5 we e sampled as pa o a ci izen science p ojec . B) Ría de A ousa: ca ied ou be ween Oc obe 2020 and Ma s 2021. The poin s loca ed on he coas o his es ua y we e sampled by expe echnicians in coas al sampling who usually ca y ou bo h sampling wo k on he selec ed beaches and ad ice o he shellfish sec o as pa o hei daily du ies. 3. Resul s and discussion Fig. 4a shows he spa ial dis ibu ion o accumula ions o an emission pe iod o one yea and a hal ob ained om he compu a ional model. Knowing he empo al dis ibu ion o accumula ions o each coas al seg- men , he o al accumula ion was calcula ed. The P25, P50, P75 and P95 (25 h, 50 h, 75 h, 95 h pe cen iles) we e used o es ablish he h esholds o e y low, low, medium, high and e y high accumula ions. The a eas wi h he lowes deb is accumula ion a e a e he ou e mos , i.e., hose mos exposed o oceanic ac ion. On he o he hand, g ea e accumula ions appea owa ds he inne pa o he es ua y. Gi en ha he cu en s en e om he sou h, in a no heas e ly di ec ion, and d ain owa ds he no h in a no hwes e ly di ec ion, he main accumula ions appea in he coas al segmen s o ien ed owa ds he sou h and sou heas , coinciding wi h his mo emen . This map (Fig. 4) shows wo clea ly defined accumu- la ion zones. One in he no heas o he es ua y and he o he in he no h- wes o he cen al island (A Illa). Likewise, he en i e in e media e inne egion shows high accumula ion alues abo e P50. Tu ning now o he s udy o wind ac ion, Fig. 4b co esponds o an eas e ly wind, Fig. 4c con- side s a sou heas e ly wind ac ion and Fig. 4d conside s a no hwes e ly wind ac ion. These h ee si ua ions demons a e he significan influence o his agen . Thus, an eas e ly wind si ua ion will cause a g ea e accumu- la ion on he coas line wi h he same o ien a ion. In he case o a p edomi- nan ly no heas e ly wind, he d ainage o he es ua yis a o ed. The e o e, he inne mos zone egis e s a lowe accumula ion, while he ou e mos e- gions wi h he same o ien a ion egis e highe accumula ion alues. Fo he las si ua ion, a p edominan ly sou hwes e ly wind a o s he high ac- cumula ions on beaches wi h he same o ien a ion and loca ed in he in e- io egion. This e ec is smoo hed in he simula ion esul s, since he a e age accu- mula ion e e y 4 h has been calcula ed o he se o bu e s; and hen he empo al accumula ion pe bu e has been es ima ed. The esul s a e p e- sen ed in wo pe spec i es: (1) o he whole simula ion pe iod and (2) sep- a a ed by mon hs. 3.1. Illa de A ousa Compa ing now he moni o ing da awi h he simula ed da a, he esul s can be sepa a ed by segmen s. Fo he fi s ime pe iod ( om he 2018 Oc- obe 1 o 2019 May 1), Fig. 5 shows he esul s o nine di e en coas al line segmen s. I is impo an o no e ha he compa ison e e s o ends, no o ne da a, i.e. bo h da a se ies (simula ed and obse a ional) ha e Fig. 4. Accumula ion a ios in he Ría de A ousa along he 427 segmen s o a pe iod o one and a hal yea s ( om 2018 Oc obe 1 o 2020 Ap il 1) ob ained om he compu a ional model. The classifica ion ‘ e y low’co esponds o accumula ion alues below P25; ‘low’, be ween P25 and P50; ‘medium’, be ween P50 and P75; ‘high’, be ween P75 and P95; and ‘ e y high’, abo e P95. The black a ow shows he wind di ec ion. a) Conside s he whole ime pe iod, b) co esponds o an eas e ly wind on he 30 h Janua y 2019, c) co esponds o a sou heas e ly wind ac ion on he 7 h May 2019 and d) co esponds o a no hwes e ly wind ac ion on he 10 h Janua y 2019. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 6 been no malized acco ding o he maximum ound in each s udy egion. In all o hem he s anding ac o has been a ied o see which one es ima ed a be e accumula ion. As can be seen, in he cases whe e i di e s, he 20% s anding ac o fi s be e . Since he Illa de A ousa a ea is a mos ly ocky a ea (see Fig. 3 ) i is highly unlikely ha mussel-pegs each sand a eas. Howe e , i can be s a ed ha he compu a ional modelis able o ep oduce he gene al end o accumula ion in he s udied a eas, wi h he excep ion o S4. In his case, he bu e used o he accumula ion coun does no eli- ably ep oduce he sho eline. Because he spa ial esolu ion o he bu e s along he coas al line, he e a e cases whe e he sho e p ofile is no well enough eplica ed. In he case o S4, he bu e o ien a ion is sligh ly di e - en om he ac ual o ien a ion o he moni o ed beach. This o ien a ion causes he simula ed ci cula ion o ba ely each his egion, so ha a smalle numbe o pa icles a e always ound. Finally, i was calcula ed he ela i e con ibu ion o he wo ex e nal emission sou ces. In his case, only 3% o he pa icles ha ha e eached he coas come om ex e - nal sou ces. This makes sense, since hese sou ces will a ec he ou e mos egions, which a e he mos influenced by oceanic ac ion. So i makes sense ha only a small ac ion o he pa icles emi ed by hese sou ces will be s anded on he coas . Fig. 6 shows he mon hly compa ison be ween simula ed and mon- i o ing da a. This highe empo al esolu ion shows mo e disc epancies han in he p e ious ep esen a ion. He e i is impo an o iden i y he causes ha can p oduce dispa i y be ween simula ed and moni o ing esul s. The fi s and mos impo an is he ela ionship be ween he s udy a ea o ien a ion and he in e idal mo ion. The mo e he beach is o ien ed owa ds he idal mo emen , he la ge he numbe o com- pu a ionally accumula ed pa icles is. Howe e , he ba hyme y o he egion can lead o accumula ion alues below he compu a ional es i- ma ion. Since he ocky egions nea he coas may ac as low e en ion ac o s, he moni o ing da a on he beach may be lowe han expec ed. In he case o he Illa, his is wha happens in S2, S3, S5, S6. All hese egions a e hea ily a ec ed by in e idal mo ion. This means ha he g ea e he numbe o pa icles passing h ough an in e media e zone, he g ea e he accumula ion eco ded in ha zone, since 80% o hepa iclesp esen a es andedon hesand(in hecaseo se ing he s anding a 80%). This ac is clea ly mo e accen ua ed in S7 and S9 due o he ac ha hei sou he n o ien a ion esul s in a highe numbe o pa icles being ecei ed in bo h egions du ing ising ides. Fig. 3 shows ha all a eas o he island ha e ocky in e idal zones ha will ha e an e ec on he moni o ing da a a he beaches, as his alue is usually lowe han ha es ima ed by he model. In cases whe e he collec ion alue is highe , i may be due o a pa icula synop ic si ua ion on he pa icula collec ion day. Being he wes e n- mos a ea o he Ibe ian Peninsula, Galicia, and in pa icula he wes coas , o en su e s om a ious a mosphe ic phenomena coming om he No h A lan ic. Specifically, in he mon h o Oc obe 2018, wo hu icane e en s eached he coas o Galicia, leading o an anom- alous synop ic si ua ion. The mon h o No embe 2018 was cha ac e - ized by he appea ance o a g ea s o m, causing a ypical wind si ua ions. Finally, he mon h o Decembe 2018 was cha ac e ized by high al i ude squalls du ing he fi s 20 days o he mon h. All hese synop ic si ua ion can a ec he moni o ing da a. The case o S4, as men ioned abo e, p esen s a ep esen a ion p oblem, since he o ien a ion o he bu e ep esen a ion means ha ha dly any pa icles each his a ea. Finally, S1 and S8 a e he bes es ima ed egions du ing he whole pe iod. Respec ing o he beaching pa ame e iza ion, in his pa icula s udy, no di e ences we e obse ed. 3.2. Ría de A ousa Fo heRíadeA ousa, hesame endsasin hep e ioussubsec ion we ecalcula ed.Fo heaccumula iono e heen i esimula ionpe- iod,simila esul sa e oundin he endo hesimula edda awi h e e ence o he sampled da a (see Fig. 7). Ba aña beach eaches he maximum accumula ion alue bo h in he simula ion and in he beach collec ions, while he minimum alue is es ima ed o Ninei iños by using bo h me hodologies. Fo his beach, he moni o ed da a show a concen a ion o 0.31% wi h espec o he o al li e collec ed on all beaches sampled. I is possible o find a clea dispa i y be ween simula- ion and obse a ion o Ca ei o beach. In pa icula , his beach is lo- ca edinahighlyexposeda eao hees ua y ooceanicac ion.In addi ion, i s posi ion makes i possible ha he accumula ions ob- se ed on he beach come om poin s ou side he es ua y, pa icula ly om he mussel cul i a ion a eas o Ría de Pon e ed a. Al hough his ac has been ied o o e come his ac wi h wo ex e nal emission sou ces, he pe cen age o pa icles eaching he coas om ex e nal sou cesisonly3%.He e, hebeha io o beaching o he las s udy e- gions is simila o he p e ious case. The highes alues o ela i e ac- cumula ion occu o a beaching alue o 80% while he lowes alues a e eco ded o a beaching o 20%. The impo ance o beach o ien a ion in ela ion wi h he in e idal mo ion is obse ed in Fig. 8 o he mon hly accumula ion. This is he case o he Illa and Fon án beaches. The alues es ima ed by he model a e highe han hose collec ed on beaches because he in e idal mo emen makes i di ficul o mussel-pegs o accumula e on hese beaches. Oceanic influence mus also be aken in o accoun . As in he p e ious case, in he egions mo e exposed o oceanic ci cula ion he model will p esen a lowe accumula ion alue. Compu a ionally, o he Ca ei o and Cas iñei as egions his beha io is ound. Howe e , in he case o Ca ei o, high accumula ion alues a e eco ded in he beach collec ions. This is due o he ac ha , despi e being in one o he ou e mos a eas o he Ría, i s o ien a ion does no a o he es ua y dynamics o clean his a ea. Mo eo e , being a ocky a ea, he dynamics in his egion will be slowe in eali y. This is no he case o he model. In he case o Cas iñei as, i s o ien a ion makes he oceanic ac ion ha e a g ea e e ec , so he accumula ion alues a e low in bo h me hods. Mo eo e , i may happen, as in he case o Aduana, ha an h opogenic cons uc ions such as a dyke a e no included in he coas al managemen plans, so hey a e also no aken in o accoun in simula ions o his ype. Howe e , in eali y, hese ypes o cons uc ions ha e an e ec ha is eflec ed in he moni o ing da a. The e o e, he Lag angian model on his beach always epo s la ge alues han hose ac ually ound in he collec ed ones. Finally, o iden a ions ha do no p esen any o hese ac o s, as i is he case o Ba aña and Po co, he moni o ing da a on he beach can be well ep oduced compu a ionally. Du ing his s udy pe- iod, he mon hs o Janua y and Feb ua y we e cha ac e ized by he oc- cu ence o s o ms ha a ec ed he egion. In hese wo mon hs, Galicia su e ed om ou di e en s ong squalls. In ou compa ison, bo h mon hs s and ou o ha ing dispa i ies be ween simula ed da a and moni o ing da a. Wi h espec o beaching, i can be concluded ha no significan di e ences can be seen in mos cases. Howe e , in hose cases in which he ci cula ion a o s he appea ance o a g ea e numbe Fig. 5. Compa ison o ends in he accumula ion o mussel-pegs o di e en sho e segmen s in he Illa de A ousa om Oc obe 2018 o Ap il 2019. Simula ions ha e been pe o med o di e en beaching alues: 20% (ligh o ange), 50% (medium o ange) and 80% (da k o ange). The moni o ing da a a e shown in b own. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 7 o pa icles o e a gi en s udy egion, he pe cen age es ablished mus be aken in o accoun , since he highe he pe cen age, he g ea e he dis- pa i y be ween esul s wi h di e en beaching may appea , as in he case o Po co beach in he mon hs o Oc obe and Ma ch. These esul s indica e he MOHID-Lag angian model is able o es ima e easonably well he accumula ion o s icks bo h on he Illa de A ousa and on he Ría de A ousa. The e a e pa icula cases, such as some synop ic si ua- ions, which cause dispa i ies be ween he model esul s and he Fig. 6. Compa ison o ends in he accumula ion o mussel-pegs o di e en sho e segmen s in he Illa de A ousa o a mon hly ep esen a ion om Oc obe 2018 o Ap il 2019. a) Oc obe 2018, b) No embe 2018, c) Decembe 2018, d) Janua y 2019, e) Feb ua y 2019, ) Ma s 2019, g) Ap il 2019. Simula ions ha e been pe o med o di e en beaching alues: 20% (ligh o ange), 50% (medium o ange) and 80% (da k o ange). The moni o ing da a a e shown in b own. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 8 moni o ing da a collec ed on he beach. Ano he sou ce o dispa i y is he ba hyme y o he accumula ion zones, whe e a high esolu ion model should be implemen ed. 4. Conclusions The g ow h o plas ic inpu s in o he seas and open ocean is a majo p oblem. Howe e , floa ing plas ic deb is can be used as an ideal ace o de e mine a eas o accumula ion as well as hei poin s o o igin. This ype o s udy is possible h ough he use o Lag angian me hods. This pape p esen s a ecen ly de eloped Lag angian ool ha has been alida ed using a floa ing plas ic; mussel-pegs. No e ha , al- hough heda a alida iono hesemodelshas obedoneona egional scale, hey can be implemen ed on an in e na ional scheme. Fu he - mo e, he obus ness o his s udy lies in he ac ha he emi ing sou ces a e well known. Fo his eason, i is easible o use such mon- i o ing da a o alida e he MOHID-Lag angian model. Howe e , he p ecise compu a ional ep oduc ion o his si ua ion is complex, since he ac ual emission om he a s is unknown, as mussel s icks a e no emi ed a each mussel collec ion and he mussel collec ion is no pe o med pe iodically. In addi ion, he esidence ime o he mussel s icks is no de e mined and i is no possible o di e en ia e whe he he o igin o a pa icula plas ic s ick comes om he a o his es ua y o om ano he . Fu he mo e, he whole ecep o poin s a e equally p obable o ecei e ma ine li e accumula ion, i.e. all sho eline seg- men s can eco d accumula ion wi h he same p obabili y. Thus, he model igno es ocky egions. The e o e, i is possible o find acc e ion zones in ocky egions whe e no eal acc e ion occu s. In addi ion, du - ing sampling, he sa e y o he sample s was always he op p io i y. Thus, sampling was always conduc ed du ing dayligh hou s and was suspended in cases whe e sa e y could be comp omised (ad e se wea he condi ionsandCOVIDp o ocols)Howe e ,asi isbasedon he oppo uni y gi en by ci izen science and he sampling wo k o he p o essionals, he eco ds we e made weekly on di e en days o he week. This means ha , in his case, i will be mo e di ficul o cha ac- e ize he possible wea he si ua ions ep esen ed by he expe imen al da a. Fo each o he beaches sampled, a specific o mwasde eloped o eco d in o ma ion on mussel-pegs accumula ion a eas. In addi ion, he samplings we e comple ed wi h pho og aphs o e i y he eliabil- i y o he da a. All hese ac o s can ha e an impac on he compa ison esul s. In his s udy i has been used a egional Lag angian model o s udy he accumula ion o ma ine li e in he Ría deA ousa. Th ee di e en ime sce- na ios ha e been simula ed: (1) one and a hal yea s o see he global end; (2) se en mon hs o compa e wi h he moni o ing da a in he Illa de A ousa and (3) six mon hs o compa e wi h he moni o ing da a in he Ría de A ousa. This kind o s udies a e highly ele an o es whe he compu a- ional models, in his case he MOHID-Lag angian, a e capable o e- p oducing he obse a ional esul s. The fi s esul s show he spa ial dis ibu ion o ma ine li e in he Ría de A ousa. Using he P25, P50, P75 and P95, he accumula ion zones along he coas line can be quali- a i ely di e en ia ed. The a eas mos exposed o oceanic ac ion a e hose wi h he lowes accumula ion a ios. Due o he ci cula ion o cu en s, he la ges accumula ions occu a he ou e mos o he cu - en sin hees ua y.I wasalso ound ha windhasag ea influence on he modelling. The acc e ion zones a y depending on he wind di- ec ion. As o he compa ison be ween he model and he moni o ing da a collec ed on he coas s, highly a o able esul s ha e been ound o bo h s udy pe iods. In bo h scena ios i was possible o ep oduce hegene al endo eacha ea o heen i e imepe iodwi h heexcep- iono wo egionsin hecaseo heIlla,andonein hecaseo heRía. In he case o he Illa, hese wo dispa i ies a e di ec ly ela ed o he o i- en a ion o he bu e wi h espec o he dynamics o he es ua y. In one case, he compu a ional exposu e o he dynamics o his egion does no co espond o he eal one, a o ing he a i al o ewe pa icles in his s udy a ea. In he o he case, he o ien a ion o he bu e is o ally a o able o he oceanic ci cula ion, which causes a high numbe o pa icles o be s anded in his a ea. A simila mon hly beha io is ound o all he s udy zones, excep o he mon hs o No embe and Decembe . In hese cases, he influence o No h A lan ic hu icanes could be he cause o hese anomalous beha io s. Fo he Ría, he o e all end is well ep oduced by he Lag angian model, wi h he excep ion o Ca ei o beach. The la ge a iabili y o his egion makes i , in a compu a ional sense, one o he cleanes a eas in his s udy. Howe e , i s pa icula o ien a ion makes he accu- mula ions moni o ed in his beach highe han expec ed. On a mon hly basis, he impo ance o he o ien a ion o he beach wi h espec o he in e idal dynamics is also obse ed. Thus, he impo ance o aking in o accoun a co ec ep oduc ion o he coas al p ofile is also highligh ed. Al hough he esul s ob ained in his s udy ha e been sa is ac o y, we would like o conclude his a icle wi h a se ies o ecommenda ions o he design o u u e s udies, as well as o moni o ing da a. The knowledge o he s udy a ea is impo an o e alua e whe he i can be ep esen ed ai h ully in a compu a ional way. Spa ial esolu ion p oblems can lead o unexpec ed and e oneous esul s, as i was ou case wi h he S4 segmen . The compu a ional o ien a ion o he co e- sponding bu e caused ha only ew pa icles eached his egion, while a mo e a o able o ien a ion and close o he eal one, he accu- mula ion in his a ea could ha e been be e es ima ed. As has been shown, ba hyme y plays a majo ole and i mus be conside ed o he co ec in e p e a ion o he egion dynamics. In addi ion, i is impo an o ake his ype o compa ison wi h cau- ion, gi en ha compu a ionally we ob ain a con inuous ime se ies o ma ine deb is accumula ion a he sho eline. Howe e , he esul s ob- ained in he collec ion co espond o ce ain si ua ions ha depend on mul iple ac o s and ha may no be ep esen a i e o he empo al end o accumula ion in a pa icula egion. These ac o s can be clas- sified as en i onmen al and human a iables and bo h can ha e a de - imen al e ec on he ob ained esul s. En i onmen al ac o s a e impossible o con ol, bu i is impo an o keep hem in mind when analyzing he esul s. As we ha e seen, synop ic si ua ions a e key o lo- ca e po en ial ma ine li e accumula ion a eas so he es ablishmen o consolida ed campaigns ha conside hese condi ions o define he moni o ing pe iods is highly ecommended. Human in e e ence is eas- ily con ollable. We would like o highligh some indica ions o be aken in o accoun in u u e s udies. Mus be ensu ed ha he o al o he planned samples a e ca ied ou a he scheduled da e and ime. All sampling mus be gone on he same day and a e he peak o high ide, p e e ably wo hou s la e . This ensu es ha all floa ing pa icles Fig. 7. Six mon h compa ison o ends in he accumula ion o mussel-pegs o di e en sho e segmen s in he Ría de A ousa, om Oc obe 2020 o Ma ch 2021. Simula ions ha e been pe o med o di e en beaching alues: 20% (ligh g een), 50% (medium g een) and 80% (da k g een). The moni o ing da a a e p esen ed in da kes g een. S. Cloux e al. Science o he To al En i onmen 822 (2022) 153338 9