Analysis of the spatio-temporal evolution of dredging from satellite images: a case study in the Principality of Asturias (Spain)

Jou nal o Ma ine Science and Enginee ing A icle Analysis o he Spa io-Tempo al E olu ion o D edging om Sa elli e Images: A Case S udy in he P incipali y o As u ias (Spain) Vanesa Ma eo-Pé ez 1, Ma ina Co al-Bobadilla 2,* , F ancisco O ega-Fe nández 1 and Vicen e Rod íguez-Mon equín1   Ci a ion: Ma eo-Pé ez, V.; Co al-Bobadilla, M.; O ega-Fe nández, F.; Rod íguez-Mon equín, V. Analysis o he Spa io-Tempo al E olu ion o D edging om Sa elli e Images: A Case S udy in he P incipali y o As u ias (Spain). J. Ma . Sci. Eng. 2021,9, 267. h ps://doi.o g/10.3390/jmse9030267 Academic Edi o : Rodge Tomlinson Recei ed: 5 Feb ua y 2021 Accep ed: 25 Feb ua y 2021 Published: 2 Ma ch 2021 Publishe ’s No e: MDPI s ays neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a il- ia ions. Copy igh : © 2021 by he au ho s. Licensee MDPI, Basel, Swi ze land. This a icle is an open access a icle dis ibu ed unde he e ms and condi ions o he C ea i e Commons A ibu ion (CC BY) license (h ps:// c ea i ecommons.o g/licenses/by/ 4.0/). 1P ojec Enginee ing Depa men , Uni e si y o O iedo, 33004 O iedo, P incipali y o As u ias, Spain; [email p o ec ed] (V.M.-P.); [email p o ec ed] (F.O.-F.); [email p o ec ed] (V.R.-M.) 2Depa men o Mechanical Enginee ing, Uni e si y o La Rioja, 26004 Log oño, La Rioja, Spain *Co espondence: [email p o ec ed]; Tel.: +34-941-299-274 Abs ac : One o he undamen al asks in he main enance o po ope a ions is pe iodic d edging. These d edging ope a ions acili a e he elimina ion o sedimen s ha he coas al dynamics in oduce. D edging ope a ions a e inc easingly es ic i e and cos ly due o en i onmen al equi emen s. Unde s anding he condi ion o he seabed be o e and a e d edging is essen ial. In addi ion, de e mining how he seabed has beha ed in ecen yea s is impo an o conside when planning u u e d edging ope a ions. In o de o analyze he beha io o sedimen anspo and he changes o he seabed due o sedimen a ion, s udies o li o al dynamics a e conduc ed o model he deposi ion o sedimen s. Ano he me hodology ha could be used o analyze he eal beha io o sedimen s would be o s udy and compa e po ba hyme ies collec ed pe iodically. The p oblem wi h his me hodology is ha i equi es nume ous ba hyme ic su eys o p oduce a su icien ly signi ican analysis. This s udy p o ides an e ec i e solu ion o ob aining a dense ime se ies o ba hyme y mapping using sa elli e da a, and enables he pas beha io o he seabed o be examined. The me hodology p oposed in his wo k uses Sen inel-2A (10 m esolu ion) sa elli e images o ob ain his o ical ba hyme ic se ies by he de elopmen o a andom o es algo i hm. F om hese his o ical ba hyme ic se ies, i is possible o de e mine how he seabed has beha ed and how he en y o sedimen s in o he s udy a ea occu s. This me hodology is applied in he Po o Lua ca (P incipali y o As u ias), ob aining sa elli e images and ex ac ing successi e ba hyme y mapping u ilizing he andom o es algo i hm. This wo k e eals how once he dock was d edged, he sedimen s we e edeposi ed and he seabed eco e ed i s le el p io o d edging in less han 2 mon hs. Keywo ds: Lua ca po ; andom o es ; d edging ac i i ies; sedimen anspo ; sa elli e images 1. In oduc ion D edging ac i i ies a e commonly used in coas al a eas o main ain he designed dep h o na iga ion channels o basins and o emo e deposi ed sedimen s. These ac i i ies include no only he p ocesses o emo ing sedimen om he bo om, bu also hei subsequen anspo o ano he loca ion. D edging o po s imp o es access and exi condi ions o ca go and passenge ships [ 1 , 2 ]. This is impo an o ma ine sa e y, emo ing sedimen ha has accumula ed in channels in o de o main ain he designed dep hs o he exis ing acili ies [ 3 , 4 ]. Howe e , despi e i s impo ance in ma i ime wo ks and i s e ec on economic and social de elopmen , and he cu en ad ances in d edging echniques, he sil a ion o po s con inues o be one o he leas unde s ood b anches o coas al enginee ing. I is o g ea impo ance o po s, in main aining and imp o ing i s dep h and, also, in de eloping new acili ies o c ea ing new po s [ 5 ]. Thus, i is pa icula ly impo an o s udy he sedimen a ion ha akes place in he basins and en ance channels o po s. Excessi e sedimen a ion can impai he unc ioning o he po and lead o a dec ease in J. Ma . Sci. Eng. 2021,9, 267. h ps://doi.o g/10.3390/jmse9030267 h ps://www.mdpi.com/jou nal/jmse J. Ma . Sci. Eng. 2021,9, 267 2 o 18 economic ac i i ies [ 6 – 8 ]. The cons uc ion o dikes and o he coas al p o ec ion wo ks ha e a g ea in luence on hyd odynamic o ces and mo emen o sedimen in he o sho e a ea [ 9 ]. Hyd odynamic and sedimen a y condi ions can cause sedimen a ion o he po channel, which mus be emo ed by main enance d edging [10]. The Po o Lua ca is loca ed in he P incipali y o As u ias and is one o he mos impo an po s in no he n Spain. In ecen yea s he e ha e been signi ican changes in he ma ine dynamics o he po , as well as in he e osion–sedimen a ion p ocesses ha ope a e he e. To main ain na iga ion, d edging is unde aken once a yea . Technical speci ica ions include he es ima ion o sil a ion olume, loca ion, du a ion and phases o d edging wo ks o be pe o med, he d edging me hod and he loca ion o sui able sedimen dumping loca ions ha will no a ec he en i onmen ad e sely [ 11 ]. A Lua ca, sandy bo oms p e ail, al hough se e al a eas a e p esen ly co e ed by mud [ 12 ]. The pu pose o d edging he dock a ea is o gua an ee he unc ionali y o he po , which is condi ioned by he dep h o he dock. The loca ion o he Po o Lua ca a he mou h o he Rio Neg o a o s he accumula ion o sedimen a y ma e ials o lu ial o igin. The dock’s loca ion o es alls he idal dynamics ha a e necessa y o wash away he deposi ed sedimen s. Fo his eason, pe iodic d edging is necessa y o main ain he dock. This enables op imal le els o use o he di e en po a eas o be achie ed. Th oughou i s ope a ional li e, he po has aced signi ican sedimen a ion p oblems and a consequen dec ease in he dep h o he d a in i s en ance channel. The p oblem o sedimen a ion is inc easing due o he en i onmen al p oblems ha a e associa ed wi h d edging and he es ic i e egula ions imposed as a esul . This sedimen a ion p oblem may be ela ed o he posi ioning and layou o he po basin and i s en y loca ion [ 13 , 14 ]. Due o high sedimen a ion a es, he basin and en ance channel o he Po o Lua ca mus be d edged egula ly o he po o con inue i s ope a ion. Many di e en me hodologies ha e been used in an e o o unde s and he beha io o sedimen s. These ha e anged om he use o p obabilis ic design me hods [ 15 – 18 ] o di ec nume ical simula ions [ 19 – 21 ]. Di ec nume ical simula ions p o ide a ough unde s anding o he beha io o sedimen s. Howe e , i is a cos ly me hodology, because i equi es he su ey o he en i e a ea in o de o unde s and i s 3D geome y. Ano he echnique ha could be used o s udy he beha io o sedimen s is o conduc an analysis o he ba hyme ies ha we e ob ained in he s udy a ea. I he e olu ion o he seabed is known, i is possible o know whe e he sedimen deposi s a e p oduced and hei sedimen a ion a e. This echnique would also help o de e mine he need o unde ake d edging, as well as i s pe iodici y. The main limi a ion when analyzing he beha io o he seabed is he a ailabili y o ba hyme ies. Ba hyme ies a e no mally conduc ed be o e and a e d edging, bu no equen ly enough o unde s and he e olu ion o he seabed, due o hei high cos and he di icul y o ca ying hem ou . This clea ly limi s hei use in analyzing he beha io o he seabed. The e o e, o his pu pose, ba hyme ies a e ob ained om sa elli e images using he simples echniques ha o e a high le el o p ecision. In addi ion, he echniques mus be applicable in a eas o u bid and shallow wa e and p o ide apidi y and lexibili y o use. Con en ional ba hyme ic me hods usually p o ide dep h p o iles o poin measu e- men s o accep able accu acy. Howe e , hei use is cos ly and ine icien . Ba hyme ic es ima ion can be imp o ed by combining echo sounding and sa elli e da a. Remo e sens- ing echnology is used in many opog aphical s udies. This is e y use ul o si ua ions in which he dep hs o wa e a e equi ed on empo al and spa ial scales, bu a e di icul o ob ain [ 22 – 25 ]. Simple me hods ha use op ical images in es ima ing dep hs a e used by some in es iga o me hods, as well as linea eg ession algo i hms o es ima e he dep h o wa e [ 26 ]. In ecen yea s, he use o machine lea ning echniques and op ical senso s o es ima e ba hyme ies has gained popula i y. This can be a ibu ed o ad ances in he de elopmen o algo i hms, as well as g ea e a ailabili y o da a and new sa elli es. Many au ho s ha e sough o de e mine he dep hs o wa e dep hs om op ical senso s and he use o eg ession models ha we e de i ed om machine lea ning echniques [ 27 – 32 ]. The J. Ma . Sci. Eng. 2021,9, 267 3 o 18 andom o es algo i hm was used in his s udy o ob ain he ba hyme ies. This algo i hm has a numbe o ad an ages, including, low-cos , simplici y, ime-e ec i eness, and i s wide-co e age o shallow wa e . Thus, he andom o es algo i hm has been ound o be applicable o cons uc eg ession models ha ely on ba hyme y da a om sa elli e images [33–41]. Many s udies ha e concen a ed on mig a ion o sedimen s o e he long e m and how his a ec s wa e quali y and he ecosys em condi ion. They model mo emen o ma e ial ha is emo ed du ing he d edging ope a ion [ 42 ]. Howe e , op imizing he d edging ope a ion has no been p oposed. The main objec i e o his wo k is o de e mine i he beha io o sedimen a ion di e s wi h he dep h o d edging. In o he wo ds, is he e a s able mean dep h ou side o which e osion and sedimen a ion occu mo e quickly? This is possible due o he capaci y o he me hodology, which enables da a om he pas wi h a equency o 1–2 mon hs o be analyzed. This p o ides g ea in o ma ion o he seabed’s beha io . 2. S udy Si e The Po o Lua ca is si ua ed in no hwes e n Spain on he coas o he Can ab ian Sea a a longi ude o 6 ◦ 32’1” W and a la i ude o 43 ◦ 32’45” N (Figu e 1a). Lua ca has been associa ed wi h ma i ime ac i i y since i began as a ishing encla e in he 10 h cen u y. Se e al changes ha e occu ed h oughou pas yea s in Lua ca Po . In 1910, he Canouco dike (Figu e 1b), 40 m in leng h, was buil o shel e he ou e basin o he po . In 1940 a new dike, he La Enco onada dike, was buil (Figu e 1b), 124 m in leng h on he wes e n side [ 43 ]. I is necessa y o conduc d edging o he dock annually. The lack o d a means ha he boa s canno en e he po , as many cu en s a e gene a ed in he en ance channel. Thus, na iga ion in his a ea is endange ed by he isk ha he boa s will be s anded, especially du ing s o ms. The po is used gene ally by small boa s. Thus, he minimum d a s in he na iga ion channel ange om 2 o 3 m, wi h an a e age idal un o 2 m. The d a in he docking a ea is somewha less—2 m. Al hough d edging is conduc ed in he inle channels, i also a ec s he inne basin. O he wo a eas, i is he inne basin ha ecei es he mos in e en ions and is mos limi ed due o d a . Nalona, a d edge boa (Suc ion Hoppe D edge , IMO 9047453) [ 44 ], has been used o d edge he bed su ounding he po . A e he sand and mud ha e been collec ed om he bo om o he Po o Lua ca, he sand is deposi ed in he a ea o Pun a Muye es, in on o he hi d beach o he illage, in an e o o enew his a ea. Howe e , i is no e y clea whe he his discha ge loca ion bene i s he po , since a la ge amoun o he sand ha is dumped he e will be mo ed by he cu en s and e-en e he dikes. The mud, on he o he hand, is deposi ed a ano he loca ion ha is a he om he beaches. The da es on which he d edging is ca ied ou in he inne basin a e p o ided by he Po Se ice o he P incipali y o As u ias. D edging ope a ions we e ca ied ou in 2017 and 2018. In 2017, he d edging began in Oc obe , bu was s opped because o a p oblem wi h he d edge. I esumed in No embe and ended in Decembe , 2017. In 2018, d edging began on No embe 15 and con inued un il he end o he yea . In 2019, he inle channel was d edged, bu li le wo k was done in he inne basin. J. Ma . Sci. Eng. 2021,9, 267 4 o 18 J. Ma . Sci. Eng. 2021, 9, x FOR PEER REVIEW 4 o 18 Figu e 1. The Po o Lua ca: (a) posi ion o he po on he Can ab ian Sea coas ; (b) he s udy si e loca ion. 3. Da a and Me hodology 3.1. Da a In Si u Measu emen s and Sa elli e Da a Field measu emen s a e gene ally necessa y o s udy he sedimen a ion in he po , including he hyd odynamic pa ame e s o he a ea, in his case only ba hyme ies will be used as ield measu emen s, and hese ba hyme ies a e p o ided annually by he Po Se ice o he P incipali y As u ias. The ba hyme ies we e conduc ed wi h he use o a Na isound 210 single beam echo sounde , wi h a 1 cm e ical esolu ion and dual equency (Reson, Inc.: Slange up, Denma k). The speci ica ions o echo sounde Na isound 210 a e shown in Table 1. Table 1. Cha ac e is ics o echo sounde Na isound 210. Na isound 210/400 Speci ica ions F equency 190–235 kHz Po ency o ansmission 300 W Impedance 100 Ohm Echo app o al leng h 100 µs–210 kHz Dep h ange 0.5–100/400/1200 m depending on he equency Resolu ion 1 cm Accu acy 1 cm a 210 kHz (1 σ) assuming co ec sound eloci y, ansduce dep h e c. The s udy’s echo sounding da a o Lua ca we e de e mined, on 28 June 2016, 10 May 2018 and 28 May 2019. Fo he p esen wo k, XY posi ions om he su ey da a we e Figu e 1. The Po o Lua ca: ( a ) posi ion o he po on he Can ab ian Sea coas ; ( b ) he s udy si e loca ion. 3. Da a and Me hodology 3.1. Da a In Si u Measu emen s and Sa elli e Da a Field measu emen s a e gene ally necessa y o s udy he sedimen a ion in he po , including he hyd odynamic pa ame e s o he a ea, in his case only ba hyme ies will be used as ield measu emen s, and hese ba hyme ies a e p o ided annually by he Po Se ice o he P incipali y As u ias. The ba hyme ies we e conduc ed wi h he use o a Na isound 210 single beam echo sounde , wi h a 1 cm e ical esolu ion and dual equency (Reson, Inc.: Slange up, Denma k). The speci ica ions o echo sounde Na isound 210 a e shown in Table 1. Table 1. Cha ac e is ics o echo sounde Na isound 210. Na isound 210/400 Speci ica ions F equency 190–235 kHz Po ency o ansmission 300 W Impedance 100 Ohm Echo app o al leng h 100 µs–210 kHz Dep h ange 0.5–100/400/1200 m depending on he equency Resolu ion 1 cm Accu acy 1 cm a 210 kHz (1 σ) assuming co ec sound eloci y, ansduce dep h e c. The s udy’s echo sounding da a o Lua ca we e de e mined, on 28 June 2016, 10 May 2018 and 28 May 2019. Fo he p esen wo k, XY posi ions om he su ey da a we e J. Ma . Sci. Eng. 2021,9, 267 5 o 18 p ojec ed using UTM Zone 30N. As hey a e usually cos ly, hey a e no mally used only o speci ic da es. The sa elli e Sen inel-2 p o ided he da a wi h which o p edic he ba hyme y o he wa e dep h a he po (Table 2). Table 2. Da es o acquisi ion o ba hyme ies. Yea 2017 2018 2019 2020 05/01/2019 25/01/2020 24/02/2018 24/02/2019 19/02/2020 16/03/2018 10/03/2020 10/04/2018 20/04/2019 30/05/2019 24/06/2018 14/06/2019 04/07/2017 29/07/2018 24/07/2019 13/08/2017 18/08/2018 23/08/2019 02/09/2017 02/09/2018 12/09/2019 02/10/2017 02/10/2018 22/10/2019 21/11/2017 16/11/2018 21/11/2019 21/12/2017 31/12/2018 3.2. Me hodology The p ocess o ob ain he ba hyme ies is shown schema ically in Figu e 2. J. Ma . Sci. Eng. 2021, 9, x FOR PEER REVIEW 5 o 18 p ojec ed using UTM Zone 30N. As hey a e usually cos ly, hey a e no mally used only o speci ic da es. The sa elli e Sen inel-2 p o ided he da a wi h which o p edic he ba hyme y o he wa e dep h a he po (Table 2). Table 2. Da es o acquisi ion o ba hyme ies. Yea 2017 2018 2019 2020 05/01/2019 25/01/2020 24/02/2018 24/02/2019 19/02/2020 16/03/2018 10/03/2020 10/04/2018 20/04/2019 30/05/2019 24/06/2018 14/06/2019 04/07/2017 29/07/2018 24/07/2019 13/08/2017 18/08/2018 23/08/2019 02/09/2017 02/09/2018 12/09/2019 02/10/2017 02/10/2018 22/10/2019 21/11/2017 16/11/2018 21/11/2019 21/12/2017 31/12/2018 3.2. Me hodology The p ocess o ob ain he ba hyme ies is shown schema ically in Figu e 2. Figu e 2. Me hodological wo k low o de elopmen o ba hyme ic maps. 3.2.1. P ocessing Sa elli e Images The p ocessing s ep conce ns using Sen inel-2A da a as ec o o a iables o aining and i ing o he andom o es model (Figu e 2). Sen inel-2A sa elli e da a was e ie ed om ESA Cope nicus Open Access Hub. SNAP (Sen inel Applica ion Pla o m) so wa e was used o isualize and p ep ocess Sen inel-2A da a (10 m esolu ion) [45]. The da a om he Sen inel-2A sa elli e e lec ance bands (B1, B2, B3, B4, B5, B6, B7, B8, Figu e 2. Me hodological wo k low o de elopmen o ba hyme ic maps. 3.2.1. P ocessing Sa elli e Images The p ocessing s ep conce ns using Sen inel-2A da a as ec o o a iables o aining and i ing o he andom o es model (Figu e 2). Sen inel-2A sa elli e da a was e ie ed om ESA Cope nicus Open Access Hub. SNAP (Sen inel Applica ion Pla o m) so wa e was used o isualize and p ep ocess Sen inel-2A da a (10 m esolu ion) [ 45 ]. The da a om he Sen inel-2A sa elli e e lec ance bands (B1, B2, B3, B4, B5, B6, B7, B8, B8A, B9, B11, and B12) we e used o p edic he dep h o he wa e a he s udy po . All spec al bands in he Sen inel-2A images we e esampled o achie e a esolu ion o 10 m using he SNAP S2 Resampling P ocesso [ 15 , 46 , 47 ]. As a esul , a da ase wi hou geo e e encing J. Ma . Sci. Eng. 2021,9, 267 6 o 18 was ob ained, and o de e mine he posi ioning o he e e ence poin s, he geog aphical loca ion o each poin was de ined by i s longi ude and la i ude using he SNAP p og am. Then, using he WGS84 ellipsoid, a coo dina e p ojec ion was c ea ed in o de o ob ain he coo dina es in ETRS89 [ 48 ]. This sys em was also used o p ojec he posi ions ha he echo sounde p o ides. The ellipsoid p ojec ions had an a e age posi ion e o o 1 cm. The da a ha was ob ained was compa ed o he ba hyme y ha had been p ojec ed. To accomplish his, a geodesic calcula o was used o p ojec he coo dina es. This enabled he au ho s o ob ain he da a o bands ha a e associa ed wi h UTM x-y coo dina es. The andom o es algo i hm was used o assign he z coo dina e. F om hese poin s and using he QGIS so wa e ( e sion 3.14), he su ace was ob ained by digi al models o he TIN (T iangula ed I egula Ne wo k) e ain ype. T iangula ion was unde aken using linea in e pola ion [ 49 ]. The e o ha his p ocess in ol ed was small due o he absence o any g ea i egula i ies in he smoo h su aces o he seabed. The po ’s ze o se es as he e e ence poin o he z coo dina es. The o me is he minimum measu ed le el o he su ace o he wa e du ing he highes low ide in he las 15 yea s. Pixels we e assigned dimensions on he basis o hei x-y loca ions. 3.2.2. The Random Fo es Algo i hm The andom o es algo i hm [ 50 ] acili a es classi ica ion and eg ession by use o a andomized subse o p edic o s ha enable i o c ea e a a ie y o classi ica ion ees [ 51 ]. Many o he ees s a as boo s apped aining da a samples. A andom subse o p e- dic o a iables (z coo dina e) is used a each o k in he p ocess. This causes each ee o be di e en . Al hough each ee is a poo p edic o , each pai o hem p o ides a di e - en esponse. This agg ega es he p edic ions o unco ela ed ees, educing p edic ion a iance and inc easing accu acy [ 52 – 54 ]. This wo k in ol ed 100 ees and 13 a iables (Sen inel-2A sa elli e bands B1, B2, B3, B4, B5, B6, B7, B8, B8A, B9, B11, and B12, and idal). The de aul alue was adop ed as he minimum size o nodes. The andom o es algo i hm o p edic ion o ba hyme y was p o ided by he Random Fo es ( 4.6-2) R package [ 55 ]. 3.2.3. T aining and Tes ing Da ase In o de o e alua e he accu acy o he model, he da ase ha was ob ained om Sen inel-2 was di ided in o wo g oups. T aining he andom o es algo i hm equi ed 80% o he da ase (1593 da a poin s), and es ing he model used he emaining 20% ( 388 da a poin s ). Da a om echo sounding measu emen s we e used o calib a e he model wi hou any kind o elabo a ion [ 56 ]. To es he model, he da a ob ained wi h he model we e compa ed wi h ield measu emen s o 20% o he poin s. As he ide a ec ed he dep h measu emen esul s, he measu ed dep hs we e he mean sea le els (MSL) ha se ed as e e ences. This was accomplished by deduc ing he measu ed dep h du ing high ide om he MSL. The idal da a we e p o ided by he nea es idal s a ion [34]. Finally, in o de o de e mine he accu acy o he model’s es ima es o dep h, sa elli e de i ed ba hyme y maps we e compa ed wi h ha o he ield measu emen s ob ained by using he echo sounde . The esidual s a is ic be ween he sa elli e de i ed dep h (Zsa elli e) and he echo sounding measu emen s (Zecho-sounde ) we e epo ed along h ee me ics. They we e he mean absolu e e o (MAE), he oo mean squa ed e o (RMSE) and he co ela ion coe icien (adjus ed R2). They a e calcula ed by he equa ions below. MAE =1 n n ∑ i=1 |ZSa elli e −Zecho−sounde | RMSE =s1 n n ∑ i=1 (ZSa elli e −Zecho−sounde )2 R2=∑n i=1Zecho−sounde −Zecho−sounde ZSa elli e −ZSa elli e q∑n i=1Zecho−sounde −Zecho−sounde 2∑n i=1Zecho−sounde −Zecho−sounde 2 J. Ma . Sci. Eng. 2021,9, 267 7 o 18 whe e Z Sa elli e a e he dep hs ha he andom o es me hodology p edic ed om Sen inel-2 images. Zecho-sounde deno e he in si u echo sounding dep hs and nis he numbe o da a. 4. Resul s The andom o es algo i hm achie ed a good p edic i e pe o mance, wi h an MAE o 0.37, an RMSE o 0.47 and an R2o 0.974. The es ing da a se esul s appea in Table 3. Table 3. E o s a is ics epo ed in me e s p oduced by he andom o es algo i hm. Algo i hm MAE (m) RMSE (m) R2 Random o es 0.37 0.47 0.974 Figu e 3p o ides a his og am o he ela ionship be ween mean absolu e e o (MAE) and he dep h o he po o Lua ca ha was ob ained by he andom o es algo i hm. As can be seen in Figu e 3, he maximum e o (60 cm) occu ed be ween − 6 and − 8 m and he minimum (24 cm) occu ed a − 10 m. This e o is accep able o he pu pose o his s udy, which is o analyze he beha io o he seabed wi h espec o ime. J. Ma . Sci. Eng. 2021, 9, x FOR PEER REVIEW 7 o 18 RMSE=1𝑛(𝑍−𝑍)   R=∑(𝑍−𝑍                  )(𝑍−𝑍            )   ∑( 𝑍−𝑍                  )∑( 𝑍−𝑍                  )   whe e ZSa elli e a e he dep hs ha he andom o es me hodology p edic ed om Sen inel- 2 images. Zecho-sounde deno e he in si u echo sounding dep hs and n is he numbe o da a. 4. Resul s The andom o es algo i hm achie ed a good p edic i e pe o mance, wi h an MAE o 0.37, an RMSE o 0.47 and an R² o 0.974. The es ing da a se esul s appea in Table 3. Table 3. E o s a is ics epo ed in me e s p oduced by he andom o es algo i hm. Algo i hm MAE (m) RMSE (m) R2 Random o es 0.37 0.47 0.974 Figu e 3 p o ides a his og am o he ela ionship be ween mean absolu e e o (MAE) and he dep h o he po o Lua ca ha was ob ained by he andom o es algo i hm. As can be seen in Figu e 3, he maximum e o (60 cm) occu ed be ween −6 and −8 m and he minimum (24 cm) occu ed a −10 m. This e o is accep able o he pu pose o his s udy, which is o analyze he beha io o he seabed wi h espec o ime. Figu e 3. Va ia ion o MAE e o s e sus dep h (m). As an example o he esul s ha we e ob ained, he ba hyme ies be o e and a e he 2017 d edging a e ep esen ed in 3D and 2D (Figu es 4 and 5). Figu e 4 ep esen s he 3D e olu ion o he seabed o 21/11/17 (Figu e 4a), 21/12/17 (Figu e 4b) and 24/02/18 (Figu e 4c). These show ha he ele a ion o he seabed was main ained in Figu e 4a,b, bu had declined in Figu e 4c due o d edging. Figu e 3. Va ia ion o MAE e o s e sus dep h (m). As an example o he esul s ha we e ob ained, he ba hyme ies be o e and a e he 2017 d edging a e ep esen ed in 3D and 2D (Figu es 4and 5). Figu e 4 ep esen s he 3D e olu ion o he seabed o 21/11/17 (Figu e 4a), 21/12/17 (Figu e 4b) and 24/02/18 (Figu e 4c). These show ha he ele a ion o he seabed was main ained in Figu e 4a,b, bu had declined in Figu e 4c due o d edging. Figu e 5 ep esen s one o he d edging episodes by a se ies o ba hyme ies ha we e analyzed in he plane and p o ile iew. These indica e he e ec o he d edging ha was unde aken. Fi s , when d edging is ca ied ou , he e is a dec ease in he ele a ion o he basin and he inle channel (Figu e 5a,c). The ele a ion eco e s when he d edging s ops (Figu e 5b,d). This e ec is seen in bo h he inne dock (sec ion AA’) and in he inle channel (sec ion BB’). In bo h ep esen a ions (Figu es 4and 5), he a eas o g ea es a ia ion a e he inne basin and he inle channel. These a e he a eas whe e d edging akes place. Addi ionally, in Figu es 4and 5, he a eas in which he g ea es changes occu a e in he lowe igh co ne (inne dock a ea and inle channel) and in he uppe le co ne ha coincides wi h he beach a ea. Bo h a e a eas ha unde go many changes in opog aphy due o coas al dynamics. J. Ma . Sci. Eng. 2021,9, 267 8 o 18 J. Ma . Sci. Eng. 2021, 9, x FOR PEER REVIEW 8 o 18 Figu e 4. 3D ba hyme ies. Da es: (a) 21/11/17, (b) 21/12/17 and (c) 24/02/18. Figu e 4. 3D ba hyme ies. Da es: (a) 21/11/17, (b) 21/12/17 and (c) 24/02/18. J. Ma . Sci. Eng. 2021,9, 267 9 o 18 J. Ma . Sci. Eng. 2021, 9, x FOR PEER REVIEW 9 o 18 Figu e 5. A ep esen a ion o he beha io o he in e io basin he d edging on: (a) 02/10/17, (b) 21/11/17, (c) 21/12/17 and (d) 24/02/18; and c oss-sec ional p o ile (e) AA’ o he inne dock, and ( ) BB’ o he inle channel. Figu e 5. A ep esen a ion o he beha io o he in e io basin he d edging on: ( a ) 02/10/17, ( b ) 21/11/17, ( c ) 21/12/17 and (d) 24/02/18; and c oss-sec ional p o ile (e) AA’ o he inne dock, and ( ) BB’ o he inle channel. J. Ma . Sci. Eng. 2021,9, 267 16 o 18 eco e ed almos immedia ely (in he d edging o 2018 in 5 days) wi h illing a es o 589.90 and 3556.35 m 3 /day, o 2017 and 2018, espec i ely. I also can be s a ed ha he spo adic a i al o sedimen abo e he mean dep h also disappea ed wi hou ex e nal in e en ion, keeping he mean dep h be ween − 3 and − 4 m. As a esul , i can be concluded ha lowe ing he ele a ion by d edging does no seem adequa e, unless he d edging dep h is close o he a e age ele a ion ha is main ained na u ally. Fo u u e wo k, i migh be use ul o ex end his in es iga ion o o he po s. Be o e de e mining he need o d edging, i would be use ul o know i he e is a dep h beyond which i is no app op ia e o conduc u he d edging due o he a e a which sedimen e u ns. The analysis o he beha io o he bo om o he po s in ecen yea s p o ides aluable in o ma ion o he planning o main enance asks and he knowledge o hei beha io in he ace o li o al dynamics. Au ho Con ibu ions: Concep ualiza ion, V.M.-P. and F.O.-F.; so wa e and alida ion, V.M.-P.; me hodology, M.C.-B.; da a cu a ion, V.M.-P.; w i ing—o iginal d a p epa a ion, M.C.-B.; w i ing— e iew and edi ing, F.O.-F., and V.R.-M. All au ho s ha e ead and ag eed o he published e sion o he manusc ip . Funding: This wo k was unded by he Science, Technology and Inno a ion Plan o he P incipali y o As u ias (Spain) Re : FC-GRUPIN-IDI/2018/000225, which is pa ly unded by he Eu opean Regional De elopmen Fund (ERDF). Ins i u ional Re iew Boa d S a emen : No applicable. In o med Consen S a emen : No applicable. Acknowledgmen s: The au ho s would like o hank he Po Se ice and T anspo In as uc u es o he P incipali y o As u ias o hei collabo a ion in his wo k. Con lic s o In e es : The au ho s decla e he e a e no con lic o in e es . Re e ences 1. Quang T i, D.; Kandasamy, J.; Cao Don, N. Quan i a i e Assessmen o he En i onmen al Impac s o D edging and Dumping Ac i i ies a Sea. Appl. Sci. 2019,9, 1703. [C ossRe ] 2. Bolam, S.G.; Rees, H.L. Minimizing Impac s o Main enance D edged Ma e ial Disposal in he Coas al En i onmen : A Habi a App oach. En i on. Manag. 2003,32, 171–188. [C ossRe ] 3. No én, A.; Fedje, K.K.; S öm all, A.-M.; Rauch, S.; Ande sson-Sköld, Y. In eg a ed Assessmen o Managemen S a egies o Me al-Con amina ed D edged Sedimen s—Wha A e he Bes App oaches o Po s, Ma inas and Wa e ways? Sci. To al En i on. 2020,716, 135510. [C ossRe ] 4. Wang, W.; Men, C.; Lu, W. Online P edic ion Model Based on Suppo Vec o Machine. Neu ocompu ing 2008 ,71, 550–558. [C ossRe ] 5. Kho am, S.; Khalegh, M.A. A No el Hyb id MCDM App oach o E alua e Po s’ D edging P ojec C i e ia Based on In ui ionis ic Fuzzy DEMATEL and GOWPA. WMU J. Ma i . A . 2020,19, 95–124. [C ossRe ] 6. Cáce es, R.A.; Zyse man, J.A.; Pe illo, G.M.E. Analysis o Sedimen a ion P oblems a he En ance o Ma Del Pla a Ha bo . J. Coas . Res. 2016,32, 301–314. [C ossRe ] 7. Feola, A.; Lisi, I.; Salme i, A.; Ven i, F.; Ped oncini, A.; Gabellini, M.; Romano, E. Pla o m o In eg a ed Tools o Suppo En i onmen al S udies and Managemen o D edging Ac i i ies. J. En i on. Manag. 2016,166, 357–373. [C ossRe ] 8. Mahmoodi, A.; Lash eh Neshaei, M.A.; Mansou i, A.; Sha ai Bejes an, M. S udy o Cu en - and Wa e-Induced Sedimen T anspo in he Nowshah Po En ance Channel by Using Nume ical Modeling and Field Measu emen s. J. Ma . Sci. Eng. 2020 , 8, 284. [C ossRe ] 9. Chen, B.; Wang, K. Suspended Sedimen T anspo in he O sho e nea Yang ze Es ua y* *P ojec Suppo ed by he Na ional Na u al Science Founda ion o China (G an No.40576017), he Na ional Basic Resea ch P og am o China (973, P og am, G an No. 2007CB411804). J. Hyd odyn. Se . B 2008,20, 373–381. [C ossRe ] 10. Zuo, S.; Xie, H.; Ying, X.; Cui, C.; Huang, Y.; Li, H.; Xie, M. Seabed Deposi ion and E osion Change and In luence Fac o s in he Yangshan Deepwa e Po o e he Yea s. Ac a Oceanol. Sin. 2019,38, 96–106. [C ossRe ] 11. E emeije , P.L.A.; Robin Lewis, R.R. En i onmen al Impac s o D edging on Seag asses: A Re iew. Ma . Pollu . Bull. 2006 ,52, 1553–1572. [C ossRe ] 12. Flo , G.; del Bus o, J.A.; Blanco, G.F. Mo phological and Sedimen a y Pa e ns o Po s o he As u ian Region (NW Spain). J. Coas . Res. 2006,48, 35–40. J. Ma . Sci. Eng. 2021,9, 267 17 o 18 13. S ock, F.; Knipping, M.; Pin , A.; Lads ä e , S.; Delile, H.; Heiss, A.G.; Lae manns, H.; Mi chell, P.D.; Ploye , R.; S eskal, M.; e al. Human Impac on Holocene Sedimen Dynamics in he Eas e n Medi e anean— he Example o he Roman Ha bou o Ephesus. Ea h Su . P ocess. Land . 2016,41, 980–996. [C ossRe ] 14. Sha aan, M.; Ib ahim, M.G.; Iskande , M.; Mas ia, A.; Nadaoka, K. Analysis o Sedimen a ion a he Fishing Ha bo En ance: Case S udy o El-Bu ullus, Egyp . J. Coas . Conse . 2018,22, 1143–1156. [C ossRe ] 15. Pou sanidis, D.; T aganos, D.; Reina z, P.; Ch ysoulakis, N. On he Use o Sen inel-2 o Coas al Habi a Mapping and Sa elli e- De i ed Ba hyme y Es ima ion Using Downscaled Coas al Ae osol Band. In . J. Appl. Ea h Obs. Geoin . 2019 ,80, 58–70. [C ossRe ] 16. Vi o i, G.; Blondeaux, P.; Mazzuoli, M.; Simeono , J.; Calan oni, J. Sedimen T anspo unde Oscilla o y Flows. In . J. Mul iph. Flow 2020,133, 103454. [C ossRe ] 17. Finn, J.R.; Li, M.; Ap e, S.V. Pa icle Based Modelling and Simula ion o Na u al Sand Dynamics in he Wa e Bo om Bounda y Laye . J. Fluid Mech. 2016,796, 340–385. [C ossRe ] 18. Finn, J.R.; Li, M. Regimes o Sedimen -Tu bulence In e ac ion and Guidelines o Simula ing he Mul iphase Bo om Bounda y Laye . In . J. Mul iph. Flow 2016,85, 278–283. [C ossRe ] 19. Kidanema iam, A.G.; Uhlmann, M. Di ec Nume ical Simula ion o Pa e n Fo ma ion in Subaqueous Sedimen . J. Fluid Mech. 2014,750, 1–13. [C ossRe ] 20. Leon ’ye , I.O.; Aki is, T.M. Modeling o Coas al Dynamics o he Anapa Bay-Ba . Oceanology 2020,60, 279–285. [C ossRe ] 21. A manini, A.; Ca edon, V.; Righe i, M. A P obabilis ic/De e minis ic App oach o he P edic ion o he Sedimen T anspo Ra e. Ad . Wa e Resou . 2015,81, 10–18. [C ossRe ] 22. Flene , C.; Lo sa i, E.; Alho, P.; Käyhkö, J. Compa ison o Empi ical and Theo e ical Remo e Sensing Based Ba hyme y Models in Ri e En i onmen s. Ri e Res. Appl. 2012,28, 118–133. [C ossRe ] 23. Gia dino, C.; B esciani, M.; Ma a, E.; B ando, V.E. Imaging Spec ome y o Inland Wa e Quali y in I aly Using MIVIS: An O e iew . In Ad ances in Wa e shed Science and Assessmen ; Younos, T., Pa ece, T.E., Eds.; Sp inge In e na ional Publishing: Cham, Swi ze land, 2015; Volume 33, pp. 61–83. ISBN 978-3-319-14211-1. 24. Jawak, S.D.; Vadlamani, S.S.; Luis, A.J. A Synop ic Re iew on De i ing Ba hyme y In o ma ion Using Remo e Sensing Technolo- gies: Models, Me hods and Compa isons. Ad . Remo e Sens. 2015,4, 147–162. [C ossRe ] 25. Hedley, J.; Roel sema, C.; Cholle , I.; Ha bo ne, A.; He on, S.; Weeks, S.; Ski ing, W.; S ong, A.; Eakin, C.; Ch is ensen, T.; e al. Remo e Sensing o Co al Ree s o Moni o ing and Managemen : A Re iew. Remo e Sens. 2016,8, 118. [C ossRe ] 26. Lyzenga, D.R. Passi e Remo e Sensing Techniques o Mapping Wa e Dep h and Bo om Fea u es. Appl. Op . 1978 ,17, 379. [C ossRe ] 27. Cheng, N.-S.; Chiew, Y.-M. Pickup P obabili y o Sedimen En ainmen . J. Hyd aul. Eng. 1998,124, 232–235. [C ossRe ] 28. Liu, S.; Gao, Y.; Zheng, W.; Li, X. Pe o mance o Two Neu al Ne wo k Models in Ba hyme y. Remo e Sens. Le . 2015 ,6, 321–330. [C ossRe ] 29. El-Mewa i, M.; Salah, M.; Fawzi, B. Assessmen o Op ical Sa elli e Images o Ba hyme y Es ima ion in Shallow A eas Using A i icial Neu al Ne wo k Model. Am. J. Geog . In . Sys . 2018,7, 99–106. 30. Obelcz, J.; Wood, W.T.; Ph ampus, B.J.; Lee, T.R. Machine Lea ning Augmen ed Time-Lapse Ba hyme ic Su eys: A Case S udy om he Mississippi Ri e Del a F on . Geophys. Res. Le . 2020,47, e2020GL087857. [C ossRe ] 31. Tonion, F.; Pi o i, F.; Faina, G.; Pal inie i, D. A Machine Lea ning App oach o Mul ispec al Sa elli e De i ed Ba hyme y. ISPRS Ann. Pho og amm. Remo e Sens. Spa . In . Sci. 2020,3, 565–570. [C ossRe ] 32. Ma aeidopoulos, A.K.; Oikonomou, E.; Palika is, A.; Poulos, S. A Hyb id Bio-Op ical T ans o ma ion o Sa elli e Ba hyme y Modeling Using Sen inel-2 Image y. Remo e Sens. 2019,11, 2746. [C ossRe ] 33. Sagawa, T.; Yamashi a, Y.; Okumu a, T.; Yamanokuchi, T. Sa elli e De i ed Ba hyme y Using Machine Lea ning and Mul i- Tempo al Sa elli e Images. Remo e Sens. 2019,11, 1155. [C ossRe ] 34. Manessa, M.D.M.; Kanno, A.; Sekine, M.; Haida , M.; Yamamo o, K.; Imai, T.; Higuchi, T. Sa elli e-De i ed Ba hyme y Using Random Fo es Algo i hm and Wo ld iew-2 Image y. Geoplanning J. Geoma . Plan. 2016,3, 117–126. [C ossRe ] 35. Kogu , T.; Weis ock, M. Classi ying Ai bo ne Ba hyme y Da a Using he Random Fo es Algo i hm. Remo e Sens. Le . 2019,10, 874–882. [C ossRe ] 36. Yunus, A.P.; Dou, J.; Song, X.; A a , R. Imp o ed Ba hyme ic Mapping o Coas al and Lake En i onmen s Using Sen inel-2 and Landsa -8 Images. Senso s 2019,19, 2788. [C ossRe ] [PubMed] 37. Bu es, L.; Sycho a, P.; Maca, P.; Roub, R.; Ma al, S. Ri e Ba hyme y Model Based on Floodplain Topog aphy. Wa e 2019 ,11, 1287. [C ossRe ] 38. Se iawan, K.T.; Suwa gana, N.; Gin ing, D.N.B.; Manessa, M.D.M.; Angg aini, N.; Adawiah, S.W.; Julza ika, A.; Su ahman, S.; Rosid, S.; Supa djo, A.H. Ba hyme y ex ac ion om spo 7 sa elli e image y using andom o es me hods. In . J. Remo e Sens. Ea h Sci. IJReSES 2019,16, 23–30. [C ossRe ] 39. Moeinkhah, A.; Shakiba, A.; Aza akhsh, Z. Assessmen o Reg ession and Classi ica ion Me hods Using Remo e Sensing Technology o De ec ion o Coas al Dep h (Case S udy o Busheh Po and Kha g Island). J. Indian Soc. Remo e Sens. 2019 ,47, 1019–1029. [C ossRe ] J. Ma . Sci. Eng. 2021,9, 267 18 o 18 40. Ha, N.T.; Manley-Ha is, M.; Pham, T.D.; Hawes, I. A Compa a i e Assessmen o Ensemble-Based Machine Lea ning and Maximum Likelihood Me hods o Mapping Seag ass Using Sen inel-2 Image y in Tau anga Ha bo , New Zealand. Remo e Sens. 2020,12, 355. [C ossRe ] 41. Mis a, A.; Ramak ishnan, B. Assessmen o Coas al Geomo phological Changes Using Mul i-Tempo al Sa elli e-De i ed Ba hyme y. Con . Shel Res. 2020,207, 104213. [C ossRe ] 42. Hube , M.E.; Zigic, S.; Gilbe , R.; Smi h, D.; Edison, K.; Goudkamp, K.; Lang y, S.; Bu ling, M. Imp o ed D edge Ma e ial Managemen o he G ea Ba ie Ree Region. In P oceedings o he Aus alasian Po and Ha bou Con e ence; Enginee s Aus alia: Ba on, Aus alia, 2013; p. 400. 43. La a, J.L.; Lucio, D.; Tomas, A.; Di Paolo, B.; Losada, I.J. High-Resolu ion Time-Dependen P obabilis ic Assessmen o he Hyd aulic Pe o mance o His o ic Coas al S uc u es: Applica ion o Lua ca B eakwa e . Philos. T ans. R. Soc. Ma h. Phys. Eng. Sci. 2019,377, 20190016. [C ossRe ] 44. Nalona (Suc ion D edge ) Regis e ed in—Vessel De ails, Cu en Posi ion and Voyage In o ma ion—IMO 9047453|AIS Ma ine T a ic. A ailable online: h ps://www.ma ine a ic.com (accessed on 18 Sep embe 2019). 45. Sen inel Applica ion Pla o m. ESA Toolboxes, 2009. SNAP. A ailable online: h p://s ep.esa.in /main/ oolboxes/snap (accessed on 12 Janua y 2020). 46. Lana as, C.; Bioucas-Dias, J.; Galliani, S.; Bal sa ias, E.; Schindle , K. Supe -Resolu ion o Sen inel-2 Images: Lea ning a Globally Applicable Deep Neu al Ne wo k. ISPRS J. Pho og amm. Remo e Sens. 2018,146, 305–319. [C ossRe ] 47. Rumo a, L.; Mile , M.; Medak, D. Impac o Va ious A mosphe ic Co ec ions on Sen inel-2 Land Co e Classi ica ion Accu acy Using Machine Lea ning Classi ie s. ISPRS In . J. Geo-In . 2020,9, 277. [C ossRe ] 48. Son, N.-T.; Chen, C.-F.; Chen, C.-R.; Guo, H.-Y. Classi ica ion o Mul i empo al Sen inel-2 Da a o Field-Le el Moni o ing o Rice C opping P ac ices in Taiwan. Ad . Space Res. 2020,65, 1910–1921. [C ossRe ] 49. Alwhaely, U.; Hussein, M.A.; AL-Kaaby, L.F. Using GIS and Remo e Sensing Sa elli e Da a o Mapping and Moni o ing Sha Al-A ab Es ua y (ou Ba A ea) and Nea by Coas line Sou he n I aq. Al-Qadisiyah J. Pu e Sci. 2020,25, 1–21. [C ossRe ] 50. B eiman, L. Random Fo es s. Mach. Lea n. 2001,45, 5–32. [C ossRe ] 51. Reiss, H.; Cunze, S.; König, K.; Neumann, H.; K öncke, I. Species Dis ibu ion Modelling o Ma ine Ben hos: A No h Sea Case S udy. Ma . Ecol. P og. Se . 2011,442, 71–86. [C ossRe ] 52. James, G.; Wi en, D.; Has ie, T.; Tibshi ani, R. An In oduc ion o S a is ical Lea ning; Sp inge : Cham, Swi ze land, 2013; Volume 112. 53. P asad, A.M.; I e son, L.R.; Liaw, A. Newe Classi ica ion and Reg ession T ee Techniques: Bagging and Random Fo es s o Ecological P edic ion. Ecosys ems 2006,9, 181–199. [C ossRe ] 54. Pe e s, J.; Bae s, B.D.; Ve hoes , N.E.C.; Samson, R.; Deg oe e, S.; Becke , P.D.; Huyb ech s, W. Random Fo es s as a Tool o Ecohyd ological Dis ibu ion Modelling. Ecol. Model. 2007,207, 304–318. [C ossRe ] 55. Kuhn, M. Ca e : Classi ica ion and Reg ession T aining. R Package Ve sion. A ailable online: h ps://www.R-p ojec .o g (accessed on 5 Feb ua y 2021). 56. Ma eo-Pé ez, V.; Co al-Bobadilla, M.; O ega-Fe nández, F.; Ve ga a-González, E.P. Po Ba hyme y Mapping Using Suppo Vec o Machine Technique and Sen inel-2 Sa elli e Image y. Remo e Sens. 2020,12, 2069. [C ossRe ] 57. Con ede ación Hid og á ica del Can áb ico. A ailable online: h ps://www.chcan ab ico.es/las-cuencas-can ab icas/ma co- isico/hid ologia/ ios/neg o (accessed on 12 No embe 2019).