Drainage of deep Mediterranean Water, its transition through the Strait of Gibraltar: spatial and temporal variability
Abstract
La presente tesis está organizada en cuatro capítulos en los que se estudia el flujo de agua saliente del Mar Mediterráneo a través del Estrecho de Gibraltar, para ello se han abordado varios procesos relacionados con el intercambio de aguas Atlánticas y Mediterráneas a través del Estrecho. El primer objetivo es identificar los mecanismos que influyen en la ventilación de las aguas profundas Mediterráneas residentes en el Mar de Alborán. Para ello en primer lugar se investiga el papel que juegan las estructuras de escala regional en la ventilación de aguas profundas mediante datos medidos in-situ, datos de satélite y simulaciones numéricas realizadas a tal fin. De éste estudio se deduce que la dinámica del Estrecho de Gibraltar es fundamental en la ventilación de las aguas profundas del Mar de Alborán, la cual a su vez es facilitada por la presencia casi permanente del giro anticiclónico del Mar de Alborán Occidental. Seguidamente, mediante modelos numéricos regionales y a nivel de cuenca se estudia el modo en que el forzamiento mareal interfiere no solo en la circulación en el propio Estrecho sino también en la ventilación de aguas profundas desde el mar de Alborán o en los procesos de formación de estas mismas aguas. Se concluye que la dinámica mareal debe ser resulta en los modelos a escala regional y de cuenca que pretenden reproducir algunas características relevantes del intercambio a largo término y, muy probablemente, otros importantes procesos que tienen lugar más allá del Estrecho, como la formación de agua profunda en el Golfo de León. A lo largo del tercer capítulo se ha estudiado en qué medida la marea en el Estrecho de Gibraltar afecta a los procesos biogeoquímicos de la vecina cuenca del Mar de Alborán, inusualmente productiva en comparación con el resto del Mar Mediterráneo. El resultado es una productividad primaria en el Mar de Alborán que se reproduce satisfactoriamente cuando el forzamiento mareal se incluye en un modelo regional físico‐ecológico. Más específicamente se encuentra que la diferencia entre tener o no en cuenta las mareas puede suponer hasta un 60% más de producción primaria en la cuenca Oeste de Alborán. Finalmente se realiza un intensivo análisis de datos de temperatura potencial y salinidad medidos en transeptos meridionales a lo largo del Estrecho. De éste modo aplicando un análisis de clusters se determinar cuál es el patrón espacio-temporal de los hasta seis tipos de agua que intervienen en el intercambio de flujos en el Estrecho. En particular se han detectado cuatro masas de agua Mediterráneas en el flujo saliente en la sección del contorno Este del Estrecho, a pesar de que una de ellas, puede no estar presente en determinados años. La señal de estas aguas se suaviza lentamente a medida que fluyen hacia el Oeste, y lo hacen abruptamente una vez sobrepasan el umbral de Camarinal. Según lo anterior, al Oeste del umbral sería más apropiado hablar de una única agua Mediterránea. Con respecto a la variabilidad temporal se observa que las señales estacionales son más evidentes en las aguas Atlánticas, mientras que la variabilidad interanual es más notable en las aguas Mediterráneas. Estas cuestiones ofrecen una nueva visión sobre la dinámica y las propiedades del flujo Mediterráneo de agua que abandona éste mar fluyendo a través del Estrecho.
Full text
(Drenaje de las aguas profundas del mar mediterráneo, estudio de su tránsito a través del Estrecho de Gibraltar y su variabilidad espacio-temporal) Tesis Doctoral presentada por: Cristina Naranjo Rosa Dirigida por: Jesús García Lafuente Y codirigida por: José Carlos Sánchez Garrido Para obtener el grado de Doctora por la Universidad de Málaga Departamento de Física Aplicada II Grupo de Oceanografía Física Escuela de doctorado: Escuela Internacional de Doctorado en Estudios del Mar (EIDEMAR) Programa de doctorado: Dinámica de los flujos biogeoquímicos y sus aplicaciones El Director Jesús García Lafuente El Codirector José Carlos Sánchez Garrido La Doctoranda Cristina Naranjo Rosa Drainage of deep Mediterranean Water, its transition through the Strait of Gibraltar: spatial and temporal variability
AUTOR: Cristina Naranjo Rosa http://orcid.org/0000-0001-6412-2443 EDITA: Publicaciones y Divulgación Científica. Universidad de Málaga Esta obra está sujeta a una licencia Creative Commons: Reconocimiento - No comercial - SinObraDerivada (cc-by-nc-nd): Http://creativecommons.org/licences/by-nc-nd/3.0/es Cualquier parte de esta obra se puede reproducir sin autorización pero con el reconocimiento y atribución de los autores. No se puede hacer uso comercial de la obra y no se puede alterar, transformar o hacer obras derivadas. Esta Tesis Doctoral está depositada en el Repositorio Institucional de la Universidad de Málaga (RIUMA): riuma.uma.es
D.JesúsGarcíaLafuente,catedráticodelaUniversidaddeMálagayD.José CarlosSánchezGarrido,investigadordelDepartamentodeFísicaAplicadaII delaUniversidaddeMálaga HACENCONSTAR: Queeltrabajorecogidoenlamemoriatitulada“Drainageofdeep MediterraneanWater,itstransitionthroughtheStraitofGibraltar:spatial andtemporalvariability”,presentadaporCristinaNaranjoRosa,hasido realizadobajonuestrasupervisiónycumpleconlacalidadcientífica suficiente. Quelapresentetesisestáavaladaporcuatroartículos,dosdeellosya publicados,unterceroenviadotrasunasegundarevisiónmenoryunúltimo en(moderada‐menor)revisión.Todosellosenrevistasdealtoíndicede impactoensuáreaquenohansidopresentadosanteriormenteenninguna otratesis. Quereúnesobradamenteloscriteriosdecalidadestablecidosparalalectura delatesisenelprogramadedoctorado“Dinámicadelosflujos biogeoquímicosysusaplicaciones”,porelloseinformamuyfavorablemente sobreeltrabajorealizadoyseautorizasupresentaciónydefensaparaoptar algradodeDoctorInternacionalporlaUniversidaddeMálaga. ParaellofirmamoslapresenteenMálagaa07 de Septiembrede2015. Fdo.DirectorFdo.Codirector JesúsGarcíaLafuenteJoséCarlosSánchezGarrido
A mis padres, mis hermanos y nuestros dos soles, Adri y Guille
TABLEOFCONTENTS I.PUBLICATIONSARISINGFROMTHISTHESIS.............................................................3 II. GENERALINTRODUCTION.........................................................................................5 II.a.Abstract..............................................................................................................5 II.b.Stateoftheart...................................................................................................5 II.c.Theobjectivesofthisdissertationanditsorganization..................................10 II.d.Methodology....................................................................................................11 III. TheWesternAlboranGyrehelpsventilatetheWesternMediterraneanDeep WaterthroughGibraltar.................................................................................................13 IV. HowmuchdotidesaffectthecirculationoftheMediterraneanSea?Fromlocal processesintheStraitofGibraltartobasin‐scaleeffects..............................................21 V.ModelingtheimpactoftidalflowsonthebiologicalproductivityoftheAlboran Sea.................................................................................................................................31 VI. MediterraneanwatersalongandacrosstheStraitofGibraltar,characterization andzonalmodification...................................................................................................77 VII. GENERALDISCUSSION.......................................................................................109 VIII MAIN CONCLUSIONS......................................................................................... 114 IX.RESUMENENESPAÑOL.....................................................................................115 IX.a. BreveResumen............................................................................................115 IX.b. EstadodelacuestiónyMetodología...........................................................116 IX.c. Resultadosmássignificativos.......................................................................121 IX.d. Principalesconclusionesdeestamemoria..................................................127 X. List of Acronyms.....................................................................................................128 XI.References.............................................................................................................130
wit of S th e Mi l th e co o co n WI W by a th e M o ou t tha th e an a Co n Ho w wa t thro u mas s mak e hLIWco m S icily.Acc o e outflow i l lo t [2009] e outflow, o lingof M n tinentals h W a minimu m e otherthr e o rerecent t flowform a tweredi s e irpathto a lysisof t n ductivity‐ w ever,th e t ersbeyo n Figure2.Sch u ghtheeaste r s identifiedby e thisfigure. m ingfrom t o rdingto M i n2003a n suggeste d theWint e M odified A h elfofth e isidenti m ofsalinit e eMedite lyMillot edbydiff e s tinguisha b wardthe t heslope Temperat u e undisp u n dCSisac o emeofthez o r nboundaryof itsacronymin t heEaster n M illot’sre s n d2004. F d theparti c e rInterm e A tlantic W e northwe s fiableinth yandocc u rranean W [2014]re v e rentprop o b leeven a A tlantic. T ofthe m u re‐Depth u tableide n o ntroversi o nalandverti c thelegendat n Mediter r s ults,this T F ollowing t c ipationo f e diateWa t W ater(M A sternregi o epotential u piesalay e W aters(W M v isitedth e ortionsof a ftercros s T heautho r m ixinglin e (CTD)pr o n tification alresult[ G c aldistributio n theStraitofG theright.Encl r aneanaft T DWwou t hesame t f afourth t er(WIW) , A W)durin o nofthe W densityinterval2 e rlocated M DW,LIW e issuest fourMedi s ingthe m r supporte e sdefine d o filesoft h ofthe d G arcíaLaf u n ofMediterra ibraltar(named osemapsho w II. Gen e ldhaveb t hread,in Mediterra , thatisf o gcoldw W estern M 8<σ θ <29in betweent andTDW, ressingth t erranean m ainsillo f d thiscon d inaθ‐ S h ecastsb e ifferent u enteetal. TACinFigure sthelocation eralIntrodu c eeroverflwingtheS t eenpresent alaterp a neanwat e o rmedby w intersin M editerran e aθ‐Sdiag t heNACW seeFigur e h eideao f waterma s f Camarin a clusionin S diagram e inganaly z Mediterra n ,2011]. nean and Atla ntic waters fl o 1). Eachcolor indicatesa w ofthecastsu s c tion 8 t rait in a per e rin the the e an. ram and e 2). f an s ses a lin the by z ed. n ean o wing w ater s edto
II. GeneralIntroduction 9 TheinflowofAtlanticwaterintotheMediterraneanSea,usuallyreferred astheAtlanticJet,determinesthesurfacecirculationoftheAlboranSea,the westernmostsub‐basinoftheWesternMediterranean.Thetypicalpatternof thesurfacecirculationinthissub‐basinistheco‐existenceoftwoanticyclonic gyres,thealreadymentionedWAGandtheEasternAlborangyre(EAG)[Seco, 1959;Donguy,1962;Lanoix,1974].TheAtlanticJetfollowsameandering patharoundbothgyresandoriginatesintensedensityfrontsrepeatedly studiedintheliterature[Tintoreetal.,1988;Tintoréetal.,1991;Viúdezetal., 1996]whenitmeetsthedenserandsaltierMAWresidinginthe MediterraneanSea.Associatedwiththefrontalactivityofthesestructures, satellitedatashowhighchlorophyllconcentrationsalongtheJet[Navarroet al.,2011],suggestinganenhancedproductivitysustainedbytheageostrophic circulationinthefront[Oguzetal.,2014].Thispatternismorenoticeablein thewesternAlboranSea,wherewesterlywindscanfurthercauseupwelling alongtheSpanishshoreandgiverisetophytoplanktonblooms[Sarhanetal., 2000;Ramírezetal.,2005;Reuletal.,2005;Macíasetal.,2007;2008].These processesareessentialtoexplainwhytheAlboranSeaisthemostproductive sub‐basinoftheMediterraneanSea[Uitzetal.,2012].Theroleplayedbythe frontaldynamicsinthebiologicalproductivityoftheAlboranSeawasrecently addressedbyOguzetal.[2014]usinga3Dnumericalmodel.Theirmodeldid notincludetides,whichisaconsiderabledrawbacksincetidalcurrentscan fertilizethesurfaceinflowingwatersthroughtheenhancementofthevertical mixingatthewesternsideofCS[GarcíaLafuenteetal.,2013].Thestrong interactionbetweentidalcurrentsandtheabruptbottomtopographyofthe StraitofGibraltaroriginateshydraulictransitions[SánchezGarridoetal., 2011]thatfavorsverticalmotions,particularlyintheTangierbasin downstreamofCS(Figure1),andthesubsequentmixingoftherelative impoverishedAtlanticwaterwiththenutrientrichdeepMediterranean watersthathadbeenpreviouslyupliftedfromtheAlboranbasin.Agooddeal ofthemixedwaterreturnstowardtheMediterraneanwhenthetidereverses andcouldfueltheproductioninthephoticlayer,helpingtomaintainthehigh ratesofprimaryproductivityobservedinthenorthernarea.However,the topichasnotbeenaddressedyet.
II. GeneralIntroduction 10 II.c. The objectives of this dissertation and its organization Theprevioussectionhasraisedseveralimportantandrelevantissues concerningdifferentprocessesoftheStraitofGibraltaranditsneighboring basins,especiallytheAlboranbasinintheMediterraneanSea.Theseissues arethebodyofthisdissertation,whichaimsatprovidingresponsesto questionslikewhetherornottheWAGisinvolvedinthefinaldrainageofthe WMDWand,ifyes,whichisitsroleintheprocess(ChapterIII2),orwhichis theroleofthetidesintheStraitofGibraltarintheaspirationofthiswater (ChapterIV3).Otherappealingissuesaretheremoteeffectofthetidesinthe StraitonlargescaleprocessestakingplaceinsidetheMediterraneanSea, suchasthewinterdeepconvectionintheGulfofLion(alsoaddressedin ChapterIV)orinthefertilizationoftheAlboranbasin(ChapterV4).Thereis alsotheongoingdebateaboutthecompositionoftheoutflowingwaters throughtheStraitofGibraltar.Basedonacomprehensivedatasetacquiredin thisarearecently,ChapterVI 5describesnewoutcomesaboutthis composition.Thedissertationfinisheswithashortsummaryofthemain findingsreportedinthementionedChapters. Allthisresearchhasbeencarriedoutintheframeofthesequenceof INGRESprojectsfundedbytheSpanishMinistryofEconomíayCompetitividad (projectsREN2003_01608,CTM2006_02326andCTM2010_21229),although mostoftheefforthasbeendonewithinINGRES‐3project(CTM2010_21229), thelastalsograntedtheauthorofthepresentdissertationwitha“Formación dePersonalInvestigador”(BES‐2011‐043421)fellowship,bothofwhichare speciallyacknowledged. 2ThisChaptercorrespondstopaper“TheWesternAlboranGyrehelpsventilatetheWestern MediterraneanDeepWaterthroughGibraltar”(seesectionI). 3ThisChaptercorrespondstopaper“HowmuchdotidesaffectthecirculationoftheMediterranean Sea?FromlocalprocessesintheStraitofGibraltartobasin‐scaleeffects”(seesectionI). 4ThisChaptercorrespondstopaper“Modelingtheimpactoftidalflowsonthebiologicalproductivity oftheAlboranSea”(seesectionI). 5ThisChaptercorrespondstopaper“MediterraneanwatersalongandacrosstheStraitof Gibraltar,characterizationandzonalmodification”(seesectionI).
II. GeneralIntroduction 11 II.d. Methodology Toaddressthefirstoftheabovementionedtopics,ChapterIIIexamines thepresenceofWMDWintheoutflowbyanalyzingafive‐yearlongtime seriesofpotentialtemperaturecollectedatES.Thesensorwasplacednear thebottomtoensurethatitwasmeasuringthedensestwaterleavingthe Strait,whichistheWMDW.AltimetrydataoftheAlboranSeawereutilizedto detectsurfacestructures,suchdeanticyclonicgyres,andfollowtheir evolution,whichallowedustoinvestigatetherelationbetweentheWAGand thedrainageofWMDWatESasindicatedbytheθtimeseries.HistoricalCTD datafromMEDATLASdatabase[MEDAR‐Group,2002]wereusedtoestimate thedepthfromwhichdeepwatersintheAlboranSeacouldbesuctioned. Finally,anumericalmodelwasrunundertwodifferentconfigurations:inone ofthemtheStraitofGibraltarwasclosedandtheoutputsofthemodelwere comparedwitharuninwhichtheStraitwaspresent.Thecomparisonclarifies therolethatthedynamicswithintheStraitplaysintheaspirationofWMDW. ChapterIVinvestigatestheeffectsthatthetidalforcingintheStraitof GibraltarcouldhavenotjustinthenearAlboranbasinbutalsofarthereast,in theinterioroftheMediterraneanSea.Twoprocess‐orientedmodel experimentswereconductedtothisaim:thefirstonewasaregional(Straitof Gibraltarandneighborbasins)andthesecondonewasbasin‐wide (MediterraneamSea).Bothmodelswereruntwice,withandwithouttidal forcing.Theregionalmodelrevisitedtheeffectoftidesontheexchanged flowsthroughtheStraitandthemodificationofthehydrologicalpropertiesof theAltanticinflowbycontrastingtheoutputsfromtherunswithandwithout tides.ThelastissueisrelevantforthedeepformationprocessesintheGulfof LionssincethemixinginducedbytidesintheStraitmakesafirst preconditioningoftheAtlanticwaterthatwillbetransformedindifferent Mediterraneanwaters.Thisimportantissuewasfurtheraddressedwitha basinscalemodelwithacomparableconfigurationintheStraitregion.The comparisonbetweenbothrunshighlightstheroleofthetidesinthedeep convectionprocessesformingWMDWintheGulfofLionandalsointhe ventilationofthiswaterintheAlboranSea. Ascommentedabove,theAlboranSeadisplayshighratesofprimary productivity,especiallyifcomparedwiththerestoftheMediterraneanSea. VerticalmixinginducedbytidesintheStraitofGibraltarhasbeenproposed [GarcíaLafuenteetal.,2013]asasourceofnutrientstothisbasin.Theissue
II. GeneralIntroduction 12 isfurtherdiscussedinChapterVbytheimplementationofahighresolution circulationmodelcoupledtoanecosystemmodel.Thesameprocedure followedinChapterIVisrepeatedhere:areferencenumericalexperimentin whichtidalforcingissuppressed(butyetreproducingthebasiccirculationin theareaoftheStraitofGibraltarandAlboranSea)iscomparedwithasimilar runincludingtidalforcinginordertoevaluatethebiologicalconsequences thatthetidaldynamicsintheStraitcouldhaveonthebiologicalprocessesin theAlboranbasin. Thelastoftheworks,presentedinChapterVI,dealswiththeclassification ofthewatermassesinvolvedintheexchangethroughtheStraitofGibraltar. Theθ‐ScharacteristicsoftheMediterraneanwatersdifferveryslightlyand, moreover,theirpresenceorpositioninadeterminedsectionishighlyvariable anddependsnoticeablyonthetidalphase.TheGibraltarInternational campaign(seeChapterVIfordetails)wasconductedtoacquirenewdatawith aMovingVesselProfiler(MVP),whichallowsahighspatialresolutionand relativelyfastaccomplishmentofhydrologicaltransectsand,therefore, providesdatasetsthatminimizethatvariability.Thesedatawereprocessed byclusteranalysisspecificallydevelopedforthesedata,whichassigneda membershiptoeachpointofeveryCTDcastinfunctionofthedistanceto centroidswithθ‐S‐σθ characteristicspredefinedforeachofthepossiblewater massesthataresusceptibletoflowthroughtheStrait.Timevariabilitywas assessedusingthecollectionofCTDmeasurementsinthewesternand easternlimitingsectionsoftheStraitthathavebeenrepeatedsince2005 withintheINGRESprojects. Furtherdetailsaboutthemethodologyaredescribedinmoredetailin eachofthefollowingChapters.
III.The Western Alboran Gyre helps ventilate the Western Mediterranean Deep Water through Gibraltar Cristina Naranjo n , Jesu ´s Garcı ´a Lafuente, Jose C. Sa ´nchez Garrido, Antonio Sa ´nchez Roma ´n, Javier Delgado Cabello Physical Oceanography Group, ETSI Telecomunicacio ´n, University of Ma ´laga, Campus Teatinos, 29071 Ma ´laga, Spain article info Article history: Received 27 April 2011 Received in revised form 19 October 2011 Accepted 24 October 2011 Available online 29 October 2011 Keywords: Western Mediterranean Deep Water Deep water ventilation Strait of Gibraltar Alboran Gyre abstract Variable properties of the Mediterranean outflow and the variability of the Western Alboran Gyre are analyzed by means of 5-year long time series of near bottom potential temperature at Espartel sill in the Strait of Gibraltar and altimetry data in the Alboran Sea. Geostrophic velocity at the southern edge of the gyre and potential temperature at Espartel sill are significantly correlated (correlation coefficient 0.67), suggesting that the intensification of the Alboran Gyre favors the ventilation of Western Mediterranean Deep Water. The analysis of historical temperature profiles shows that Western Mediterranean Deep Water in the Alboran Sea can be suctioned from a layer between 500 and 700 m depth for typical changes of the gyre intensification. 1. Introduction The question of how the Western Mediterranean Deep Water (WMDW) is ventilated through the Strait of Gibraltar is of key importance for the renewal of the Mediterranean Sea. WMDW has to flow over Camarinal sill (CS, see Fig. 1) whose depth (290 m) is less than the depth of the interface between WMDW and the overlying waters in the Alboran Sea. The Mediterranean outflow is formed by the WMDW, the Levantine Intermediate Water (LIW, by far the most important contributor) and other waters of Mediterranean origin (Millot, 2009). WMDW is the densest, hence the deepest, of all the Mediterranean waters and therefore the most energydemanding to be drained out to the Atlantic Ocean over the depths of CS. In a pioneering paper by Stommel et al. (1973), using Bernoulli equation arguments, showed that WMDW could be aspired from depths as great as 700 m in the Alboran basin, a possibility confirmed few years later by Whitehead (1985) in laboratory simulations and by Kinder and Parrilla (1987) who presented for thefirsttimeexperimentalevidenceofWMDWwestofCSthus proving that this water had overflowed the sill. Bryden et al. (1994) and Vargas et al. (2006) showed that the outflow over CS occurs in a pulsating way driven by tidal forces that move back and forth a volume of water three to five times larger than the mean flow (Garcı ´a Lafuente et al., 2000). Tide is also the main source of energy to suction deep waters located in the eastern approach of the Strait and to bring them west of CS into the Atlantic Ocean. Actually, the ventilation of WMDW would comprise two separated, though linked, set of processes. The first one would include all processes leading to make WMDW available for suction in the westernmost part of the Alboran basin close to the Strait of Gibraltar or, in other words, to carry WMDW towards the eastern approach of the Strait. Garcı ´aLafuenteetal.(2009)identified some of these processes such as the re-filling of the western Mediterranean basin with newly formed WMDW, or favorable meteorological conditions able to diminish the net barotropic flow through the StraitofGibraltar,orthepresenceofalargeWesternAlboranGyre (WAG). The second set of processes comprise all physical forcing that influence the total current over CS since the larger the current here, the greater the depth from which deep water in the eastern approach of the Strait can be suctioned. Tidal strength (the fortnightly cycle) plays a relevant role in this mechanism, a fact acknowledged by Kinder and Bryden (1990) who suggested that pure WMDW overflows CS during all spring tide periods. Bryden and Stommel (1982) presented currentmeter observations at 500 m depth in the southwestern Alboran Sea (see the black triangle in Fig. 1) showing a rather permanent flow of WMDW (mean¼4.674.3 cm s 1 ) heading towards the Strait of Gibraltar with potential temperature ( y hereafter) ranging from 12.81 1Cto 13.01 1C, typical of this water. They also observed a marked upwards slope of the isotherms towards the African shore, suggesting that the main path of WMDW ventilation follows the southern part of the Alboran Sea, a result further confirmed by numerical models (Parrilla et al., 1986;Speich et al., 1995 and references therein) and reproduced by the model used in this work as well. On the basis Published in Deep-Sea Research I doi:10.1016/j.dsr.2011.10.003 III. The WAG helps ventilate the WMDW through Gibraltar 13
of these observations Bryden and Stommel (1982) put forward for the first time the idea that the WAG may facilitate the drainage of WMDW, as the gyre provides energy to uplift deep water to depths where it can be easily aspired when they eventually reach the eastern entrance of the Strait. New experimental evidence of direct WMDW suction over CS has been provided by Garcı ´a Lafuente et al. (2007),whoshowedtwo cold pulses typical of WMDW in a time series of y measured at Espartel sill (ES, Fig. 1) during the winters of years 2005 and 2006. More recently, Garcı ´a Lafuente et al. (2009),usingalongerseries y at ES, re-visited the role of the WAG in facilitating the aspiration of WMDW through the Strait. They carried out an Empirical Orthogonal Function analysis of altimetry data of the Alboran Sea and showed that the time coefficients of the second empirical mode were significantly correlated with y at ES. Since this mode captured a significant fraction of the temporal variability of the WAG, they conclude that the presence of a well developed gyre facilitated the suction of WMDW—identified by relative minima of y in the time series—through ES (and, hence, through CS) supporting the Bryden and Stommel (1982) hypothesis. However, part of the WAG variability in Garcı ´a Lafuente et al. (2009) was also explained by the first empirical mode, which was uncorrelated with y at ES. The question deserved further attention and has been addressed in this work usinglongtimeseriesof y and Sea Surface Height (SSH) from altimetry in the Alboran Sea. The Massachusetts Institute of Technology general circulation model (MITgcm, Marshall et al., 1997)has also been run to illustrate some basic aspects of the whole process. 2. Data and data processing 2.1. In situ measurements In October 2004 an oceanographic station was deployed in ES to monitor the Mediterranean outflow. A conductivity–temperature (CT) probe installed 10 m above the sea floor measured the y –S characteristics of the outflow at a sampling rate of 30 min. Due to its proximity to the bottom, the probe detected the densest water flowing out from the Mediterranean Sea. Six conductivity– temperature–depth (CTD) casts along the cross-section CA (Fig. 1) were carried out in June 2009 to show the spatial structure of the Mediterranean water masses in the eastern part of the Strait. The CTD probe was lowered as close as possible to the bottom to register the characteristics of the deepest, densest water. Historical CTD data of the Alboran Sea from the MEDATLAS database (MEDAR group, 2002) have also been used in this study. As in Garcı ´a Lafuente et al. (2009), y series at ES is used here to follow the variability of the WMDW in the outflow. Salinity was another possibility that was discarded because temperature is a better choice to discriminate WMDW against LIW, the other important water mass in the outflow. The y series shows large tidal fluctuations that act as noise for long-term variability on which this work focuses. Following Garcı ´a Lafuente et al. (2007, 2009), tides have been removed by selecting the sample of minimum y within each semidiurnal cycle, which decimates the series to about a value every 12 h. Since it will be correlated to SSH data from AVISO, whose sampling interval is a week, y series has been further decimated to weekly values by extracting the coldest sample every week. The resulting series will be referred to as y min series hereafter. 2.2. Satellite data Geostrophic surface velocities derived from altimetry provided by AVISO have been used to determine the WAG variability. These data have spatial and temporal resolutions of 1/4 degree and one week, respectively, and span the period from October 2004 to October 2009, which coincides with the CT time series. For the purpose of this work, the variable U SWAG defined as the spatial average of the zonal component of the geostrophic velocity in the 6°W30’ 5°W30’ 4°W30’ 3°W 35°N 30’ 36°N 30’ 9°W3 °W3 °E 33°N 36°N 39°N 42°N ES TB CS CA Fig. 1. Map of the Strait of Gibraltar. ES and CS indicate the location of Espartel and Camarinal sills; TB is Tangier Basin and CA indicates the Ceuta-Algeciras CTD section mentioned in the text. Dashed gray line shows the model section used in Fig. 3. Dashed black rectangle in the Alboran Sea indicates the area where MEDATLAS profiles have been selected. Arrows depict the Western Alboran Gyre and indicate surface velocity in August 2008 from altimetry data. Crosses indicate the positions to compute U SWAG (see text Section 2.2). Black triangle indicates the currentmeters’ site in Bryden and Stommel (1982). The diamond represents the Alboran Island. 158 III. The WAG helps ventilate the WMDW through Gibraltar 14
grid points marked in Fig. 1 has been used as a proxy of the WAG variability. Large absolute values of U SWAG correspond to a welldeveloped gyre that, in our hypothesis, has more potential to uplift WMDW from greater depths in the Alboran Sea. 3. Experimental results 3.1. The WAG and the temperature in ES Fig. 2 shows the time evolution of U SWAG and y min and their lowpassed contributions. Its mean value is 24 cm s 1 with a standard deviation (STD) of 711 cm s 1 , and does not exhibit a well-defined seasonality although the low-passed series suggests greater westward velocity in summer months. Taking into account that AVISO geostrophic velocities are inferred from sea level gradients, summer maximum would correspond to a well-developed WAG, in good agreement with previous works dealing with the WAG variability (Vargas-Ya ´n ˜ez et al., 2002). More than six year long time series of y min in ES shows cold peaks close to 13 1C(Fig. 2b) that roughly match the time evolution of U SWAG . Warmer values above 13.2 1Carefoundbytheendofthe different years when U SWAG also presents (absolute) minimum values associated with a weaker WAG. The cross correlation coefficient between both series is 0.46 at the 95% significance level. The coefficient raises to 0.61 if high-frequency variability is removed (low-passed series) and it further increases to 0.67 if y min series is shifted back 2 weeks with respect to U SWAG . This fact is interpreted as a time-delayed response if the WAG intensity is a driving force for uplifting WMDW in the Alboran Sea. 3.2. Spatial evolution of potential temperature depth The mean value of y min at ES is 13.10 1C, a value that can be taken as representative of the WMDW that leaves the Strait of Gibraltar through ES. Using long time series of simultaneous observations in CS and ES, Garcı ´aLafuenteetal.(2011)showed that vigorous tidal mixing in Tangier basin (TB, Fig. 1) increases the WMDW temperature by 0.12 1C on its way from CS to ES, so that the representative y min at CS would be 12.98 1C. Eastwards of CS the cross-area of the outflow increases markedly and the Mediterranean layer flows rather slowly. This fact reduces greatly water mixing, so we can assume that y remains constant along streamlines. Fig. 3ashows that, in the eastern approach of the Strait, WMDW still maintains the tendency of remaining attached to the African side and that y ¼12.98 1C can be found in the depth range 300–600 m, shallower in the south. Historical MEDAR y data in the southern half of the western Alboran Sea (Fig. 4) show that this water resides, on average, at 520 m depth, although the banking of WMDW against the African slope would allow this water to be found at shallower depths more to the south. This description shows the continuous uplifting of the WMDW from the Alboran Sea to the west until it finally crosses ES and flows out to the North Atlantic Ocean. 3.3. Temporal variability Taking the STD as representative of the typical variability, y min at ES would be 13.1070.05 1C although Fig. 2b shows that it fluctuates between 13.01 1C and an anomalous maximum of 13.26 1C by the end of 2009. Correcting these values for mixing in Tangier basin, y min at CS would be 12.9870.05 1C (from 12.93 1C to 13.03 1C interval), which would be also representative of y min variability eastwards of CS in our hypothesis of very reduced mixing. Isotherms 12.93 1C and 13.03 1C are marked with −60 −40 −20 0 20 u (cm/s) 2005 2006 2007 2008 2009 2010 13 13.1 13.2 13.3 T (ºC) Fig. 2. (a) Gray line: U SWAG time series derived from altimetry data. Black line: low-pass filter signal. (b) Gray line: y min time series in Espartel sill. Black line: lowpass filter signal. Dashed line: the mean minimum and maximum values of y, 13.05 and 13.15 1C, respectively (y7STD). 12.89 12.98 13.1 13.3 14 12.93 13.03 Depth (m) 800 700 600 500 400 300 200 100 0 12.8 12.9 13 13.5 Depth (m) 35.9 35.95 36 36.05 36.1 36.15 800 700 600 500 400 300 200 100 0 Fig. 3. (a) Potential temperature measure in cross-section CA marked in Fig. 1, x-axis is latitude (1N); CTD data were collected from 10:16 to 14:09 h in June 16th, 2009. (b) Model results in a similar section (dashed gray line in Fig. 1). III. The WAG helps ventilate the WMDW through Gibraltar 15
gray lines in Fig. 3a and indicate the layer in the eastern part of the Strait that would be evacuated to the North Atlantic under the typical fluctuations of the forcing responsible for the observed variability at ES. It occupies the depth range between 670 m and 250 m depth. However, extreme y min values at ES of 13.01 1Cin Fig. 2a, which would reduce to 12.89 1C at CS and eastwards, would displace the lower bound to the sea bottom (Fig. 3a), suggesting that all the Mediterranean layer in the eastern approach of the Strait is potentially drainable to the Atlantic under extreme forcing. Further east, the depth range associated with the STD-size fluctuations at ES is from 440 to 600 m if we only consider the mean profile (gray portion of the mean profile in Fig. 4) but increases from 230 to 780 m if the variability of the profile, indicated by the dashed lines, is taken into account (doubleheaded arrow in Fig. 4). In case of extreme fluctuations, this depth could exceed 800–900 m. 4. Numerical model The correlation between U SWAG and y min supports the hypothesis of Bryden and Stommel (1982) that the WAG helps to uplift WMDW. For that, the WMDW in the western Alboran Sea must flow westwards along the African slope in the same direction as the overlaying Atlantic layer in the southern part of the WAG. The observations of Bryden and Stommel (1982) confirm they do but whether this preferred path of the WMDW is determined by the general circulation of the Mediterranean Sea (remote influence) or it is more a consequence of the water exchange through the Strait (local influence) remains unclear. A second issue to investigate is to which extent the uplifting of WMDW in the southern Alboran Sea is directly influenced by the dynamic of the Strait. Both topics have been addressed by process-oriented numerical simulations using the MITgcm. The model domain extends from 91Wto11E and is horizontally discretized by a curvilinear orthogonal grid with minimum grid size in the Strait ( D x, D y¼500 m),gradually increasing to 4–5 km in the Alboran Sea, and 8–10 km near the open boundaries (Fig. 5a). The vertical spatial coordinate is discretized with 46 z-levels. The model was run from a very simplified initial state in which the Mediterranean side was filled with two water masses representing LIW and WMDW of y –Scharacteristics (13.2 1C, 38.55) and (12.8 1C, 38.45), respectively. Initially LIW occupied the 600 uppermost meters, the rest being filled with WMDW. The Atlantic side was occupied by Atlantic water of (15.5 1C, 36.2) y –S characteristics. Both basins, initially at rest, were separated by a barrier that was removed at t¼0 (lock-exchange initial condition). Orlanski radiation boundary condition (Orlanski, 1976) for both velocity and tracers was imposed at the open boundaries. No other external forces (meteorological, tides) were applied. The model evolves to a final state characterized by a well developed WAG occupying the western Alboran Sea encircled by a swift Atlantic jet of 1 ms 1 typical speed (Fig. 5b) that leaves the domain by the eastern boundary to form eventually the Algerian current. The deep circulation (Fig. 5c) shows a vein of LIW/WMDW at 500 m depth that enters the western Alboran basin along the trough located north of the Alboran Island and flows southwestwards to the African continental slope where it veers towards the Strait to contribute to the outflow. Its mean speed at the grid points used to compute U SWAG is 4.8 cm s 1 , quite close to the mean value of 4.6 cm s 1 provided by Bryden and Stommel (1982) at the same depth at a short distance to the west. The spatial distribution of isotherms across the Alboran basin (Fig. 5d) reproduces the pattern reported by these authors and illustrates the uplift of WMDW in the south where it is noticeably shallower (o300 m) than it was initially (600 m). Fig. 3b shows that the model also accumulates WMDW in the southern part of the Strait, matching the observations in Fig. 3a. Since no forcing was imposed on the open boundaries, this pattern would be linked to the dynamics of the exchange rather than to remote forcing acting across the boundaries. To assess the direct influence of the Bernoulli aspiration caused by high velocities over CS on the uplifting of WMDW in the Alboran Sea, another numerical experiment was carried out in which the WAG was isolated from the strait dynamics by closing the strait at t¼100 day, 50 days before the situation presented in Fig. 5b. The surface circulation 50 days later is yet characterized by a large WAG (Fig. 6a), now extending more to the north as it is not constrained by the Atlantic jet. Fig. 6b and c shows that the depth of the isotherm 12.85 1C is fairly similar in the closed and open-strait experiments except for the small area in front of the Strait’s entrance where the closed-strait simulation shows it deeper and somewhat displaced northeastwards (consequence of the new enlarged shape of the WAG). Farther away from this area the circulation appears to be quite independent of the strait status indicating that the gyre is mainly responsible for the banking of WMDW against the African slope. On the other hand, the shallowness of the isotherm in front of the strait in the openstrait run is attributable to the Bernoulli suction, whose effect is only noticeable around this area. 5. Discussion and conclusions The analysis of the different datasets in this work has showed the good correlation between the geostrophic velocity of the WAG 12.7 12.8 12.9 13 13.1 13.2 1000 900 800 700 600 500 400 300 Depth (m) 15 20 25 1000 200 400 600 800 1200 Z0 θ (°C) Fig. 4. Spatially-averaged vertical profiles of potential temperature below 260 m depth from the MEDATLAS database in the southern part of the western Alboran basin (the inset shows the whole profile). Dashed lines indicate 71 STD, shadow area indicates de y min interval variation in ES corrected by the mixing factor. 160 III. The WAG helps ventilate the WMDW through Gibraltar 16
and the near-bottom potential temperature at ES, which gives support to the hypothesis that a well developed WAG helps ventilate WMDW from the Alboran Sea and, hence, from the Mediterranean Sea. The ventilation is a two-step process starting with the uplift of WMDW produced by the WAG, whose final result is to facilitate the arrival of WMDW to the eastern entrance of the strait from where it is eventually aspired by the enhanced currents over CS. Once the sill has been surpassed, the WMDW flows without major topographic restrictions towards the Atlantic Ocean while mixing with the surrounding waters. Some attempts have been made to investigate the depth from which WMDW can be aspired by this two-step mechanism. Using y as a tracer, historical data from MEDAR in the southern half of the western Alboran basin (Fig. 4) indicate that y min ¼12.98 1C (the mean value of y min at CS) is found at z 0 ¼520 m. This would be the representative depth from which the coldest/densest WMDW can be suctioned. Numerical results presented in Fig. 5c strongly suggest that water at 500 m depth flowing at 4.8 cm s 1 leaves the Mediterranean Sea. The coincidence in turn suggests that this velocity would correspond to the minimum value able to uplift resident WMDW in Alboran over the sills of the Strait. When the time variability of y min , scaled by its STD at ES, is incorporated into the analysis then z 0 becomes a layer between 440 and 610 m (thick gray line in Fig. 4). If the correlation between y min and U SWAG stands, the upper bound of this layer would coincide with the maximum depth from which a weakened WAG of U SWAG ¼11 cm s 1 (the mean, 24 cm s 1 , minus the STD, 13 cm s 1 ) could uplift, on average, water for feeding the outflow. Similarly, the lower bound would be the maximum depth from which a strengthened WAG of U SWAG ¼37 cm s 1 (meanþSTD) would suction water. Obviously the WMDW characteristics in the suctioned water will be much clearer in the second case. Under extreme situations, both the lower and upper ends of the depth interval will change. For instance, judging from the y transect in Fig. 3a, the lower bound can move downwards to 800–900 m if we consider the extreme situations that forces the periodically observed presence of y min as low as 13.01 1C in the Mediterranean outflow at ES (Fig. 2b). On the contrary, there are unusual situations in which the WAG in the western Alboran Sea disappears and is replaced by a coastal jet of Atlantic water flowing attached to the African shore (Vargas-Ya ´n ˜ez et al., 2002). In this case, the downwards transfer of momentum would act to stop the flow of WMDW towards the Strait while the general circulation would facilitate a major drainage of the LIW residing at intermediate depths in the northern part of the Alboran basin, thus changing markedly the proportion of the different Mediterranean waters in the outflow. It must be noticed that the water exchange is responsible for the WAG generation, which in turn uplifts WMDW from the deep part of the Alboran basin and makes it available for suction at the eastern approach of the Strait. From this point of view, the two-step process of WMDW evacuation from the western Alboran 8°W6 °W4 °W2 °W0 ° 33°N 34°N 35°N 36°N 37°N 38°N 39°N 30’ 5°W30’ 4°W30’ 3°W 30’ 5°W30’ 4°W30’ 3°W 20’ 40’ 36°N 20’ 40’ 0.15 m/s 12.9 13 13.1 13.5 14. 5 15 Depth (m) 35.2 35.4 35.6 35.8 36 36.2 36.4 36.6 1500 1000 500 0 S S Fig. 5. (a) Computational grid used in the study (only 50% of the grid points are shown for the sake of clarity), the black rectangle encloses the Strait of Gibraltar and the Western Alboran Sea. (b) Modeled surface velocity. The diamond represents Alboran Island. (c) Same as (a) at z¼500 m depth. Isobaths z¼100, 200, y, 2000 m are given by gray lines. (d) Cross section S, (panel c) of temperature and zonal velocity. Light contours are for negative velocities. Darker contours indicate areas of uo0.02 m s 1 . III. The WAG helps ventilate the WMDW through Gibraltar 17
Tides modify the hydrological properties of inflowing water The most apparent local effect of tides in the SoG is the enhanced mixing driven by shear instabilities and turbulence, whose most noticeable results are the thickening of the interfacial layer and the modification of the hydrological properties of their water. Both issues are addressed in this section using the RGSM model under tidal and non-tidal runs. The upper panels of Fig. 3show the interfacial layer thickness produced by each run, calculated by fitting a hyperbolic tangent function to the vertical salinity profiles, following Sannino et al. (2007). Lower panels indicate the depth of the middle interface, which coincides with the inflection point of the fitted hyperbolic tangent function. The panels show that the interfacial layer is shallower in the tidal run and, consequently, covers a greater horizontal extension. Despite this increment, west of CS the thickness of the interfacial layer is much the same in both runs, but the layer becomes markedly thicker (up to 90 m in certain regions) eastwards of CS in the tidal run (Fig. 3a and b). As analyzed in Garcia-Lafuente et al. (2013), the reason for this pattern is the eastward advection of the waters that had been mixed in the Tangier basin during the previous rising tide and in the subsequent release and eastward progression of the internal hydraulic jump formed leeward of CS, which provides energy for maintaining high rates of mixing. The shoaling of the interface is primarily noticeable in north-eastern part of the Strait, where the depth of the middle interface changes from 70 m in non-tidal run to near the surface in the tidal one. The joint effect of the shoaling and thickening of the interface in the eastern half of the SoG in the tidal run is to carry water from below to the surface layers in a more effective way than in the non-tidal run. In other words, AWs are expected to be saltier and colder under tidal forcing and, hence, denser and less buoyant. These AWs are directly advected into the adjacent Alborán Sea basin so that the tidal run gives colder temperatures not only within the dimensions of the SoG but also beyond its limits. Fig. 4 shows the surface temperature difference between the tidal and non-tidal simulations. The greatest differences are found in the eastern exit of the SoG and along the expected path of the Atlantic Jet in the western Alborán Sea, which confirms that in the tidal run the Jet carries colder water because of the enhanced tidal mixing in the SoG. Even when reduced, this difference is still detectable in the area of the Almería-Orán front at the eastern end of the Alborán basin and along the path of the Algerian current further east. These visual results are confirmed by computing the mean temperature of the incoming AWs across AC section: the inflow is 0.37 °C colder in the tidal run (15.42 °C versus 15.79 °C) and also 0.47 units saltier (36.63 versus 36.16). The joint effect is an inflow 0.45 kg m 3 denser in the case of the tidal run and, consequently, the flux of advected buoyancy into the Alborán Sea diminishes. For instance, the advected buoyancy within the 100 upper meters of the water column, computed at the AC section, is 1.13 times lower in the tidal run (2.505 ms 2 versus 2.235 ms 2 ), which may have far-field significant consequences, as the cold and salty signature generated in the SoG is exported by the Algerian current to the interior of the MS. Fig. 2. Atlantic and Mediterranean transports versus the salinity of the isohaline used to calculate it. Black and gray lines correspond to tidal and no-tidal runs, respectively. Arrows indicate the maximum transport with the same color code. Only one arrow is seen at ES because the isohaline that maximizes the transport is the same in both runs. Table 1 Computed transports (in Sv) at ES, CS and AC sections (see Fig. 1). The first two columns indicate the isohalines that maximizes the transports in each section (see Fig. 2). The third and fourth columns are the mean transport in the tidal and non-tidal runs, respectively, and the fifth column gives the increment (percentage) that tides cause on the transports. The term ( ul hl), which is representative of the slow-varying term in Eq. (1), is shown in the sixth column. Eddy-fluxes (u0 lh0 l) and the percentage of the total flow that they account for are shown in the last two columns. S Tid S No-Tid Q Tid Q No-Tid D (%) u i h i u 0 i h 0 i %u 0 i h 0 i ES 36.6 36.6 0.988 0.935 5.7 0.957 0.031 3.0 CS 37.3 37.1 0.844 0.793 6.4 0.522 0.322 38 AC 37.5 37.3 0.893 0.823 8.5 0.880 0.013 1.4 IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 24
Tides favor Mediterranean deep water ventilation through the SoG The GMSM model outputs are examined in this section to address the influence of tides on the draining of the WMDW toward the Atlantic Ocean through the shallow SoG. Prior to investigating the topic, it is convenient to check the model performance, which can be assessed from Fig. 5. The left panel shows the mean temperature and salinity in the AC section (see Fig. 1) provided by the model, whereas panel b) presents observations collected in a nearby cross section. The hydrological properties of the modeled water masses as well as their spatial distribution match remarkably well the observations, particularly in the lower layer, which supports the use of the GMSM model to accomplish the study of the WMDW drainage. The influence of tides is assessed by comparing the volume of WMDW flowing across the AC section computed for the tidal and non-tidal runs. In the present study, water colder than 13 °C (potential temperature) in the SoG area is considered as WMDW. This criterion, which has been traditionally assumed in the literature (Bryden and Stommel, 1982; Kinder and Parrilla, 1987; Kinder and Bryden, 1990; Garcia-Lafuente et al., 2007, 2009; Naranjo et al., 2012), is convenient in this case because the isotherm h=13°C is sufficiently far from the bottom in both the model and the observations (see Fig. 5) to ensure that it always lies well outside of the bottom boundary layer. Therefore, the estimated flows of WMDW are not critically dependent on the physics of the bottom boundary layer, which is not as well resolved as the ocean interior in the numerical models. Consequently, this weakness is not a major concern for our results. Table 2 shows the estimated WMDW flow across AC section using the five last years of the GMSM hindcast. Results for the entire 5-year period along with the values for each single year are presented in Table 2. The 5-year average is 28% (0.051 Sv) Fig. 3. (a) Interface thickness, in meters, for the non-tidal run. (b) Same as (a) for the tidal run. (c) Mean depth of the interface, in meters, for the non-tidal run. (d) Same as (c) for the tidal run. See text for details about the way they are computed. Fig. 4. Surface temperature difference (h tides h no-tides ,in°C) between the tidal and non-tidal run in the Alborán Sea. Fig. 5. (a) Mean potential temperature and salinity in section AC (see Fig. 1) derived from the GMSM model outputs. The filled color contours are potential temperature (color bar on the right), while the labeled black contours show the salinity. The white contour is the isotherm h=13°C used as the upper limit of WMDW (see text). (b) CTD profiles collected at the – nearly – same section as in panel (a). The color bar is again for temperature, the dashed black contours are salinity, and the solid black line is the isotherm h=13°C. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 25
higher in the tidal run, although the percentage fluctuates between 10% and 50% depending on the year. The relevant result regarding this study is the visible enhancement of the WMDW aspiration in the tidal run, but the high year-to-year variability during the simulated period, which we ascribe to the internal variability of the MS and of the WMDW formation processes should also be noted. A similar procedure has been carried out to estimate the WMDW flow through section WA at 3°W in the Alborán Sea (see Fig. 1). Although the overall estimates point at a slight increase of WMDW volume transport toward the SoG in the tidal run, the analysis is not statistically conclusive. The fact that the comparison of tidal and non-tidal runs does not clearly detect the effect of tides on the WMDW flow through this section, located around 200 km to the east of the SoG, is interpreted as the weakening with the distance of the direct suction by tides. Basin scale processes, WMDW formation Deep convection that leads to the formation of WMDW in the northwestern MS is among the most relevant oceanographic processes taking place in this Sea and, as such, it has been extensively discussed in the literature (MEDOC Group, 1969; Schott et al., 1996; Herrmann et al., 2008; Gascard, 1991; Smith et al., 2008; Marshall and Schott, 1999). The convection process involves several phases. The first one is the preconditioning phase (Gascard and Richez, 1985; MEDOC Group, 1969), which comprises the densification processes (buoyancy losses) that the Atlantic surface water undergoes since entering the MS through the SoG until it sinks in winter during the second phase of the convection process. Section ‘Tides modify the hydrological properties of inflowing water’ of this study has shown that the AW inflow is colder and saltier, i.e., less buoyant, when tides are included in the model and that those signatures are carried eastwards by the Algerian current (Fig. 4). In other words, the AW in the tidal run is more strongly preconditioned than in the non-tidal run. This section investigates whether or not this difference at the source point (SoG) has a far-field influence on the formation of WMDW in the Gulf of Lion area (Fig. 1). The outputs of the GMSM model run with and without tides are compared to address the issue. Two bulk variables or proxies are analyzed herein: the mixed layer depth (MLD) and the surface area susceptible to participate in the deep convection event (hereinafter DCA). The MLD at each grid point has been defined as the distance from the surface to the depth where the vertical diffusion coefficient reaches a threshold value of 0.04 m 2 s 1 , following Herrmann et al. (2008). DCA has been estimated as the area where the surface water is denser than 1029.10 kg m 3 , a criterion widely used in the literature of deep convection (Schott et al., 1996; Smith et al., 2008; Pinardi et al., 2013). Fig. 6 shows the MLD averaged from January to the end of April for every year from 1963 to 1967. Maximum mean MLD reaches 1500 m during 1963, 1964 and 1965, values that are quite similar to those presented by Schott et al. (1996) and Herrmann et al. (2008) in the same area. Except for 1963 MLD in the tidal run is always greater than in the non-tidal run, which is expected if tidal runs better precondition the inflowing AW. Fig. 7 presents the evolution of the estimated DCA during the winter months (January to end of April). The greatest DCA is reached in March 1965 (almost 10 10 10 m 2 in the tidal run, 2 10 10 m 2 more than in non-tidal run) followed by 1963; during this year, the DCA is larger in the non-tidal run as was the case for the MLD as well. During the other years, however, the tidal run provides larger DCA than the nontidal run. A deep water formation rate has been calculated following Lascaratos and Nittis (1998) and is presented in Fig. 8a. Since it combines the former MLD and DCA variables, the comparison of tidal and non-tidal runs does not provide new results but confirms the previous ones, the non-tidal run gives higher formation rates in 1963 and lower rates in all the other years. The results presented in Figs. 6 and 7 deserve some comentaries despite the topic not being the scope of this work. Focussing on the tidal run, the model simulations predict a great year-to-year variability in all variables, namely MLD, DCA and rate of formation. Notice the reduced DCA in years 1966 and 1967, which is in agreement with the also diminished MLD during these years shown in Fig. 6, which in turn gives the minimum formation rate of barely 0.2 Sv in 1967 (Fig. 8a). On the opposite end is 1965 when the rate of formation approaches 4 Sv, a rather high value that would correspond to an exceptionally productive year. In any case, they fall inside the interval of deep water formation rates reported by other authors that ranges between 0.3 and 6 Sv (Tziperman and Speer, 1994;Krahmann, 1997; Castellari et al., 2000; Herrmann et al., 2008;Béranger et al., 2009;Beuvier et al., 2012;Pinardi Table 2 Time-averaged outflow of water with h<13°C across the AC section (see Fig. 1), which is identified with WMDW in this study. The first row is the five-year average and the remaining rows indicate the year-averaged WMDW outflow. Transport with and without tides are shown in the first two columns. Last column gives the increment between tidal and non-tidal run ( D Q) and its percentage (%). Period Block 1: h<13°C Q TID (Sv) Q NOT (Sv) D Q(Sv)/% (1963–1968) 0.233 0.182 0.051/28 1963 0.091 0.060 0.031/51 1964 0.285 0.236 0.049/21 1965 0.307 0.230 0.077/33 1966 0.334 0.303 0.031/10 1967 0.211 0.158 0.052/33 Fig. 6. Mixed layer depth in meters (color bar) in the Gulf of Lion area (see Fig. 1) computed using a threshold value for the vertical diffusion coefficient of 0.04 m 2 s 2 . Upper and lower rows correspond to tidal and non-tidal runs, respectively. The contours represent the January-to-April (both included) average. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 26
et al., 2013; Schott et al., 1994, 1996; Send et al., 1995; Schroeder et al., 2008; Durrieu de Madron et al., 2013). Fig. 8b gives the time evolution of the estimated volume of deep water resident in the region of the Gulf of Lion as a function of time. The large formation rate in 1963 filled up the bottom layer, which started draining out once the winter formation processes had finished. The way it did is different for the tidal run, in which the volume decrease stopped after a very short time of drainage and remained constant thereinafter, and the non-tidal run that featured the exponential-like decay expected in this type of physical problem, where the rate of draining must be proportional to the volume of water that is being drained. Notice that, after this year, the drainage after the refilling of the basin predicted in the tidal run shapes an exponential curve reasonably well. We lack of a suitable explanation for the anomalous drainage in 1963, which could be the reason explaining the apparent paradox (according to our hypothesis) of why the non-tidal run predicts a higher rate of formation this year than the tidal run. The footprint of the winter of 1964 is the small cusp, more visible in the tidal run, while 1966 and 1967 with such small rates of deep water formation, hardly leave any signal in the curve. In contrast, 1965 doubled the volume of deep water stored in the Gulf of Lion due to the extraordinary rate of formation, much greater in the tidal run as seen in Fig. 8a. Discussion and conclusion This work examines the effect of tides in the SoG on several oceanographic processes whose spatial scales range from local to basin-wide. The most significant outcome is their contribution to increase the mean exchange that is described here, but there are other effects intuitively related to tidal forcing that have not been thoroughly addressed to our knowledge yet. Some of them are investigated in the different sections of this work by comparing the outputs of two numerical models which have been run with and without tidal forcing. As expected, the model results confirm that the inclusion of tides generate eddy fluxes that increase the long-term exchange. Despite its very similar contribution in the different sections of the SoG (6–8%, see Table 1), the way the increment is achieved differs between sections. At the main CS section, eddy fluxes play a key role and represent a fundamental process to increase the long-term exchange, a result that agrees with previous findings (Bryden et al., 1994; Vargas et al., 2006). At the boundary sections of AC and ES the contribution of eddy fluxes is very small although the increment of the long-term flow due to tides is similar or even higher than at CS (8.5% at AC versus 6.4% at CS, see Table 1). A remarkable result that stems from the weakness of the eddy fluxes at the boundary sections is that the flows could be estimated satisfactorily there using only the slowly-varying term in Eq. (1).It is an interesting outcome for experimental studies because the computation of eddy fluxes from observations poses serious challenges. Actually, some experimental studies (Sanchez-Román et al., 2009; Garcia-Lafuente et al., 2000) have made already use of this flow property. A second effect of the tides is the enhancement of the mixing between Mediterranean and Atlantic waters within the strait itself. The energy for mixing is mainly released in the various supercritical-to-subcritical flow transitions occurring in the Tangier basin (Sanchez-Garrido et al., 2011). The final outcome is the thickening and shoaling of the interfacial layer in the SoG, which is favored by the propagation of nonlinear internal waves and entrainment of MW by the Atlantic jet, an issue that has been recently addressed by Garcia-Lafuente et al. (2013). These processes are nearly inhibited in the absence of tides, a fact that is reflected by the very thin interfacial mixing layer in the non-tidal simulation (Fig. 3). In addition to the significant effect that the shoaling of the interface may have on biological communities, the tidally-induced mixing also makes the Atlantic jet saltier (0.47 units) and colder (0.37 °C). This water is finally advected to the Alborán Sea which therefore shows colder surface waters almost everywhere, although it is more visible along the mean path of the Atlantic jet and, particularly as it exits the SoG (Fig. 4). As shown recently by Sanchez-Garrido et al. (2013), the 4 °C colder surface water in this area obtained in the tidal simulation has its origin possibly in the advection of positive shear vorticity generated by the interaction of tidal currents with the solid northern boundary of the SoG. If so, this signature would be more related to the local doming of isotherms associated with the enhancement of the cyclonic circulation rather Fig. 7. Time evolution of the Deep Convection Area in the Gulf of Lion (see Fig. 1) from January to the end of April of every year. Solid and dashed lines represent the tidal and non-tidal runs respectively. Fig. 8. (a) Deep water formation rate (in Sv) calculated as (V A V B )/T, where V A is the maximum volume in a particular year, V B is the minimum volume before the convection event and T= 1 year = 3.15 10 7 s(Lascaratos and Nittis, 1998). (b) Deep water volume measured in the volume control (Fig. 1). Black and gray colors correspond to the tidal and non-tidal run, respectively. IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 27
than to the direct advection of colder water from the strait. The downstream temperature anomaly in the Alborán Sea would be a consequence of this process, at least partially, but it does not modify its tidal origin. Of particular interest is the fact that the cold signature is still clearly visible in the Almería-Orán front and Algerian current, at the eastern exit of the Alborán Sea (dark blue strip over this area in Fig. 4). We hypothesize that the denser AW produced by mixing in the SoG in the tidal run facilitates the formation of WMDW in the Gulf of Lion. The comparison of two bulk variables, namely MLD and DCA, suggests that indeed, the tidal run tends to produce more volume of WMDW. However the first year of the hindcast does not behave so, as the non-tidal run produced more volume of WMDW (Fig. 8). As mentioned in Section ‘Basin scale processes, WMDW formation’, we lack an explanation for this behavior, which is further confounded by the fact that this is the only year when the drainage of the WMDW out of the control volume in the tidal run does not follow an expected exponential decay with time (Fig. 7b). Therefore, the results concerning 1963 must be interpreted with caution. What both models predict in a similar way is the marked year-to-year variability driven by the atmospheric forcing, a variability that has also been found in other studies (Herrmann et al., 2008; Pinardi et al., 2013). The last issue addressed of whether or not tides favor the ventilation of the deep WMDW layer has a positive answer according to our results. Table 2 indicates that the outflow of WMDW (defined as the water colder than h=13°C in the neighborhood of the SoG) increases by nearly 30% in the tidal run. This percentage is greater than the 6–7% increment of the long-term outflow due to tides (column 5 in Table 1). The difference in percentages suggests that the drainage of WMDW is specially aided by tides and it is more favored than any of the other Mediterranean waters participating in the outflow.Table 2 also shows an interannual variability that is apparently related to the variability of the WMDW reservoir in the Gulf of Lion revealed in Fig. 8. For instance, the outflow of WMDW reaches its maximum in 1966, a year after the large WMDW formation that occurred in 1965. Taking into account the time the signal will take totravelfromtheGulfofLiontotheSoG,thisdelayseemsreasonable (Garcia-Lafuente et al., 2007, 2009). The trend of the WMDW stored in the control volume in the Gulf of Lion is to diminish after 1965 (Fig. 7a), a trend that seemsto be followed by the outflow of WMDW with a year delay (Table 2). In conclusion, our study has provided evidence that tides in the SoG has local (increase of the long-term exchange, noticeable tidally-driven mixing, eventually exported to the MS), regional or short-range (colder and saltier inflow, enhanced aspiration of WMDW) and long-range (influence in deep convection processes) influences. Of all them, the last one is more open to debate because a convincing conclusion requires much longer simulations that include tides in the forcing terms, which are computationally unaffordable at this moment. Acknowledgments This work is a contribution to the Spanish funded National Project INGRES 3 (CTM2010-21229) with partially financial support of Project P08-RNM-3738 from Plan Andaluz de Investigación. Cristina Naranjo acknowledges the fellowship BES-2011-043421 from the Ministry of Economy and Competitiveness – Spain. The Mediterranean tidal simulation has been carried out thanks to a CINECA ISCRA grant. The authors are also grateful to the Supercomputing and Bioinformatics (SCBI) Center of the University of Malaga and to CRESCO Supercomputing Facilities at ENEA (www.cresco.enea.it). This is the publication no. 62 from CEIMAR Publication Series. References Armi, L., Farmer, D., 1985. The internal hydraulics of the Strait of Gibraltar and associated sills and narrows. Oceanologica Acta 8, 37–46. Armi, L., Farmer, D.M., 1986. Maximal two-layer exchange through a contraction with barotropic net flow. Journal of Fluid Mechanics 164, 27–51. Artale, V., Calmanti, S., Carillo, A., Dell’Aquila, A., Herrmann, M., Pisacane, G., Ruti, P.M., Sannino, G., Struglia, M.V., Giorgi, F., Bi, X., Pal, J.S., Rauscher, S., The PROTHEUS Group, 2009. An atmosphere-ocean regional climate model for the Mediterranean area: assessment of a present climate simulation. Climate Dynamics 35, 721–740. http://dx.doi.org/10.1007/s00382-009-0691-8. Baschek, B., Send, U., Garcia-Lafuente, J., Candela, J., 2001. Transport estimates in the Strait of Gibraltar with a tidal inverse model. Journal of Geophysical Research 106 (C12), 31033–31044. http://dx.doi.org/10.1029/2000JC000458. Béranger, K., Testor, P., Crépon, M., 2009. Modelling water mass formation in the Gulf of Lion (Mediterranean Sea). In: Briand, F. (Ed.), Dynamics of Mediterranean Deep Waters, CIESM Workshop Monogr, vol. 38. Mediterr. Sci. Comm., Monaco, pp. 91–100. Beuvier, J., Béranger, K., Lebeaupin-Brossier, C., Somot, S., Sevault, F., Drillet, Y., Bourdallé-Badie, R., Ferry, N., Lyard, F., 2012. Spreading of the Western Mediterranean Deep Water after winter 2005: time scales and deep cyclone transport. Journal of Geophysical Research 117, C07022. http://dx.doi.org/ 10.1029/2011JC007679. Bray, N.A., Ochoa, J., Kinder, T.H., 1995. The role of the interface in exchange through the Strait of Gibraltar. Journal of Geophysical Research 100 (C6), 10755–10776. http://dx.doi.org/10.1029/95JC00381. Bryden, H.L., Kinder, T.H., 1991. Steady two-layer exchange through the Strait of Gibraltar. Deep Sea Research Part A 38, S445–S463. http://dx.doi.org/10.1016/ S0198-0149(12)80020-3. Bryden, H.L., Stommel, H.M., 1982. Origin of the Mediterranean outflow. Journal of Marine Research 40, 55–71. Bryden, H.L., Candela, J., Kinder, T.H., 1994. Exchange through the Strait of Gibraltar. Progress in Oceanography 33, 201–248. http://dx.doi.org/10.1016/00796611(94)90028-0. Carrère, L., Lyard, F., 2003. Modeling the barotropic response of the global ocean to atmospheric wind and pressure forcing-comparisons with observations. Geophysical Research Letters 30, 1275. http://dx.doi.org/10.1029/ 2002GL016473. Castellari, S., Pinardi, N., Leaman, K., 2000. Simulation of the water mass formation processes in the Mediterranean Sea: influence of the time frequency of the atmospheric forcing. Journal of Geophysical Research 105 (C10), 24157–24181. http://dx.doi.org/10.1029/2000JC900055. Durrieu de Madron, X., Houpert, L., Puig, P., Sanchez-Vidal, A., Testor, P., Bosse, A., Estournel, C., Somot, S., Bourrin, F., Bouin, M.N., Beauverger, M., Beguery, L., Calafat, A., Canals, M., Cassou, C., Coppola, L., Dausse, D., D’Ortenzio, F., Font, J., Heussner, S., Kunesch, S., Lefevre, D., Le Goff, H., Martín, J., Mortier, L., Palanques, A., Raimbault, P., 2013. Interaction of dense shelf water cascading and open-sea convection in the northwestern Mediterranean during winter 2012. Geophysical Research Letters 40, 1379–1385. http://dx.doi.org/10.1002/ grl.50331. Farmer, D.M., Armi, L., 1986. Maximal 2-layer exchange over a sill and through the combination of a sill and contraction with barotropic flow. Journal of Fluid Mechanics 164, 53–76. Farmer, D.M., Armi, L., 1988. The flow of Mediterranean Water through the Strait of Gibraltar. Progress in Oceanography 21 (1), 1–103. Garcia-Lafuente, J., Vargas, J.M., Plaza, F., Sarhan, T., Candela, J., Bascheck, B., 2000. Tide at the eastern section of the Strait of Gibraltar. Journal of Geophysical Research 105 (C6), 14197–14213. http://dx.doi.org/10.1029/2000JC900007. Garcia-Lafuente, J., Delgado, J., Vargas, J.M., Vargas, M., Plaza, F., Sarhan, T., 2002a. Low-frequency variability of the exchanged flows through the Strait of Gibraltar during CANIGO. Deep Sea Research Part II: Topical Studies in Oceanography 49 (19), 4051–4067. http://dx.doi.org/10.1016/S0967-0645(02)00142-X. Garcia-Lafuente, J., Alvarez Fanjul, E., Vargas, J.M., Ratsimandresy, A.W., 2002b. Subinertial variability in the flow through the Strait of Gibraltar. Journal of Geophysical Research 107 (C10), 3168. http://dx.doi.org/10.1029/ 2001JC001104. Garcia-Lafuente, J., Sanchez Roman, A., Díaz del Río, G., Sannino, G., SanchezGarrido, J.C., 2007. Recent observations of seasonal variability of the Mediterranean outflow in the Strait of Gibraltar. Journal of Geophysical Research 112, C10005. http://dx.doi.org/10.1029/2006JC003992. Garcia-Lafuente, J., Delgado, J., Sanchez-Roman, A., Soto, J., Carracedo, L., Díaz-delRío, G., 2009. Interannual variability of the Mediterranean outflow observed in Espartel sill, western Strait of Gibraltar. Journal of Geophysical Research 114, C10018. http://dx.doi.org/10.1029/2009JC005496. Garcia-Lafuente, J., Sanchez-Roman, A., Naranjo, C., Sanchez-Garrido, J.C., 2011. The very first transformation of the Mediterranean outflow in the Strait of Gibraltar. Journal of Geophysical Research 116, C07010. http://dx.doi.org/10.1029/ 2011JC006967. Garcia-Lafuente, J., Pozas, E.B., Sanchez-Garrido, J.C., Sannino, G., Sammartino, S., 2013. The interface mixing layer and the tidal dynamics at the eastern part of the Strait of Gibraltar. Journal of Marine Systems 117, 31–42. http://dx.doi.org/ 10.1016/j.jmarsys.2013.02.014. Gascard, J.C., 1991. Open ocean convection and deep water formation revisited in the Mediterranean, Labrador, Greenland and Weddell Seas. In: Chu, P.C., Gascard, J.C. (Eds.), Deep Convection and Deep Water Formation in the Oceans. IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 28
Elsevier Oceanography Series, 57, Amsterdam, pp. 157–182, http://dx.doi.org/ 10.1016/s0422-9894(08)70066-7. Gascard, J.C., Richez, C., 1985. Water masses and circulation in the western Alborán Sea and in the Straits of Gibraltar. Progress in Oceanography 15 (3), 157–216. Helfrich, K.R., 1995. Time-dependent two-layer hydraulic exchange flows. Journal of Physical Oceanography 25 (3), 359–373. Herrmann, M., Somot, S., Sevault, F., Estournel, C., Déqué, M., 2008. Modeling the deep convection in the northwestern Mediterranean Sea using an eddypermitting and an eddy-resolving model: case study of winter 1986–1987. Journal of Geophysical Research 113, C04011. http://dx.doi.org/10.1029/ 2006JC003991. Kinder, T.H., Bryden, H., 1990. Aspiration of deep waters through straits. In: Pratt, L.J. (Ed.), The Physical Oceanography of Sea Straits. Kluwer Acad., Norwell Mass., pp. 295–319, http://dx.doi.org/10.1007/978-94-009-0677-8_14. Kinder, T.H., Parrilla, G., 1987. Yes, some of the Mediterranean outflow does come from great depth. Journal of Geophysical Research 92 (C3), 2901–2906. Krahmann, G., 1997. Saisonale und zwischenjhrliche Variabiltim westlichen Mittelmeer-Analyse historischer Daten. Dissertation, Univ. Kiel, Kiel, Germany, 168 pp. Lascaratos, A., Nittis, K., 1998. A high-resolution three-dimensional numerical study of intermediate water formation in the Levantine Sea. Journal of Geophysical Research 103 (C9), 18497–18511. http://dx.doi.org/10.1029/98JC01196. Leith, C.E., 1968. Diffusion approximation for two-dimensional turbulence. Physics of Fluids 11, 671. http://dx.doi.org/10.1063/1.1691968. Macías, D., Martin, A.P., Garcia-Lafuente, J., Garcia, C.M., Yool, A., Bruno, M., Echevarría, F., 2007. Analysis of mixing and biogeochemical effects induced by tides on the Atlantic–Mediterranean flow in the Strait of Gibraltar through a physical–biological coupled model. Progress in Oceanography 74 (2), 252–272. http://dx.doi.org/10.1016/j.ocean.2007.04.006. Macías, D., Bruno, M., Echevarría, F., Vázquez, A., García, C.M., 2008. Meteorologically-induced mesoscale variability of the North-western Alboran Sea (southern Spain) and related biological patterns. Estuarine, Coastal and Shelf Science 78, 250–266. http://dx.doi.org/10.1016/j.ecss.2007.12.008. Marshall, J., Schott, F., 1999. Open-ocean convection: observations, theory, and models. Reviews of Geophysics 37 (1), 1–64. http://dx.doi.org/10.1029/ 98RG02739. Marshall, J., Hill, C., Perelman, L., Adcroft, A., 1997a. Hydrostatic, quasi-hydrostatic, and nonhydrostatic ocean modeling. Journal of Geophysical Research 102 (C3), 5733–5752. http://dx.doi.org/10.1029/96JC02776. Marshall, J., Adcroft, A., Hill, C., Perelman, L., Heisey, C., 1997b. A finite-volume, incompressible Navier Stokes model for studies of the ocean on parallel computers. Journal of Geophysical Research 102, 5753–5766. http://dx.doi.org/ 10.1029/96JC02775. MEDOC Group, 1969. Observations of formation of deep-water in the Mediterranean Sea. Nature 227, 1037–1040. Mikolajewicz, U., 2011. Modeling Mediterranean Ocean climate of the last glacial maximum. Climate of the Past 7, 161–180. http://dx.doi.org/10.5194/cp-7-1612011. Naranjo, C., Garcia-Lafuente, J., Sanchez-Garrido, J.C., Sanchez-Roman, A., Delgado Cabello, J., 2012. The western Alborán gyre helps ventilate the western Mediterranean deep water through Gibraltar. Deep Sea Research Part I: Oceanographic Research Papers 63, 157–163. http://dx.doi.org/10.1016/ j.dsr.2011.10.003. Oddo, P., Adani, M., Pinardi, N., Fratianni, C., Tonani, M., Pettenuzzo, D., 2009. A nested Atlantic–Mediterranean Sea general circulation model for operational forecasting. Ocean Science Discussions 5 (461–473), 2009. http://dx.doi.org/ 10.5194/os-5-461-2009. Pacanowski, R.C., Philander, S.G.H., 1981. Parameterization of vertical mixing in numerical models of tropical oceans. Journal of Physical Oceanography 11 (11), 1443–1451. Pinardi, N., Zavatarelli, M., Adani, M., Coppini, G., Fratianni, C., Oddo, P., Bonaduce, A., 2013. Mediterranean Sea large-scale low-frequency ocean variability and water mass formation rates from 1987 to 2007: a retrospective analysis. Progress in Oceanography. http://dx.doi.org/10.1016/j.pocean.2013.11.003 (in press). Sammartino, S., Garcia-Lafuente, J., Sanchez-Garrido, J.C., De los Santos, F.J., ÁlvarezFanjul, E., Naranjo, C., Bruno, M., Calero, C., 2014. A numerical model analysis of the tidal flows in the Bay of Algeciras, Strait of Gibraltar. Continental Shelf Research 72, 34–46. http://dx.doi.org/10.1016/j.csr.2013.11.002. Sanchez-Garrido, J.C., Sannino, G., Liberti, L., Garcia-Lafuente, J., Pratt, L., 2011. Numerical modeling of three-dimensional stratified tidal flow over Camarinal Sill, Strait of Gibraltar. Journal of Geophysical Research 116, C12026. http:// dx.doi.org/10.1029/2011JC007093. Sanchez-Garrido, J.C., Garcia-Lafuente, J., Álvarez-Fanjul, E., Sotillo, M., de-losSantos, F.J., 2013. What does cause the collapse of the Western Alborán Gyre? Results of an operational ocean model. Progress in Oceanography 116, 142–153. http://dx.doi.org/10.1016/j.pocean.2013.07.002. Sanchez-Román, A., Sannino, G., Garcia-Lafuente, J., Carillo, A., Criado-Aldeanueva, F., 2009. Transport estimates at the western section of the Strait of Gibraltar: a combined experimental and numerical modeling study. Journal of Geophysical Research 114, C06002. http://dx.doi.org/10.1029/2008JC005023. Sannino, G., Bargagli, A., Artale, V., 2004. Numerical modeling of the semidiurnal tidal exchange through the Strait of Gibraltar. Journal of Geophysical Research 109, C050115. http://dx.doi.org/10.1029/2003JC002057. Sannino, G., Carillo, A., Artale, V., 2007. Three-layer view of transports and hydraulics in the Strait of Gibraltar: a three-dimensional model study. Journal of Geophysical Research 112, C03010. http://dx.doi.org/10.1029/2006JC003717. Sannino, G., Sanchez Garrido, J.C., Liberti, L., Pratt, L., 2014. Exchange flow through the Strait of Gibraltar as simulated by a r -coordinate hydrostatic model and a z-coordinate nonhydrostatic model. In: Borzelli, G.L.E., Gac ˇic ´, M., Lionello, P., Malanotte-Rizzoli, P. (Eds.), The Mediterranean Sea: Temporal Variability and Spatial Patterns. John Wiley & Sons Inc., Oxford. http://dx.doi.org/10.1002/ 9781118847572.ch3. Schott, F., Visbeck, M., Send, U., 1994. Open ocean deep convection, Mediterranean and Greenland Seas. In: Malanotte-Rizzoli, P., Robinson, A.R. (Eds.), Ocean Processes on Climate Dynamics: Global and Mediterranean Examples. Kluwer Acad., Norwell, Mass., pp. 203–225. Schott, F., Visbeck, M., Send, U., Fischer, J., Stramma, L., Desaubies, Y., 1996. Observations of deep convection in the Gulf of Lions, northern Mediterranean, during the winter of 1991/92. Journal of Physical Oceanography 26 (4), 505– 524. http://dx.doi.org/10.1175/1520-0485(1996) 026<0505:OODCIT>2.0.CO;2. Schroeder, K., Ribotti, A., Borghini, M., Sorgente, R., Perilli, A., Gasparini, G.P., 2008. An extensive western Mediterranean deep water renewal between 2004 and 2006. Geophysical Research Letters 35, L18605. http://dx.doi.org/10.1029/ 2008GL035146. Send, U., Schott, F., Gaillard, F., Desaubies, Y., 1995. Observation of a deep convection regime with acoustic tomography. Journal of Geophysical Research 100 (C4), 6927–6941. http://dx.doi.org/10.1029/94JC03311. Smith, R.O., Bryden, H.L., Stansfield, K., 2008. Observations of new western Mediterranean deep water formation using Argo floats 2004–2006. Ocean Science 4 (2), 133–149. Stommel, H., Bryden, H., Mangelsdorf, P., 1973. Does some of the Mediterranean outflow come from great depth? Pure and Applied Geophysics 105 (1), 879– 889. Tziperman, E., Speer, K., 1994. A study of water mass transformation in the Mediterranean Sea: analysis of climatological data and a simple three-box model. Dynamics of Atmospheres and Oceans 21, 53–82. http://dx.doi.org/ 10.1016/0377-0265(94)90004-3. Vargas, J.M., Garcia-Lafuente, J., Candela, J., Sanchez, A.J., 2006. Fortnightly and monthly variability of the exchange through the Strait of Gibraltar. Progress in Oceanography 70 (2), 466–485. http://dx.doi.org/10.1016/j.pocean. 2006.07.001. Vázquez, A., Bruno, M., Izquierdo, A., Macías, D., Ruiz-Cañavate, A., 2008. Meteorologically forced subinertial flows and internal wave generation at the main sill of the Strait of Gibraltar. Deep Sea Research Part I: Oceanographic Research Papers 55 (10), 1277–1283. http://dx.doi.org/10.1016/j.dsr. 2008.05.008. Vázquez, A., Flecha, S., Bruno, M., Macías, D., Navarro, G., 2009. Internal waves and short-scale distribution patterns of chlorophyll in the Strait of Gibraltar and Alborán Sea. Geophysical Research Letters 36 (23), L23601. http://dx.doi.org/ 10.1029/2009GL040959. Wang, D.P., 1993. The Strait of Gibraltar Model: internal tide, diurnal inequality and fortnightly modulation. Deep Sea Research Part I: Oceanographic Research Papers 40 (6), 1187–1203. http://dx.doi.org/10.1016/0967-0637(93)90133-N. Wesson, J.C., Gregg, M.C., 1994. Mixing at Camarinal sill in the Strait of Gibraltar. Journal of Geophysical Research: Oceans 99 (5), 9847–9878. http://dx.doi.org/ 10.1029/94JC00256 (1978–2012). Whitehead Jr., J.A., 1985. A laboratory study of gyres and uplift near the Strait of Gibraltar. Journal of Geophysical Research 90 (C4), 7045–7060. http:// dx.doi.org/10.1029/JC090iC04p07045. IV. How much do tides affect the circulation of the Mediterranean Sea? From local processes in the SoG to basin-scale effects 29
V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea Jos´e C. S´anchez-Garrido1, C. Naranjo,1, D. Mac´ıas,2, J. Garc´ıa-Lafuente1, and T. Oguz3 Corresponding author: J. C. S´anchez-Garrido, Physical Oceanography Group, University of M´alaga, ETSI Telecomunicaci´on, 29071, M´alaga, Spain. (jcsanc[email protected]) 1Physical Oceanography Group, University of M´alaga, M´alaga, Spain. 2European Commission, Joint Research Center, Institute for Environment and Sustainability, Water Research Unit, Ispra, Italy. 3Institute of Marine Sciences Middle East Technical University, Erdemli, Turkey. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 31
Abstract. The control of phytoplankton production by tidal forcing in the Alboran Sea is investigated with a high-resolution ocean circulation model coupled to an ecosystem model. The aim of the modeling efforts was to elucidate the role of tides in sustaining the high biological productivity of the Alboran Sea, as compared with the rest of the Mediterranean sub-basins. It is shown that tidal forcing accounts for an increase of phytoplankton biomass and primary productivity in the basin of about 40% with respect to a non-tidal circulation, and about 60% in the western Alboran Sea alone. The tidal dynamics of the Strait of Gibraltar is shown to be the primary factor in determining the enhancement of productivity, pumping nutrients from depth to the photic zone in the Alboran Sea. Model results indicate that the biological implications of the propagating internal tides are small. These results imply that nutrient transports through the Strait of Gibraltar have to be parametrized in ocean models that do not resolve tides in order to properly represent the biochemical budgets of the Alboran Sea. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 32
1. Introduction The Alboran Sea (AS; Fig. 1a) is the first sub-basin that the jet of Atlantic Water (AW) entering through the Strait of Gibraltar (SoG) encounters in its way along the Mediterranean. Its surface circulation is fairly variable in both space and time, but its most typical and classically described configuration is that consisting of the Atlantic jet meandering around and feeding two mesoscale anticyclonic gyres, the so-called Western and Eastern Alboran Gyres (hereinafter WAG and EAG). Typical velocities along the jet are as large as 1-2 m s−1, and to a first approximation can be assumed to be geostrophically maintained by the strong density front between the relatively fresh incoming AW and the saltier ambient water (modified AW). As elsewhere in the ocean, such fronts feature a secondary cross-front ageostrophic circulation characterized by large vertical velocities and with potential to sustain high levels of biological productivity [e.g., Spall, 1995; Nagai et al., 2008]. In biological terms, the AS is one of the most (if not the most) productive sub-basins of the Mediterranean [e.g., Uitz et al., 2012], and there are a number of physical reasons for that. The most obvious one lies in its active frontal activity and, not surprisingly, field observations and ocean color images reveal enhanced levels of chlorophyll (chl) concentration along the jet [Navarro et al., 2011]. The potential of the frontal jet ageostrophic circulation to sustain phytoplankton production has been further confirmed in the recent process-oriented model study by Oguz et al. [2014]. Another important factor is the wind-driven circulation, particularly that forced by westerlies that frequently drive strong V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 33
by ambient nutrient, light, and temperatures [Geider et al., 1997]. A perfect matching between model and remote sensing observations is not then to be expected, although general spatial patterns should be similar. 3. Experimental procedure In a first experiment the model is run for three years without tidal forcing and with the rest of the forcing fields corresponding to the period 2011-2013. The first year of simulation (2011) is taken as spin-up time and is therefore excluded from the analysis. The aim of this run is to provide a basic circulation and ecosystem state on which later investigate the role of tides on the ecosystem. It will be referred as the reference experiment (REF). In a second run the model includes tides (TID experiment) and their biological implications are evaluated. This run covers only three months, a sub-period of the REF run. The reason to conduct a shorter run is largely motivated by the fact that tidal forcing makes the model substantially more expensive in computational terms. On the other hand, a period of three months that encompasses several spring-neap tidal cycles (this cycle is the main source of low-frequency tidal variability) is considered suitable to provide a good estimation of the effect of tides on biology. 4. Results 4.1. REF experiment The results of the REF run are examined in order to assess the ability of the model to reproduce the general circulation and basic biological patterns of the AS. Likewise other regional models that do not incorporate tides nor explicitly resolve the exchange flow V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 40
through the SoG are successful in capturing such basic features [e.g. Lazzari et al., 2012], thus great discrepancies between model and observations are not expected. The mean sea surface height (SSH) of the simulation (black contours in Fig. 2a) reveals the presence of both gyres, the WAG and EAG. The WAG has larger size and features greater geostrophic currents than its eastern counterpart, as noted by the sharper crossgyre SSH gradients. This result stems partially from the fact that the WAG is a rather permanent feature during the simulation, while the EAG exhibits more variability in both size and location (sometimes it is even absent). Figure 2a also displays the twoyear climatology of the surface chl and reflects the oligotrophic nature of the gyres. As expected chl concentration is high in the vicinity of the jet that runs eastwards with its axis centered along the SSH=0 contour (thick solid line). This is particularly true for the cyclonic side of the frontal jet (to the north of the AS), a pattern that is consistent with a cross-front ageostrophic circulation causing downwelling at the anticyclonic side of the front, and upwelling in its cyclonic side. This creates an overall meridional gradient of chl with increasing concentration to the north of the basin. Additionally, there is a decrease of chl from west to east with the richest area being the northwestern region of the AS (NWAS), in agreement with field and remote sensing observations [see, e.g., Sarhan et al., 2000; Reul et al., 2005; Mac´ıas et al., 2007a]. Oguz et al. [2014] attribute the eastwards decreasing pattern of chl to the gradual weakening of the density front, and thereby of its potential to fuel phytoplankton production, by accumulative mixing between incoming and ambient waters. The spatial pattern of the surface chl in the REF run has been compared with remotely sensed chl obtained from the GlobColour Project (http://www.globcolour.info/). V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 41
They produce ocean color maps (Level-3) by merging data from the sensors SeaWiFS, MODIS and MERIS. The obtained chl is for case 1 waters where phytoplankton concentration prevails over inorganic particles and the applied algorithm is based on weighted averaging merging techniques. The spatial resolution of the downloaded composite images is 1/48◦and consist of 8-day merged fields. The climatology computed during the analyzed period (Fig. 2b) displays the main features reproduced by the REF run, including the oligotrophic gyres and the biomass-rich NWAS, although with somewhat different characteristics. For instance, the chl-rich region of the NWAS is larger than in the model. Moreover, chl concentration in the WAG is considerably higher than in the EAG, both in its core and specially at the outer edges, and the west-to-east decreasing pattern of chl appears more marked than in the model. Another noticeable difference is observed over the continental shelf to the north west of the SoG, where the satellite image shows the greatest chl concentration of the domain (more than 1 mg m−3), whereas in the model this area is not particularly rich. These discrepancies are to a great extent due to tides, as we shall see later. The time variability of the model and satellite chl has been also compared by computing spatial aggregate median values in the basin (Fig. 3). For the computation the model outputs were interpolated onto the satellite data grid and the cloud cover mask was applied to the interpolated chl for a more reliable comparison of the two data sets. Satellite chl (box-plot) exhibits a seasonal cycle with minimum values of chl by mid summer and maximum by the beginning of March and reflects the two regimes identified by Garc´ıaGorriz and Carr [1999], with a fall-to-spring bloom (November-March) and a non-bloom period (May-September). The model (red line) captures the timing and the duration of V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 42
these regimes, the agreement being better in year 2013. Less satisfactory is the agreement between the timing of wind-driven bloom events due to westerlies that induce coastal upwelling along the north coast of the AS. An example is the bloom occurring in the model by July 2012 (see zonal wind stress in Fig. 4b) that is not reflected in the observations. Conversely, there are some events of remarkably high chl concentration in the satellite records that are not captured by the model and could be associated with coastal upwelling as well, like the chl peak observed by the end of May 2013. There is then a mismatch in the timing of these blooms that could be ascribed to the coarse resolution of the forcing winds. Nonetheless, this is not of particular concern for the results discussed in this paper that focus on tidal processes. Another noteworthy characteristic of the AS (and also of the rest of the Mediterranean) is the presence of a permanent deep chl maximum (DCM) with the possible exception of the late winter [Siokou-Frangou et al., 2010]. Its mean depth is about 30 m and it is shallower than in the rest of the Mediterranean, presumably because of the higher productivity and hence lower seawater transparency of the AS. In the model the maximum concentration of phytoplankton biomass is found at z=−34 ±25 m (Fig. 4a), it deepens during summer, and shoals towards the late winter, which is in agreement with the mentioned behaviour of the DCM. The most dramatic vertical excursions of the DCM take place, however, during the mentioned upwelling events when, regardless the season the chl maximum reaches the very surface. According to the model this occurs by mid July 2012 and early September 2013 (Fig. 4a). The last variable examined has been the depth-integrated PP (PPint). In situ measurements reflects the great variability of the AS in both space and time with estimated PPint V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 43
ranging in a broad interval, typically between 100-300 g C m−2y−1[see Siokou-Frangou et al., 2010, and the references therein]. Such variability has been largely attributed to the presence of the frontal jet in which a PPint of up to 475 g C m−2y−1has been reported [1300 mg C m−2d−1;Lohrenz et al., 1988]. The basin-averaged PPint in the REF run (136.1±24.6 g C m−2y−1) is in the lower range of the reported values, and, as the DCM, exhibits both seasonal and wind-driven variability (Fig. 4b). 4.2. TID experiment In the TID experiment the model was integrated with tidal forcing from February to June 2013 and starting from REF outputs as initial conditions. Approximately after one month of simulation the differences between the biochemical variables diagnosed in the REF and TID runs, including PP, remained stable over time, indicating that the ecosystem adjusted to the new hydrodynamic conditions after that period. The first 45 days of model outputs were then ruled out and so the final TID data set covered the last 2.5 months of the experiment (from mid March to the beginning of June). It was found that tides increased the productivity of the AS by about 40%, from a mean value of 126.6 g C m−2y−1in REF to 176.7 g C m−2y−1in TID (Fig. 5a,b), this later value more in the range of reported estimates. The increase was not uniform over the basin, the ecosystem of its western half being particularly affected by tides. In the area between the SoG and 3◦W, encompassing the NWAS and the WAG, the PP increment with respect to REF was 60% (from 141.4 g C m−2y−1in REF to 226.1 g C m−2y−1in TID) whereas in the eastern AS (east of 3◦W) it was only 11% (from 111.1 to 124.1 g C m−2y−1). Zones that become particularly productive with tides are the NWAS, where PPint locally exceeds 500 g C m−2y−1, and also a small area of the continental shelf northwest of the strait in which V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 44
the productivity is largely tidally driven (more than twice the REF value). Phytoplankton biomass increases approximately in the same proportion (Fig. 5c,d). Regarding the spatial pattern of chl displayed in satellite images, the inclusion of tides in the TID run results in a spatial distribution that agrees with observations much better than the REF run. Of particular concern is the west-to-east decreasing pattern of remote sensing images (Fig. 6c) that is satisfactorily reproduced by the TID run (Fig. 6b) but not by the REF run (Fig. 6a). The TID run also gives rise to a localized spot of high chl concentration in the aforementioned region of the continental shelf northwest to the SoG (Fig. 6b), although it is still unable to reproduce the high chl signal along the Spanish coast displayed in the satellite image (Fig. 6c), a fact that we attribute to nutrient loads from the Guadalquivir river that were not prescribed in the model. 5. Analysis Here we analyze the candidate physical processes for explaining the tidally driven increase of the AS productivity. First we assess the hypothesis that tidal flows in the SoG fertilize the Atlantic inflow. The tidal dynamic of the strait is examined and nutrient transports into the AS will be computed for both the REF and TID simulations. Attention will be also paid to other tidal processes with possible biological consequences such as propagating internal waves. 5.1. Tidal dynamics and nutrient transports through the SoG The configuration of the flow through the SoG as simulated in the REF run is shown in Fig. 7, displaying salinity, velocity, and nutrient concentration (nitrate) along the strait. The flow consists of an undercurrent of nutrient-rich and salty MW (>8 mmol NO3m−3; V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 45
S > 37) flowing westwards, and AW flowing to the AS (S < 37), which can be further partitioned into the nutrient-depleted Surface Atlantic Water (SAW; <3 mmol NO3m−3) and the moderately nutrient-rich North Atlantic Central Water (NACW; ≈5−7 mmol NO3m−3). The NACW is fresher but denser than the overlying SAW lies to the west of the Camarinal Sill at 150-200 m depth, with a salinity of S≈36.2−36.4. With this configuration the nutrient budget of the inflow is largely dependent upon the volume of NACW that can overpass the sill and participate in the inflow, as well as on the ability of the Atlantic current to entrain nutrients from the underlying MW. Figure 7 suggests that the second mechanism is at work 30 km to the east of the sill, in the so-called Tarifa Narrows (TN in inset of Fig. 1b), where the inflow accelerates in response to the constricted lateral boundaries. The steady exchange in the REF run is notably modified in the TID run as it can be seen in the time evolution of the flow during a tidal cycle presented in Fig. 8. Figure 8a roughly corresponds to the time of maximum tidal flow towards the Atlantic (flood tide; see barotropic velocity over the sill in the bottom panel) and reflects important differences with respect to the steady flow configuration described above. The response of the flow to the tidal discharge is the expected for an hydraulically controlled sill flow exposed to a quasi-steady barotropic forcing [Long, 1954; Lawrence, 1993] both upstream (to the right) and downstream (left) of the bottom obstacle. In the upstream side the flow adjusts to the tidal forcing by raising the interface between Atlantic and Mediterranean Waters, in a clear manifestation of upstream influence exerted by the sill, whereas the most noticeable downstream effect is the enhancement of vertical mixing, as noted by the homogenization of both salinity and nutrient concentration over the water column. Such enhanced of V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 46
mixing is closely related to the increase of the internal Froude number at the downstream side of the sill [S´anchez-Garrido et al., 2011]. The biological implications of the tides are revealed in the subsequent stages of the tidal cycle, when the tidal flow turns to the east (ebb tide; Fig. 8b,c). Tidal forcing is sufficiently strong to invert the direction of the Mediterranean current over the sill, temporally facilitating the eastwards advection of both NACW and MW, the later returning to the eastern flank of the sill with the tidal reversal (note positive velocities all over the sill crest in Fig. 8c). In fact, Garc´ıa-Lafuente et al. [2013] estimate that as much as 30% of the pool of mixed water formed west of the sill during the flood tide, a mixture of NACW and MW, can be subsequently pumped to the near surface in the eastern part of the strait, providing a nutrient supply to the AS. A second way in which tides can fertilize the inflowing waters is through the enhancement of diapycnal mixing to the east of the sill, driven in part by the eastwards propagation of the internal tidal bore generated at the sill (note it, for instance, at x=22 km in Fig. 8c). The strong time dependence of the flow is transferred to the nutrient budget of the inflowing waters, as noticed in Fig. 9 that shows the time-space dependence of nitrate integrated in the uppermost 94 m of the water column. The maximum nutrient content is periodically reached in Camarinal Sill (x=0) around the end of the flood tide as a result of the shoaling of the interface (see Fig. 8a), and moves eastwards during the ebb tide at a velocity that can be estimated in 1.5 m s−1from the slope of the contours in the time-space diagram of Fig. 9a. Garc´ıa-Lafuente et al. [2013] described a similar eastwards propagation pattern of the thickness of the interface mixing layer and ascribed it to the physical processes mentioned above, namely, advection of mixed water formed V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 47
west of the sill and enhanced entrainment of MW driven by the propagating baroclinic bore. The same processes are very likely involved in the eastward propagation of nutrients from the sill with tidal modulation. This behaviour differs from the REF run, in which the nutrient budget of the inflow hardly exhibits time dependence. It is from the Tarifa Narrows that it increases eastwards (Fig. 9b), due to the enhanced entrainment of MW driven by the spatial acceleration of the inflow in this location, as remarked earlier. Most importantly, the nutrient content of the inflow at the eastern mouth of the SoG (x=45 km) is greater in the TID experiment, oscillating between 350-550 mmol NO3m−2, against the more constant value of 200-250 mmol NO3m−2obtained in REF, which results in different nutrient transports into the AS in the two model simulations (Fig. 10). The nitrate transport in TID (Fig. 10a; gray line) has a mean ±std value of 2.28±1.28 kmol s−1, nearly 75% greater than in REF (1.31±0.28 kmol s−1; black dashed line). The greater variability of the TID series is largely accounted for the M2tidal constituent, although a noticeable fortnightly signal can be also distinguished, with greater and smaller transports during spring and neap tides, respectively, as highlighted by the low-pass curve (tidal variability filtered out; black solid line). The same applies to phosphate, although differences between the two simulations are slightly more moderate (0.20±0.10 and 0.15± 0.03 kmol s−1in TID and REF, respectively; Fig. 10b). Note that these discrepancies are of particular concern because the input of nutrients provided by tides is within the photic zone and so can be utilized by phytoplankton. This becomes even more obvious by noting that the nutrients transport in TID are even greater with respect to those in REF when computed in the upper part of the photic zone (−48 < z < 0; Fig. 10c,d), where high levels of light can further stimulate phytoplankton growth. The tidal pumping V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 48
of nutrients from the sill is therefore a clear candidate to explain the increase of biomass and productivity obtained in the TID simulation. 5.2. Tides and tidal-related processes in the AS It is also feasible that the different biological patterns of the TID and REF runs are accounted for, at least partially, by local tidal currents in the AS, rather than in the adjacent SoG. For instance, a distinctive feature of the tidal experiment is the internal tide that progresses eastwards through the SoG into the AS after being generated in the Camarinal Sill (Fig. 8). Such internal waves induce large vertical velocities (Fig. 11) that can potentially modify phytoplankton growth by exposing it to different levels of irradiance through vertical advection. Another tidal-related process that bears consideration refers to lateral boundary flows. S´anchez-Garrido et al. [2013] found that the back and forth motion of tides generate shear vorticity on the strait’s lateral boundary layers (positive on the north, negative on the south) that can lead to a wake of coherent submesoscale eddies carried by the Atlantic jet [see also La-Violette, 1984]. The flow along the wake is quite ageostrophic and can stimulate productivity, much like observed in island wakes [Hammer and Hauri, 1981; Hern´andez-Le´on, 1991; Hasegawa et al., 2004]. In order to clarify whether these processes are biologically relevant, a third experiment has been conducted in which the physics is as in the REF run but with modified ecosystem equations so that biochemical tracers are nudged towards their TID values within the SoG. More especially, a Newtonian nudging term τ−1(B−Btid) was added to the right hand side of eq. 1, where Btid denotes the value of the biochemical tracer obtained in the tidal simulation (daily mean fields), and τis the relaxation time scale. For a sufficiently small τthe nudging term acts more like a forcing rather than a relaxation term, forcing the V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 49
Table 1 displays the values given to all described parameters together with the most relevant ones involved in zooplankton and detritus equations; further details can be found in the referred papers. Acknowledgments. The data of this work are available upon request to the corresponding author (jcsanc[email protected]). This study was funded by Consejer´ıa de Econom´ıa, Innovaci´on, Ciencia y Empleo (Junta de Andaluc´ıa, Spain) under the Research Project MOCBASE (RNM-1540). JCSG was supported by a Juan de la Cierva Postdoctoral Grant (JCI-2012-13451) from the Spanish Ministry of Economy and Competitiveness. CN acknowledges a postgraduate fellowship (BES-2011-043421) from the same Ministry. The model calculations were conducted in the facilities of the Supercomputing and Bioinnovation Center of the University of M´alaga. We thank the ESA for distributing the GlobColour data used in this paper. NCEP Reanalysis data were provided by the NOAA/OAR/ESRLPSD, Boulder, Colorado, USA. ASCAT data were obtained from the Centre de Recherche et d’Exploitation Satellitaire (CERSAT), at IFREMER, Plouzan´e (France). References Brandt, P., W. Alpers, and J. O. Backhaus (1996), Study of the generation and propagation of internal waves in the strait of gibraltar using a numerical model and synthetic aperture radar images of the european ers 1 satellite, J. Geophys. Res.,101(C6), 14,237– 14,252. Dutkiewicz, S., M. J. Follows, and J. G. Bragg (2009), Modeling the coupling of ocean ecology and biogeochemistry, Global Biogeochem. Cycles,23(4), n/a–n/a, doi: V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 56
10.1029/2008GB003405. Echevarr´ıa, F., L. Zabala, A. Corzo, G. N. L. Prieto, and D. Mac´ıas (2009), Spatial distribution of autotrophic picoplankton in relation to physical forcings: the gulf of C´adiz, Strait of Gibraltar and Albor´an Sea case study, J. Plankton Res.,31(11), 1339– 1351. Evans, M. A., S. MacIntyre, and G. W. Kling (2008), ”internal wave effects on photosynthesis: Experiments, theory, and modeling ”, Limnol. Oceanogr.,53(1), 339–353. Follows, M. J., S. Dutkiewicz, S. Grant, and S. W. Chisholm (2007), Emergent biogeography of microbial communities in a model ocean, Science,315(5820), 1843–1846, doi:10.1126/science.1138544. Garc´ıa-Gorriz, E., and M.-E. Carr (1999), The climatological annual cycle of satellitederived phytoplankton pigments in the Alboran Sea, Geophys. Res. Lett.,26(19), 2985– 2988, doi:10.1029/1999GL900529. Garc´ıa-Lafuente, J., E. Bruque, J. C. S´anchez-Garrido, G. Sannino, and S. Sammartino (2013), The interface mixing layer and the tidal dynamics at the eastern part of the Strait of Gibraltar, J. Marine Syst.,117-118(0), 31–42, doi: http://dx.doi.org/10.1016/j.jmarsys.2013.02.014. Geider, R. J., H. L. McIntyre, and T. M. Kana (1997), Dynamic model of phytoplankton growth and acclimation: responses of the balanced growth rate and the chlorophyll a: carbon ratio to light, nutrient-limitation and temperature, Mar. Ecol. Prog. Ser.,148, 187–200. Griffies, S. M., and R. W. Hallberg (2000), Biharmonic friction with a smagorinsky-like viscosity for use in large-scale eddy-permitting ocean models, Mon. Wea. Rev.,128, V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 57
2935–2946. Hammer, W. M., and I. R. Hauri (1981), Effects of island mass: Water flow and plankton pattern around a reef in the great barrier reef lagoon, Australia, Limnol. Oceanogr.,26, 1084–1102. Hasegawa, D., H. Yamazaki, R. G. Lueck, and L. Seuront (2004), How islands stir and fertilize the upper ocean, Geophys. Res. Lett.,31(16), n/a–n/a, doi: 10.1029/2004GL020143. Hern´andez-Le´on, S. (1991), Accumulation of mesozooplankton in a wake area as a causative mechanism of the island-mass effect, Mar. Biol.,109(1), 141–147, doi: 10.1007/BF01320241. Holloway, G., and K. Denman (1989), Influence of internal waves on primary production, J. Plankton Res.,11(2), 409–413. Huertas, I. E., A. F. R´ıos, J. Garc´ıa-Lafuente, A. Makaoui, S. Rodr´ıguez-G´alvez, A. S´anchez-Rom´an, A. Orbi, J. Ru´ız, and F. F. P´erez (2009), Anthropogenic and natural co2exchange through the Strait of Gibraltar, Biogeosciences,6(4), 647–662, doi:10.5194/bg-6-647-2009. Hundsdorfer, W., and R. A. Trompert (1994), Method of lines and direct discretization: A comparison for linear advection, Appl. Numer. Math.,13(6), 469–490, doi: http://dx.doi.org/10.1016/0168-9274(94)90009-4. Kalnay (1996), The NCEP/NCAR 40-year reanalysis project, Bull. Amer. Meteor. Soc., 77, 437–470. La-Violette, P. E. (1984), The advection of submesoscale thermal features in the Alboran Sea Gyre, J. Phys. Oceanogr.,14, 550–565. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 58
Lawrence, G. A. (1993), The hydraulics of steady two-layer flow over a fixed obstacle, J. Fluid Mech.,254, 605–633. Lazzari, P., C. Solidoro, V. Ibello, S. Salon, A. Teruzzi, K. B´eranger, S. Colella, and A. Crise (2012), Seasonal and inter-annual variability of plankton chlorophyll and primary production in the Mediterranean Sea: a modelling approach, Biogeosciences,9(1), 217–233, doi:10.5194/bgd-8-5379-2011. Leichter, J. J., G. Shellenbarger, S. J. Genovese, and S. R. Wing (1998), Breaking internal waves on a Florida (USA) coral reef: A plankton pump at work?, Mar. Ecol. Prog. Ser., 166, 83–97. Lohrenz, S. E., D. A. Wiesenburg, I. P. DePalma, K. S. Johnson, and D. E. Gustafson (1988), Interrelationships among primary production, chlorophyll, and environmental conditions in frontal regions of the western Mediterranean Sea, Deep-Sea Res. Oceanogr., A,35(5), 793 – 810, doi:http://dx.doi.org/10.1016/0198-0149(88)90031-3. Long, R. R. (1954), Some aspects of the flow of stratified fluids. ii. experiments with a two-fluid system, J. Fluid Mech.,6, 97–115. Losch, M., V. Strass, B. Cisewski, C. Klaas, and R. G. Bellerby (2014), Ocean state estimation from hydrography and velocity observations during eifex with a regional biogeochemical ocean circulation model, J. Marine Syst.,129(0), 437–451, doi: http://dx.doi.org/10.1016/j.jmarsys.2013.09.003. Mac´ıas, D., G. Navarro, F. Echevarr´ıa, C. M. Garc´ıa, and J. L. Cueto (2007a), Phytoplankton pigment distribution in the northwestern Alboran Sea and meteorological forcing: A remote sensing study, J. Mar. Res.,65(4), 523–543, doi: doi:10.1357/002224007782689085. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 59
Mac´ıas, D., A. Martin, J. Garci´a-Lafuente, C. Garc´ıa, A. Yool, M. Bruno, A. V´azquezEscobar, A. Izquierdo, D. Sein, and F. Echevarr´ıa (2007b), Analysis of mixing and biogeochemical effects induced by tides on the AtlanticMediterranean flow in the Strait of Gibraltar through a physicalbiological coupled model, Prog. Ocenogr.,74(2-3), 252 – 272. Maillard, C. (2002), MEDAR/MEDATLAS: a database for Mediterranean and Black Sea projects, 16, 115-117 pp., CIESM, IFREMER/SISMER, Centre de Brest, B.P.70, 29280 Plouzane France. Marshall, J., C. Hill, L. Perelman, and A. Adcroft (1997a), Hydrostatic, quasi-hydrostatic, and nonhydrostatic ocean modeling, J. Geophys. Res.,102(C3), 5733–5752, doi: 10.1029/96JC02776. Marshall, J., A. Adcroft, C. Hill, L. Perelman, and C. Heisey (1997b), A finite-volume, incompressible Navier Stokes model for studies of the ocean on parallel computers, J. Geophys. Res.,102(C3), 5753–5766, doi:10.1029/96JC02775. Muacho, S., J. da Silva, V. Brotas, and P. Oliveira (2013), Effect of internal waves on nearsurface chlorophyll concentration and primary production in the nazar canyon (west of the iberian peninsula), Deep Sea Research Part I: Oceanographic Research Papers, 81(0), 89–96. Nagai, T., A. Tandon, N. Gruber, and J. C. McWilliams (2008), Biological and physical impacts of ageostrophic frontal circulations driven by confluent flow and vertical mixing, Dyn. Atmos. Oceans.,45(3-4), 229 – 251. Navarro, G., A. V´azquez, D. Mac´ıas, M. Bruno, and J. Ru´ız (2011), Understanding the patterns of biological response to physical forcing in the Alboran Sea (western MediterV. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 60
ranean), Geophys. Res. Lett.,38(23), n/a–n/a, doi:10.1029/2011GL049708. Oguz, T., D. Mac´ıas, J. Garc´ıa-Lafuente, A. Pascual, and J. Tintor´e (2014), Fueling plankton production by the meandering frontal jet in the Alboran Sea (Western Mediterranean), PloS ONE,9(11), e111,482. Pacanowski, R. C., and S. G. H. Philander (1981), Parameterization of vertical mixing in numerical models of tropical oceans, J. Phys. Oceanogr.,11, 1443–1451. Pan, X., G. Wong, F.-K. Shiah, and T.-Y. Ho (2012), Enhancement of biological productivity by internal waves: observations in the summertime in the northern south china sea, Journal of Oceanography,68(3), 427–437. Ram´ırez, T., D. Cort´es, J. Mercado, M. Vargas-Yaez, M. Sebasti´an, and E. Liger (2005), Seasonal dynamics of inorganic nutrients and phytoplankton biomass in the {NW} alboran sea, Est. Coast. Shelf Sci.,65(4), 654–670. Reul, A., V. Rodr´ıguez, F. Jim´enez-G´omez, J. Blanco, B. Bautista, T. Sarhan, F. Guerrero, J. Ru´ız, and J. Garc´ıa-Lafuente (2005), Variability in the spatio-temporal distribution and size-structure of phytoplankton across an upwelling area in the NW-Alboran Sea, (W-Mediterranean), Cont. Shelf Res.,25 (5-6), 589 – 608, doi: http://dx.doi.org/10.1016/j.csr.2004.09.016. S´anchez-Garrido, J. C., G. Sannino, L. Liberti, J. Garc´ıa Lafuente, and L. Pratt (2011), Numerical modeling of three-dimensional stratified tidal flow over Camarinal Sill, Strait of Gibraltar, J. Geophys. Res.,116(C12), n/a–n/a, doi:10.1029/2011JC007093. S´anchez-Garrido, J. C., J. Garc´ıa-Lafuente, E. ´ Alvarez-Fanjul, M. Garc´ıa-Sotillo, and F. J. de los Santos (2013), What does cause the collapse of the Western Alboran Gyre? Results of an operational ocean model, Prog. Ocenogr.,116(0), 142–153, doi: V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 61
http://dx.doi.org/10.1016/j.pocean.2013.07.002. Sangr`a, P., G. Basterretxea, J. L. Pelegr´ı, and J. Ar´ıstegui (2001), Chlorophyll increase due to internal waves in the shelf-break of Gran Canaria Island (Canary Islands), Sci. Mar.,65, 89–97. Sannino, G., J. C. S´anchez Garrido, L. Liberti, and L. Pratt (2014), Exchange flow through the Strait of Gibraltar as simulated by a sigma-coordinate hydrostatic model and a zcoordinate non-hydrostatic model, in The Mediterranean Sea: Temporal Variability and Spatial Patterns, pp. 25–50, John Wiley and Sons, Inc., doi:10.1002/9781118847572.ch3. Sarhan, T., J. Garc´ıa-Lafuente, M. Vargas, J. M. Vargas, and F. Plaza (2000), Upwelling mechanisms in the northwestern Alboran Sea, J. Marine Syst.,23(4), 317–331, doi: http://dx.doi.org/10.1016/S0924-7963(99)00068-8. Siokou-Frangou, I., U. Christaki, M. G. Mazzocchi, M. Montresor, M. Ribera d’Alcal´a, D. Vaqu´e, and A. Zingone (2010), Plankton in the open Mediterranean Sea: a review, Biogeosciences,7(5), 1543–1586, doi:10.5194/bg-7-1543-2010. Spall, M. A. (1995), Frontogenesis, subduction, and cross-front exchange at upper ocean fronts, J. Geophys. Res.,100(C2), 2543–2557. Steinacher, M., F. Joos, T. L. Fr¨olicher, L. Bopp, P. Cadule, V. Cocco, S. C. Doney, M. Gehlen, K. Lindsay, J. K. Moore, B. Schneider, and J. Segschneider (2010), Projected 21st century decrease in marine productivity: a multi-model analysis, Biogeosciences, 7(3), 979–1005, doi:10.5194/bg-7-979-2010. Uitz, J., D. Stramski, B. Gentili, F. D’Ortenzio, and H. Claustre (2012), Estimates of phytoplankton class-specific and total primary production in the Mediterranean Sea from satellite ocean color observations, Global Biogeochem. Cycles,26(2), n/a–n/a, V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 62
doi:10.1029/2011GB004055. Vlasenko, V., J. C. S´anchez-Garrido, N. Stashchuk, J. Garc´ıa-Lafuente, and M. Losada (2009), Three-dimensional evolution of large-amplitude internal waves in the Strait of Gibraltar, J. Phys. Oceanogr.,39, 2230–2246. Wesson, J. C., and M. C. Gregg (1994), Mixing at Camarinal Sill in the Strait of Gibraltar, J. Geophys. Res.,99 (C5), 9847–9878, doi:10.1029/94JC00256. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 63
33oN 34oN 35oN 36oN 37oN 38oN 39oN IBERIAN PENINSULA AFRICA Alboran Sea Strait of Gibraltar Gulf of Cadiz a) 8oW 6oW 4oW 2oW 0o 2oE 32oN 33oN 34oN 35oN 36oN 37oN 38oN 39oN b) Depth (m) 4500 4000 3500 3000 2500 2000 1500 1000 500 0 6oW 50’ 40’ 30’ 20’ 48’ 54’ 36oN 6’ 12’ TN Depth (m) 1200 1000 800 600 400 200 0 CS Figure 1. a) Model domain and computational grid. For the sake of clarity only one third of the total grid lines are shown. The rectangle encloses the Alboran Sea, the selected area in which some variables have been spatially averaged (see text for details). b) Model bathymetry with a zoom area of the Strait of Gibraltar displayed in the inset. Two relevant bathymetric features have been labelled: Camarinal Sill (CS), and Tarifa Narrows (TN). V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 64
6oW 5oW 4oW 3oW 2oW 1oW 30’ 35oN 30’ 36oN 30’ 37oN a) chl (mg m−3) MODEL (REF) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 6oW 5oW 4oW 3oW 2oW 1oW b) OBSERVATION Figure 2. a) Model derived climatology of surface chl for the period 2012-2013 (REF experiment; median values). Black contours display the model mean SSH. The contour interval is 1.5 cm, with solid and dashed lines corresponding to positive and negative values, respectively. The thick solid line corresponds to SSH=0. b) Surface chl climatology derived from remote sensing images corresponding to the period 2012-2013 (median values are shown). In the computation of the climatologies, satellite data were interpolated onto the model grid and its cloud mask was applied to the model outputs for a better comparison of the two data sets. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 65
Distance from CS (km) a) −20 0 20 40 mmol NO3 m−2 50 150 250 350 450 550 650 750 850 −2 0 2 0 5 10 15 20 25 30 Time (hours) Vel. (m/s) Distance from CS (km) b) −20 0 20 40 0 5 10 15 20 25 30 mmol NO3 m−2 CS EastWest 100 125 150 175 200 225 250 275 Figure 9. Temporal and spatial dependence of the depth-integrated nitrate along the central axis of the Strait of Gibraltar. Nitrate concentration is integrated from z=-94 to z=0, R0 −94NO3dz. Panel a) corresponds to the TID run, with the leftmost axis showing the batrotropic (depthintegrated) velocity over the Camarinal Sill, while b) corresponds to the REF run. Time is given in hours from March 15th 2013. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 72
0 2 4 6 8 mol s−1 (x 103) NO3 (−94<z<0) a) 0 1 2 3 mol s−1 (x 103) NO3 (−48<z<0) c) Apr May Jun 0 0.2 0.4 0.6 mol s−1 (x 103) PO4 (−94<z<0) b) Apr May Jun 0 0.1 0.2 0.3 mol s−1 (x 104) PO4 (−48<z<0) d) Figure 10. a) Nitrate transport at the eastern end of the Strait of Gibraltar (x= 40 km in Fig.7) as simulated in the TID (gray line; 3-hourly data) and REF (black dashed line) runs. The black solid line if the low-pass TID series (tidal variability removed). The transport is computed in the first 94 m of the water column, roughly encompassing the inflowing Atlantic layer. b) same as a) but with the transport computed in the first 48 m of the water column. c) and d) and as a) and b) but for phosphate. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 73
5oW 4oW 3oW 2oW 1oW 35oN 30’ 36oN 30’ 37oN w (m h−1) −40 −20 0 20 40 Figure 11. a) Synoptic field of vertical velocity at z=−88.5m, as simulated in the TID experiment. 30’ 35oN 30’ 36oN 30’ 37oNPPint NUDGING gCm−2y−1 0 100 200 300 400 500 6oW 5oW 4oW 3oW 2oW 1oW Figure 12. Mean depth-integrated primary production obtained in the nudging experiment. The polygon around the Strait of Gibraltar encloses the area where the nudging term described in the text has been applied. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 74 6oW
Parameter Symbol Value Units Maximum growth rates µmax1, µmax22.5, 1.4 d−1 Phytoplankton mortality rates mp 1, mp 20.1,0.1 d−1 Phytoplankton sinking rates wp 1, wp 20.5,0 m d−1 Phosphate half-saturation const. κP O41, κP O423.5·10−2,1.5·10−2µM P Nitrate+Nitrite half-saturation const. κIN1, κIN20.56,0.24 µM N Ammonium half-saturation const. κNH41, κNH420.28,0.12 µM N Ammonium inhibition ψ4.6 (µM N)−1 PAR saturation coeff. kpar1, kpar21.2·10−2,1.2·10−2(µEin m−2s−1)−1 PAR inhibition coeff. kinhib1, kinhib210−3,6·10−3(µEin m−2s−1)−1 Temperature curve coeff. A1.04 Temperature normalization coeff. τ1, τ23, 0.3 Zooplankton mortality rate mz3.3·10−2d−1 Maximum grazing rates gmaxa, gmaxb0.2, 3.3·10−2d−1 Grazing half-saturation const. Kp0.1 µM P DOM remineralization rates rDOP , rDON 10−2,10−2d−1 POM remineralization rates rP OP , rP ON 2·10−2,2·10−2d−1 POM sinking rate wP OM 10 m d−1 Table 1. Ecosystem model parameters and given values. Subscripts 1 and 2 refers to different phytoplankton species. V. Modeling the impact of tidal flows on the biological productivity of the Alboran Sea 75
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 77 Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. Cristina Naranjo(1), Simone Sammartino(1), Jesús García-Lafuente(1), María J. Bellanco(2), Isabelle Taupier-Letage(3). (1) Grupo de Oceanografía Física, Universidad de Málaga, 29071 Málaga, Spain (2) Centro Oceanográfico de Cádiz, Instituto Español de Oceanografía (IEO), 11006 Cádiz, Spain. (3) Aix Marseille Université, CNRS, Université de Toulon, IRD, MIO UM 110, Antenne de la Seyne, 83507 La Seyne, France Abstract Hydrological data collected in the Strait of Gibraltar have been used to examine the distribution and spatial-temporal evolution of the water masses in the area. The spatial variability has been addressed by means of a clustering method that determines the affinity of a collection of Temperature-Salinity samples to one of the water masses involved in the exchange. The method, which has been applied to a nearly-synoptic data set, highlights the clear evolution of the Mediterranean Waters as they flow westward through the Strait. While up to four different Mediterranean Waters are spatially distinguishable east of the main sill of Camarinal in the Strait, most of their differentiating characteristics are eroded after flowing over this restrictive topography due to mixing. West of the sill, therefore, speaking of a unique Mediterranean Water seems more appropriate. The same applies to the North Atlantic Central Water flowing in the opposite direction, which is noticeably modified along its path to the Mediterranean Sea, most of its transformation taking place in the Camarinal sill surroundings. A series of repeated transects carried out in the eastern and western sides of the Strait, provided a temporal analysis of the water masses evolution: the temporal variability manifests seasonality in the surface waters, while interannual signal is mainly detected in the deeper water masses. It is worth remarking the statistically significant positive trend of Western Mediterranean Deep Water (0.009ºC/year) and Winter Intermediate Water (0.03ºC/year), with the latter showing also intermittent occurrence in the Strait. Keywords: Strait of Gibraltar, Mediterranean outflow, water masses, cluster analysis.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 78 1. Introduction In the Mediterranean Sea (MedS, hereinafter) the Atlantic Water (AW) that flows in through the Strait of Gibraltar (SoG) is modified by evaporation and transformed into Mediterranean water, saltier and denser, which ends up flowing out through the SoG to the Atlantic Ocean. A simplified Mediterranean basin is schematized by an eastern and a western basins connected by the Strait of Sicily. In the eastern basin, Levantine Intermediate Water (LIW) is formed through open-sea convection. In the western basin, more specifically in the Gulf of Lion, Western Mediterranean Deep Water (WMDW) is formed by deep convection. It was known since long ago that the LIW was a permanent contributor to the outflow. However, the possibility that the WMDW was participating significantly in the outflow was first presented by Stommel et al. (1973), who attributed its presence to the Bernouilli aspiration of this water from great depth in the MedS over the main sill of Camarinal in the SoG. Subsequently, other authors have revisited the topic and stressed this thought [Bryden and Stommel, 1982; Gascard and Richez, 1985; Whitehead, 1985; Kinder and Parrilla, 1987; Kinder and Bryden, 1990; Millot and Taupier-Letage, 2005; J. García Lafuente et al., 2007; Naranjo et al., 2012; Naranjo et al., 2014] At present, it is accepted that this deep water is a permanent part of the outflow. Studies dealing with the outflow within and nearby the SoG used to focus on the two main Mediterranean Waters (MWs hereinafter), the LIW and the WMDW [Pettigrew, 1989; Bray et al., 1995; J. García Lafuente et al., 2007], which are easily identified by the maximum and minimum potential temperature, respectively, in the densest part of the θ-S diagram [Gascard and Richez, 1985]. Recent efforts made to clarify the hydrological characteristics of the water masses leaving the MedS through the SoG have suggested the presence of other Mediterranean water masses, more specifically, the Tyrrhenian Dense Water (TDW) and the Winter Intermediate Water (WIW) [Rhein et al., 1999; Millot et al., 2006; Millot, 2009, 2014b, a]. The first is formed by the mixing of old WMDW residing in the Tyrrhenian Sea with newly entered LIW flowing into the western MedS through the Strait of Sicily [Rhein et al., 1999; Millot et al., 2006]. The WIW is seasonally formed by convection of cooled modified Atlantic Water under severe winter condition along the continental shelf of the Liguro-Provençal sub-basin and Catalan Sea [Conan and Millot, 1995; Vargas-Yáñez et al., 2012]. At its source, it is the coolest water in the Western MedS [Salat and
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 79 Font, 1987; Lopez Jurado et al., 1995; Millot, 1999] and it is easily detected in any θ-S diagram by a minimum of potential temperature between potential density anomaly σθ=28.0 and σθ=29.0 [Millot, 2014a]. The volume of formed WIW has been reported to show marked interannual fluctuations [Pinot et al., 2002; Monserrat et al., 2008], the case of no formation being non-discardable [Pinot et al., 2002; Ribó et al., 2015]. These MWs are rather well differentiated (when present) at the eastern side of the SoG [Fuda et al., 2000; Millot, 2009], but the question remains as whether or not they are still distinguishable at the western part of the SoG once the Mediterranean outflow has crossed the Camarinal sill. The reason behind this noticeably different spatial distribution of the water masses in both halfs of the SoG (East-West) is the outstanding tidal dynamics in the area [Candela et al., 1990; Bryden et al., 1994; García-Lafuente et al., 2000; J. García Lafuente et al., 2007], which is strongly enhanced in the surrondings of Camarinal sill and westwards of it [Wesson and Gregg, 1994; J. C. Sánchez Garrido et al., 2008; J.C. Sánchez Garrido et al., 2011].The barotropic tidal curre nts interact with the SoG’s topography, mainly with Camarinal sill, to produce a remarkable internal tide [Candela et al., 1990; Bryden et al., 1994; García-Lafuente et al., 2000] that in turn gives rise to dissipation rates that are amongst the highest found in the world ocean [Wesson and Gregg, 1994]. The supercritical-to-subcritical flow transitions at the different critical (in hydraulic sense) sections, that happen not only in Camarinal sill but also oceanwards of it at specific times of the tidal cycle, drive that enhanced mixing [J.C. Sánchez Garrido et al., 2011; Jesús García Lafuente et al., 2013], which is the responsible for the fading out of the water masses identities in the western half of the SoG. On the other hand, Millot [2014a], using schematic mixing lines in the Mediterranean zone of a θ-S diagram, has proposed that the four MWs can be still detected as far as at 6º 15’ W to the West of the main sill and, even, traced along the Gulf of Cadiz in the Atlantic Ocean. This stand point differs from the widespread view of a Mediterranean Water that exits the SoG as a rather well mixed plume with typical properties of θ~13ºC and S~38.4 [M Baringer, 1993; M O N Baringer and Price, 1997] in which the different MWs water masses are not discernable. With the aim of provide a clear and standardized method to classify the water masses in the SoG, this work proposes a statistical method to automatically classify every water mass involved in the exchange. Two sets of data, described in Section 2, were specifically collected in the SoG area
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 80 to address the topic. The first dataset was acquired during the Gibraltar International Campaign (GIC, Section 2.1) and the second one throughout the lifespan of the INGRES projects (Section 2.2) funded by the Spanish Government. Section 3 describes the data processing, paying special attention to the description of the proposed method of analysis (Section 3.2). The hydrological information contained in these two sets of data has been exploited in different ways in this study. GIC data were collected during a very short period and allow us to make a quasi-synoptic description of the water masses distribution in the SoG. On the contrary, INGRES data gather samples spanning a rather long period of time and have the potential of addressing the time variability and evolution of the water masses. Section 4 discusses both topics in Subsections 4.1 (GIC) and 4.2 (INGRES) respectively. Finally, Section 5 summarizes the findings and conclusions of the study. 2. Data 2.1. CTD and MVP data from Gibraltar International Campaign In the framework of the international Hydrochanges programme sponsored by the Commission Internationale pour l'Exploration Scientifique de la Méditerranée (Mediterranean Science Commission, CIESM) and supported by the HyMeX programme, the French Mediterranean Institute of Oceanography carried out the Gibraltar International Campaign on board the R/V Tethys II from the 4th to the 6th July 2012. The cruise was aimed at obtaining high resolution Conductivity-Temperature-Depth (CTD) profiles along the transects showed in Figure 1 in order to give an accurate water mass characterization and distribution of Mediterranean waters within the SoG. Except for section R5, a Moving Vessel Profiler (MVP) was employed; this instrument allows semi-autonomous sampling of the water column with very high spatial resolution (horizontal averaged resolution is 500m while vertically resolution is 1m). A drawback of the MVP is its limited range depth (~350m). Transect R5 and a repetition of transect R2 were sampled with a CTD probe (SBE 911plus CTD, sampling frequency of 24Hz) that reached the seafloor. The CTD vertical profiles in these transects, however, are substantially further apart than MVP profiles (typical distance between casts ranging from 1 to 3 Km).
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 81 Figure 1. Map of the Strait of Gibraltar showing bathymetric contours, in meters. The black dots and red asterisks indicate the location of the vertical profiles along the 5 sampled sections for MVP and CTD data in GIC campaign (R1-R5), respectively. Blue circles represent the two CTD sections regularly repeated in the INGRES project (TAC and TES). The main sills of Espartel (ES) and Camarinal (CS), the small Tangier Basin (TB) between them and the Tarifa Narrows (TN) are also indicated. The inset shows the location of the Strait (SoG) between the Alboran Sea (AS), the westernmost basin of the Mediterranean Sea, and the Gulf of Cadiz (GoC) in the Atlantic Ocean. 2.2. Historical CTD data from INGRES project The INGRES projects were initiated in 2004 with the objective of monitoring the Mediterranean outflow and its variability in response to subinertial and longer-term forcing as well as the hydrological properties of the densest and, hence, deepest Mediterranean water leaving the MedS. At the time of this study the monitoring, which is planned to be kept on position sine die, is still in progress. Whenever the station was serviced (every 4 or 6 months) and weather permitting, CTD transects were accomplished. Among them, transects labeled TES and TAC in Figure 1 have been repeatedly sampled during the lifespan of INGRES projects. They make up an unevenly distributed time series since the meteorological conditions often prevented the accomplishment of one or both transects. Overall, TES was sampled 15 times and TAC 12 times (details about the dates when these transects were collected are shown in Table 1). TAC, located at the eastern entrance of the SoG, is the last MedS transect where the MWs may be found as unmixed as in the interior of the Alboran Sea, the westernmost basin of the MedS. Further west, the enhanced turbulence associated with the tidal dynamics and the very hydrodynamics of the exchange [Wesson and Gregg, 1994; J.C. Sánchez Garrido et al.,
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 88 4. Results 4.1. GIC dataset The tidal variability in the SoG, subdued by semidiurnal frequencies, makes the water masses pattern be dependent on the time of the tidal cycle when the transect was accomplished [J. García Lafuente et al., 2007] and the tidal phase during which the sampling was carried out must be specified for each transect. This information is provided by the sea level oscillation in Tarifa (see Figure 1). In this regard, it is interesting to remind that the barotropic semidiurnal tide in the SoG behaves like a standing wave [García-Lafuente et al., 2000; J. García Lafuente et al., 2007] and that the tidal flow goes westwards during the rising tide (low to high water, or flood tide) and eastwards during the falling tide (high to low water, or ebb tide). On the other hand, the GIC sampling was accomplished during a relatively short period of time (4th to 6th of July, 2012). In some sense, the observations are synoptic for lower frequency fluctuations (subinertial or seasonal/interannual variability) and they should reflect the water mass composition in the SoG during that period of time, despite the tidal variability. The results of the analysis of the GIC transects are presented from east to west in Figure 3 and Figure 4. During this survey, no traces of WIW were observed and the cluster analysis identifies only five water masses: WMDW, TDW, LIW, NACW and SAW. Figure 3 corresponds to the easternmost transect, R5, where the less mixed MWs that enter the SoG from the MedS are expected to be found. The transect was carried out from south to north with a CTD probe during the ebb tide, the last station being completed shortly after the slack tide of low water (Figure 3a). The θ-S values of the densest water sample were 12.92ºC-38.48, which corresponds to WMDW (Figure 3c). Figure 3d shows that this water stacks up in the south while the LIW layer is thicker in the north, a spatial pattern that agrees with Parrilla et al. [1986] and Naranjo et al. [2012]. Millot [2014b] pointed out that LIW (and also TDW to some extent) flows counterclockwise and is pushed to the north due to the Coriolis force, while the WMDW is compelled to flow attached to the southern shore preferably, where the incoming AWs, NACW in particular, are accumulated due to the Coriolis effect (Figure 3d). Such a distribution facilitates the mixing of WMDW with AWs in the south part, a fact reflected by the mixing lines of the southernmost stations of the transect, which head
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 89 directly towards the AWs region from the vicinity of the WMDW centroid (Figure 3c). This is not so for the sations located further north in the transect where the mixing lines bend towards the LIW centroid before heading towards the AWs, showing that it is mainly the LIW and not the WMDW nor the TDW that mixes with the AWs. The remaining transects discussed below were sampled with the MVP. The way the instrument samples the water column results in shorter times to complete a transect and higher spatial resolution (Figure 4). However, the maximum sampled depth is less than the one reached by CTD. Next transect to the west is R2 (panels in column I of Figure 4). It is located to the east of CS so that the water masses have not been exposed yet to the strong mixing happening in the sill area [Wesson and Gregg, 1994; J.C. Sánchez Garrido et al., 2011]. Moreover, it is not far from R5 and no significant differences are thus expected. That is the case for AWs, which depict a similar pattern (Figure 4-I-d). However, R2 transect was sampled during the rising tide and near the high tide (Figure 4-I-a). At this moment of the tide, the WMDW and TDW are allowed to reach shallower depths due to the interface rising associated with the internal tide [Bryden et al., 1994; Sánchez Román et al., 2012]. The situation is just the opposite of R5 sampling. The most interesting feature, however, is the spatial differentiation of the MWs (Figure 4-I-c,d) with WMDW occupying the southern part and TDW and LIW the northern area, although the latter is found at intermediate depths all over the transect. The following transect R1 is still east, although near, of CS (R1, Figure 4II). It is also close to R2 so a certain similarity between them is expected. But differences are apparent in the south due to the inversion of the tidal flow. R2 was accomplished from south to north and the southern stations were sampled ~1h before the high tide during the flood tide (Figure 4-I-a), while R1 was accomplished from north to south with the stations in the south done during the ebb tide, ~1h after the high tide (Figure 4-II-a). During the ebb phase the interface between AWs and MWs sinks nearby CS [Sánchez Román et al., 2008], giving rise to a considerably thicker AWs layer (Figure 4-II-d). The northern half of both sections were sampled under similar tidal conditions near the high water, when the interface is at its shallowest position [Sánchez Román et al., 2012], and show similar accumulation of LIW and TDW and a very thin layer of AWs. The spatial differentiation showed in Figure 4-I-c for R2 is easily recognisable in Figure 4-II-c for transect R1 as well, which is another remarkable similarity.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 90 Figure 3. (a) Tidal oscillation at Tarifa displaying the time of the CTD casts (red dots) in the R5 transect, which was accomplished from south to north. (b) θ-S diagram showing the CTD data of the R5 transect, contours lines indicate potential density anomaly. The centroids of the different water masses are marked with asterisks and the colour scale on the right identify the different casts by their latitude. (c) Zoom of the MWs area of the θ-S diagram. (d) Results of the cluster analysis where each colour represents the cluster associated with a water mass according to the legend on the right.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 91 Figure 4. Zonal evolution, from East to West, of the thermohaline properties for the different water masses involved in the exchange at the SoG. The figure is divided in four columns named I, II, III and IV and each of them is subdivided in rows. The upper row, (a), shows the sea level during the sampling of each transect, with the red circles indicating the time of the different casts. The second row, (b), displays the θ-S diagram of the whole section. The names and locations of the defined water masses are indicated. The third row, (c), is a zoom of the θ-S diagram that focuses on the MWs, contours lines are σθ. Finally, the last row, (d), shows the classification of the water masses in the section provided by the cluster analysis.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 92 Next transect to the west is R4 (Figure 4-III), which is already west of CS. The θ-S diagram shows two noticeable differences with regard to the three previous transects. Firstly, the θ-S curves bend towards the NACW centroid, implying a much greater impact of this water mass that now spreads downwards to 200m depth in the south (Figure 4-III-d). Secondly, the spatial differentiation of the MWs has disappeared and now they nearly lay along a single mixing line. This is an obvious outcome of the strong mixing in the Tangier basin [J.C. Sánchez Garrido et al., 2011] which makes the MWs lose their specific identity to a great extent. The cluster algorithm only returns one kind of MWs in this transect, LIW in this case, which is somewhat misleading in view of the θ-S diagram in Figure 4-III-c. Therefore, it requires clarification. The algorithm situates the water samples in a cluster, which is the one with the greater percentage of the water mass defined by the corresponding centroid. According to the chosen metrics (Eq. 1 and 2), the deep water samples in this transect (and also in the next one, R3, commented below) have similar proportions of the three MWs but a slightly higher proportion of LIW (Figure 5 1 ). Should we have displaced any of the centroids of the MWs by a tiny distance, the algorithm would have possibly returned a different prevailing cluster. The reasonable conclusion is that the MWs are hardly distinguishable once the Mediterranean outflow has passed CS and that the sensible option is to speak of a unique “Mediterranean water”. On the other hand, it is noteworthy here the effect of adding σθ to the metrics. Should it not be included, all the deep water would have been classified as TDW with an overwhelming percentage, as can be easily deduced from Figure 4-III-c. Its inclusion in the metrics makes the algorithm work more realistically in the sense that the actual sampled water is more likely to be the result of local mixing between LIW and WMDW (and TDW as well) than the outcome of the individual contribution of TDW. 1 Figures like Figure 5 are helpful in order to supplement the information displayed in Figures like Figure 4 or 7. However, and in order to keep the length of the manuscript within a reasonable limit, such Figures are not included in the text, but they are offered as supplementary material.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 93 Figure 5. Percentage of the MWs in transect R4 (Figure 4, column III). The sum of the three contributions gives the 100% in the Mediterranean layer. The colour scales go from 0 to 50% for the sake of clarity. Contours represent the σθ. The westernmost section R3, located in the western exit of the SoG, shows large similarity with the previous one. The fading out of the spatial differentiation of the MWs already detected in R4 is now more evident (Figure 4-IV-c), and so it is the prevalence of the NACW in the Atlantic layer (Figure 4-IV-d). Mixing lines are organised along two well-depicted directions, from MWs to NACW, and from NACW to SAW (Figure 4-IV-b), indicating that direct mixing of MWs with SAW does not happen any longer. Notice that this mixing can be partially detected in the previous transect R4 (Figure 4-III-b), this feature being almost the only difference among the two westernmost transects. The commentaries about the outputs of the cluster algorithm regarding the MWs made for R4 still apply in this transect. The previous discussion has focussed on the spatial evolution of the different water masses as they flow through the SoG. Table 4 presents the θ-S values of the coldest water sampled in every transect, which is taken as a proxi of the MWs, in order to illustrate their transformation along its path to the Atlantic Ocean. To this regard, two remarks are noteworthy. The first one is about the density of the MWs: all the coldest samples are the densest samples too, except for section R1 where the densest sample has σθ=29.090 while the coldest sample has σθ=29.087 (see also Figure 4-IIc). The second and more important remark is that the MVP does not reach the bottom so that the values reported in Table 4 should not be identified with the coldest/densest water in the section, which might not have been sampled. Even so, the regular east-west spatial trend of temperature is quite suggestive of the erosion the MWs undergo along the SoG. Table 4 also displays the coldest sample with salinity lower than 36.5, which is the best example representing NACW at each transect, in order to show the alteration of this water in its way towards the MedS. None of the former
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 94 concerns apply to this water mass which has been correctly sampled by the MVP. In the case of the MWs, θ increases and S decreases towards the west, the greatest jumps occurring between the transects R1 and R4 that surround CS, thus stressing the importance of this area as a source of turbulence [Wesson and Gregg, 1994; J.C. Sánchez Garrido et al., 2011]. The same applies to the North Atlantic Central Water flowing in the opposite direction, since both θ and S tend to increase as the water flows eastward. Once again the main changes happen in the surroundings of CS, although the rising of θ and, in particular , of S still continues from R2 to R5. R5 (east) R2 R1 R4 R3 (west) MWs θ (ºC) 12.92 12.94 12.97 13.05 13.07 S (psu) 38.48 38.47 38.46 38.43 38.40 NACW θ (ºC) 14.72 14.03 14.45 13.37 13.64 S (psu) 36.48 36.15 36.24 35.95 36.03 Table 4. The two first rows show the potential temperature and salinity, respectively, of the coldest MW sample observed in the transect (MWs block). Third and fourth rows show the potential temperature and salinity of the coldest sample with salinity less than 36.5 (NACW block). The different columns correspond to the different transects, which have been organised from east (R5) to west (R3), see Figure 1 for details. 4.2. INGRES Transects TES and TAC (see Figure 1) have been sampled repeatedly since 2004 and they are more regularly accomplished since 2011. Figure 6 presents the θ-S diagrams of both transects confirming the already mentioned evolution of the water masses as they progress through the SoG: the fading of the spatial differentiation between MWs from TAC (east) to TES (west) and the erosion of the NACW signal from TES (west) to TAC (east). In particular, the reddish colours in the inset of Figure 6b illustrates the fact that LIW, WIW and TDW flow preferably across the northern half of the SoG while the WMDW flows attached to the southern slope. The inset of Figure 6a shows how the former pattern is lost at TES.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 95 Figure 6. θ–S diagram showing the CTD data collected at TES (a) and TAC (b) transects during the INGRES project from 2004 up to the present. The colour scale indicates latitude and the contour lines shows σ θ . Black stars mark the θ–S pairs of the water masses involved in the exchange (see Table 2). The insets zoom in the Mediterranean water zone of the diagram. 4.2.1. Spatial distribution Figure 7a-b show the mean potential temperature and salinity distributions at TES and TAC sections, respectively, which have been obtained by averaging all the transects collected within INGRES. Figure 7c-d show the results of applying the cluster analysis to the same sets of data. The cluster analysis at TAC transect shows the averaged spatial distribution of the six water masses involved in the exchange (Figure 7d). Contrary to GIC, INGRES data recorded WIW, which now is identified flowing attached to the north shore just below a very thin layer of NACW. The remaining MWs display the same spatial pattern as in GIC: the WMDW, easily identified by θ<13ºC in Figure 7d, resides in the deepest layer and preferably stacked up in the southern half of the transect; the TDW and LIW, which appear as a salty wedge encircled by the isohaline 38.485 (grey line in Figure 7b), occupy an intermediate layer that thickens to the north. Any of these MWs is saltier (S>38.4, see Figure 7b) than the rather mixed MW at TES (Figure 7a), a result that can only be explained by the entrainment of AW by the Mediterranean outflow west of Camarinal Sill, as discussed in Jesús García Lafuente et al. [2011]. The two AWs are at the top of the water column in a layer that thickens from 100m in the north to 150m in the south. The presence of NACW is appreciably reduced as it experiences a marked mixing with respect to the TES transect (compare Figure 7c and Figure 7d). The blue colour identifying NACW in Figure 7d must be then interpreted as the sample being closer to NACW than to any other water mass and not as if it were aside the point marking the pure NACW.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 96 At TES, the cluster analysis only detects three water masses (Figure 7c), in agreement with the results obtained from the GIC data in this area. The bottom layer is occupied by the MWs, which the cluster analysis identifies as LIW (the same cautionary comments on the identification of MWs as LIW made for R4 transect in the previous Section apply here). These MWs, whose averaged salinity is 38.2 with maxima of 38.4, flow mainly through the southern channel below 250 meters, the volume flowing through the northern channel being much smaller. NACW is the prevailing water mass, occupying a layer from 50 to 250 meters in the southern channel (Figure 7c). Obviously not all this layer is NACW. It includes its mixture with the overlying (SAW) and underlying (LIW) waters, with the NACW entering in greater proportion than the others. Actually, Figure 7a shows a core of minimum salinity around 150-200m depth in the southern channel, which would be the depth where the purest NACW is flowing. Figure 7. Averaged potential temperature (colour scale) and salinity (labelled contours) of the whole dataset collected at TES (a) and TAC (b). Panels (c) and (d) show the distribution of water masses in these transects provided by the cluster analysis. Contours display the potential temperature. Insets show the location of the cast in each transect. ºC ºC
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 97 4.2.2. Temporal fluctuations in the core of the water masses This section addresses the time variability of the thermohaline characteristics found in INGRES data. To this aim, we have selected representative samples of each water mass with the same criterion for a given transect, although the criterion may change slightly from the west (TES) to the east (TAC) transect, as explained below. At TAC the criterion to define each water mass must be selected carefully, as the four MWs detected in this transect have only very small thermohaline differences. The SAW is selected as the warmest sample. The most appropriate criterion for the NACW will be to take the freshest sample, but NACW has been widely altered by mixing and the freshest water criterion may be not applicable at TAC (which however is pretty suitable in the TES section, see θ-S diagram of Figure 6b). Even more, the mixing could have been so important that speaking of NACW makes no clear sense. Thus, we only admit the presence of NACW if a clear minimum of salinity with respect to the overlying SAW is observed in the vertical profile, otherwise we ignore this water mass, even if temperature/salinity dots in the θ-S diagram bend gently towards the mark representing the NACW. Regarding the MWs, WMDW is determined as the coldest sample, TDW as the saltiest one, and LIW as the warmest whenever its salinity exceeds 38.4. As for the WIW, it is identified as the coldest sample between σθ=28.0 and σθ=29.0 (Millot, 2014), provided that it is visually detected in the θ-S diagram previously (that is, whenever the relative minimum around the WIW position in the diagram is positively identified, see inset in Figure 6b). Figure 8 shows the series of these representative samples at TAC transect, which are displayed along with the depth where the sample was found (in brackets) and the location of the profile (see labels on top of Figure 7). The seasonality of SAW is recognisable at TAC despite being more intense in the TES section (compare Figure 8a with Figure 9a). As for the NACW, it was positively identified only 4 out of 11 times (Figure 8b). In all these occasions its core mass was always found in the southern casts (cast 2 or 3) at depths between 20 and 85m, quite shallower than at TES, where the NACW core was between 150 and 200m (Figure 7a). Regarding MWs, except for 2012 when WIW was not observed, the four water masses were positively identified during all cruises. The lightest one is the WIW, which is at the top of the Mediterranean layer between 150 and 260m and it is detected in casts near the northern shore, in agreement with
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 104 Acknowledgement This work is a Spanish contribution funded by the National Project INGRES 3 (CTM2010-21229), a French contribution to the HyMeX and MOOSE programmes funded by MISTRALS, and to the CIESM HYDROCHANGES Programme (ciesm.org/marine/programs/hydrochanges.htm). Cristina Naranjo acknowledges the fellowship BES-2011-043421 from the Ministry of Economy and Competitiveness – Spain, and M. Jesús Bellanco acknowledges the pre-doc fellowship IEO-FPI 2011/10. We are also grateful to the crews of the R/V Angeles Alvariño and the R/V TETHYS II, and to G. Rougier (MIO) and Deny Malengros (SAM/MIO) who operated the MVP. The MVP has been funded through the CETSM project. References Baringer, M. (1993), Mixing and dynamics of the Mediterranean outflow(Ph. D. Thesis). Baringer, M. O. N., and J. F. Price (1997), Mixing and Spreading of the Mediterranean Outflow, Journal of Physical Oceanography, 27(8), 16541677, doi:10.1175/1520-0485(1997)027<1654:MASOTM>2.0.CO;2. Ben Ismail, S., K. Schroeder, C. Sammari, G. P. Gasparini, M. Borghini, and L. Aleya (2014), Interannual variability of water mass properties in the Tunisia–Sicily Channel, Journal of Marine Systems, 135(0), 14-28, doi:http://dx.doi.org/10.1016/j.jmarsys.2013.06.010. Bethoux, J. P., B. Gentili, J. Raunet, and D. Tailliez (1990), Warming trend in the western Mediterranean deep water, Nature, 347(6294), 660-662. Borghini, M., H. Bryden, K. Schroeder, S. Sparnocchia, and A. Vetrano (2014), The Mediterranean is becoming saltier, Ocean Sci., 10(4), 693-700, doi:10.5194/os-10-693-2014. Bray, N. A., J. Ochoa, and T. H. Kinder (1995), The role of the interface in exchange through the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 100(C6), 10755-10776, doi:10.1029/95JC00381. Bryden, H. L., J. Candela, and T. H. Kinder (1994), Exchange through the Strait of Gibraltar, Progress in Oceanography, 33(3), 201-248, doi:http://dx.doi.org/10.1016/0079-6611(94)90028-0. Bryden, H. L., and H. M. Stommel (1982), Origin of the Mediterranean Outflow, Journal of Marine Research, 40, 55-71. Candela, J., C. Winant, and A. Ruiz (1990), Tides in the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 95(C5), 7313-7335, doi:http://dx.doi.org/10.1029/JC095iC05p07313. Conan, P., and C. Millot (1995), Variability of the northern current off marseilles, western mediterranean-sea, from february to june 1992, Oceanologica acta, 18(2), 193-205.
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 105 Criado-Aldeanueva, F., J. García-Lafuente, J. M. Vargas, J. Del Río, A. Vázquez, A. Reul, and A. Sánchez (2006), Distribution and circulation of water masses in the Gulf of Cadiz from in situ observations, Deep Sea Research Part II: Topical Studies in Oceanography, 53(11–13), 1144-1160, doi:http://dx.doi.org/10.1016/j.dsr2.2006.04.012. Folkard, A. M., P. A. Davies, A. F. G. Fiúza, and I. Ambar (1997), Remotely sensed sea surface thermal patterns in the Gulf of Cadiz and the Strait of Gibraltar: Variability, correlations, and relationships with the surface wind field, Journal of Geophysical Research: Oceans, 102(C3), 5669-5683, doi:10.1029/96JC02505. Font, J. (1987), The path of the Levantine intermediate water to the Alboran sea, Deep Sea Research Part A. Oceanographic Research Papers, 34(10), 1745-1755, doi:http://dx.doi.org/10.1016/0198-0149(87)90022-7. Fuda, J. L., C. Millot, I. Taupier-Letage, U. Send, and J. M. Bocognano (2000), XBT monitoring of a meridian section across the western Mediterranean Sea, Deep Sea Research Part I: Oceanographic Research Papers, 47(11), 2191-2218, doi:http://dx.doi.org/10.1016/S09670637(00)00018-2. García-Lafuente, J., J. M. Vargas, F. Plaza, T. Sarhan, J. Candela, and B. Bascheck (2000), Tide at the eastern section of the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 105(C6), 14197-14213, doi:http://dx.doi.org/10.1029/2000JC900007. García Lafuente, J., E. Bruque Pozas, J. C. Sánchez Garrido, G. Sannino, and S. Sammartino (2013), The interface mixing layer and the tidal dynamics at the eastern part of the Strait of Gibraltar, Journal of Marine Systems, 117–118(0), 31-42, doi:http://dx.doi.org/10.1016/j.jmarsys.2013.02.014. García Lafuente, J., A. Sánchez Román, G. Díaz del Río, G. Sannino, and J. C. Sánchez Garrido (2007), Recent observations of seasonal variability of the Mediterranean outflow in the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 112(C10), C10005, doi:http://dx.doi.org/10.1029/2006JC003992. García Lafuente, J., A. Sánchez Román, C. Naranjo, and J. C. Sánchez Garrido (2011), The very first transformation of the Mediterranean outflow in the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 116(C7), C07010, doi:http://dx.doi.org/10.1029/2011JC006967. Gascard, J. C., and C. Richez (1985), Water masses and circulation in the Western Alboran sea and in the Straits of Gibraltar, Progress in Oceanography, 15(3), 157-216, doi:http://dx.doi.org/10.1016/00796611(85)90031-X. Hur, H. B., G. A. Jacobs, and W. J. Teague (1999), Monthly Variations of Water Masses in the Yellow and East China Seas, November 6, 1998, Journal of Oceanography, 55(2), 171-184, doi:10.1023/A:1007885828278. Kim, K., K. R. Kim, T. S. Rhee, H. K. Rho, R. Limeburner, and R. C. Beardsley (1991), Identification of Water Masses in the Yellow Sea and the East China Sea by Cluster Analysis, in Elsevier Oceanography Series, edited
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 106 by K. Takano, pp. 253-267, Elsevier, doi:http://dx.doi.org/10.1016/S04229894(08)70100-4. Kinder, T. H., and H. L. Bryden (1990), Aspiration of Deep Waters through Straits, in The Physical Oceanography of Sea Straits, edited by L. J. Pratt, pp. 295-319, Springer Netherlands, doi:10.1007/978-94-009-0677-8_14. Kinder, T. H., and G. Parrilla (1987), Yes, some of the Mediterranean outflow does come from great depth, Journal of Geophysical Research: Oceans, 92(C3), 2901-2906, doi:10.1029/JC092iC03p02901. Krahmann, G., and F. Schott (1998), Longterm increases in western Mediterranean salinities and temperatures: Anthropogenic and climatic sources, Geophysical Research Letters, 25(22), 4209-4212, doi:10.1029/1998GL900143. Leaman, K. D., and F. A. Schott (1991), Hydrographic Structure of the Convection Regime in the Gulf of Lions: Winter 1987, Journal of Physical Oceanography, 21(4), 575-598, doi:10.1175/15200485(1991)021<0575:HSOTCR>2.0.CO;2. Lopez Jurado, J. L., J. García Lafuente, and N. Cano Lucaya (1995), Hydrographic conditions of the Ibiza channel during November 1990, March 1991 and July 1992, Oceanologica acta, 18(2), 235-243, doi:http://archimer.ifremer.fr/doc/00097/20868/. Millot, C. (1999), Circulation in the Western Mediterranean Sea, Journal of Marine Systems, 20(1–4), 423-442, doi:http://dx.doi.org/10.1016/S09247963(98)00078-5. Millot, C. (2009), Another description of the Mediterranean Sea outflow, Progress in Oceanography, 82(2), 101-124, doi:http://dx.doi.org/10.1016/j.pocean.2009.04.016. Millot, C. (2014a), Heterogeneities of inand out-flows in the Mediterranean Sea, Progress in Oceanography, 120(0), 254-278, doi:http://dx.doi.org/10.1016/j.pocean.2013.09.007. Millot, C. (2014b), Levantine Intermediate Water characteristics: an astounding general misunderstanding!, Scientia Marina, 78(2), 165-171, doi:doi: 10.3989/scimar.04045.30H. Millot, C., J. Candela, J.-L. Fuda, and Y. Tber (2006), Large warming and salinification of the Mediterranean outflow due to changes in its composition, Deep Sea Research Part I: Oceanographic Research Papers, 53(4), 656-666, doi:http://dx.doi.org/10.1016/j.dsr.2005.12.017. Millot, C., and I. Taupier-Letage (2005), Circulation in the Mediterranean Sea, in The Mediterranean Sea, edited by A. Saliot, pp. 29-66, Springer Berlin Heidelberg, doi:10.1007/b107143. Monserrat, S., J. L. López-Jurado, and M. Marcos (2008), A mesoscale index to describe the regional circulation around the Balearic Islands, Journal of Marine Systems, 71(3–4), 413-420, doi:http://dx.doi.org/10.1016/j.jmarsys.2006.11.012. Naranjo, C., J. García-Lafuente, J. C. Sánchez-Garrido, A. Sánchez-Román, and J. Delgado-Cabello (2012), The Western Alboran Gyre helps ventilate
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 107 the Western Mediterranean Deep Water through Gibraltar, Deep Sea Research Part I: Oceanographic Research Papers, 63(0), 157-163, doi:http://dx.doi.org/10.1016/j.dsr.2011.10.003. Naranjo, C., J. Garcia-Lafuente, G. Sannino, and J. C. Sanchez-Garrido (2014), How much do tides affect the circulation of the Mediterranean Sea? From local processes in the Strait of Gibraltar to basin-scale effects, Progress in Oceanography(0), doi:http://dx.doi.org/10.1016/j.pocean.2014.06.005. Parrilla, G., and T. H. Kinder (1987), The Physical Oceanography of the Alboran SeaRep., DTIC Document. Parrilla, G., T. H. Kinder, and R. H. Preller (1986), Deep and intermediate mediterranean water in the western Alboran Sea, Deep Sea Research Part A. Oceanographic Research Papers, 33(1), 55-88, doi:http://dx.doi.org/10.1016/0198-0149(86)90108-1. Pettigrew, N. R. (1989), Direct measurements of the flow of western Mediterranean deep water over the Gibraltar sill, Journal of Geophysical Research: Oceans, 94(C12), 18089-18093, doi:10.1029/JC094iC12p18089. Pinot, J. M., J. L. López-Jurado, and M. Riera (2002), The CANALES experiment (1996-1998). Interannual, seasonal, and mesoscale variability of the circulation in the Balearic Channels, Progress in Oceanography, 55(3– 4), 335-370, doi:http://dx.doi.org/10.1016/S0079-6611(02)00139-8. Rhein, M., U. Send, B. Klein, and G. Krahmann (1999), Interbasin deep water exchange in the western Mediterranean, Journal of Geophysical Research-Oceans, 104, 23495-23508, doi:10.1029/1999jc900162. Ribó, M., P. Puig, and H. van Haren (2015), Hydrodynamics over the Gulf of Valencia continental slope and their role in sediment transport, Deep Sea Research Part I: Oceanographic Research Papers, 95(0), 54-66, doi:http://dx.doi.org/10.1016/j.dsr.2014.10.004. Rohling, E. J., and H. L. Bryden (1992), Man-induced salinity and temperature increases in western Mediterranean deep water, Journal of Geophysical Research: Oceans, 97(C7), 11191-11198, doi:10.1029/92JC00767. Salat, J., and J. Font (1987), Water mass structure near and offshore the Catalan coast during the winters of 1982 and 1983, Annales Geophysicae, Series B, 5 B(1), 49-54. Salat, J., P. Puig, and M. Latasa (2010), Violent storms within the Sea: dense water formation episodes in the NW Mediterranean, Adv. Geosci., 26, 53-59, doi:10.5194/adgeo-26-53-2010. Sánchez Garrido, J. C., J. García Lafuente, F. Criado Aldeanueva, A. Baquerizo, and G. Sannino (2008), Time-spatial variability observed in velocity of propagation of the internal bore in the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 113(C7), n/a-n/a, doi:10.1029/2007JC004624. Sánchez Garrido, J. C., G. Sannino, L. Liberti, J. García Lafuente, and L. Pratt (2011), Numerical modeling of three-dimensional stratified tidal flow
VI.Mediterranean waters along and across the Strait of Gibraltar, characterization and zonal modification. 108 over Camarinal Sill, Strait of Gibraltar, Journal of Geophysical Research: Oceans, 116(C12), C12026, doi:http://dx.doi.org/10.1029/2011JC007093. Sánchez Román, A., F. Criado Aldeanueva, J. García Lafuente, and J. C. Sánchez Garrido (2008), Vertical structure of tidal currents over Espartel and Camarinal sills, Strait of Gibraltar, Journal of Marine Systems, 74(1–2), 120-133, doi:http://dx.doi.org/10.1016/j.jmarsys.2007.11.007. Sánchez Román, A., J. García Lafuente, J. Delgado, J. C. Sánchez Garrido, and C. Naranjo (2012), Spatial and temporal variability of tidal flow in the Strait of Gibraltar, Journal of Marine Systems, 98–99(0), 9-17, doi:http://dx.doi.org/10.1016/j.jmarsys.2012.02.011. Santinelli, C., A. Ribotti, R. Sorgente, G. P. Gasparini, L. Nannicini, S. Vignudelli, and A. Seritti (2008), Coastal dynamics and dissolved organic carbon in the western Sardinian shelf (Western Mediterranean), Journal of Marine Systems, 74(1–2), 167-188, doi:http://dx.doi.org/10.1016/j.jmarsys.2007.12.005. Severin, T., P. Conan, X. Durrieu de Madron, L. Houpert, M. J. Oliver, L. Oriol, J. Caparros, J. F. Ghiglione, and M. Pujo-Pay (2014), Impact of openocean convection on nutrients, phytoplankton biomass and activity, Deep Sea Research Part I: Oceanographic Research Papers, 94(0), 62-71, doi:http://dx.doi.org/10.1016/j.dsr.2014.07.015. Smith, R. O., H. L. Bryden, and K. Stansfield (2008), Observations of new western Mediterranean deep water formation using Argo floats 2004–2006, Ocean Sci., 4(2), 133-149, doi:10.5194/os-4-133-2008. argasez, M ., . laza, . Garc a-Lafuente, T. Sarhan, J. M. Vargas, and P. Vélez-Belchi (2002), About the seasonal variability of the Alboran Sea circulation, Journal of Marine Systems, 35(3–4), 229-248, doi:http://dx.doi.org/10.1016/S0924-7963(02)00128-8. Vargas-Yáñez, M., P. Zunino, K. Schroeder, J. L. López-Jurado, F. Plaza, M. Serra, C. Castro, M. C. García-Martínez, F. Moya, and J. Salat (2012), Extreme Western Intermediate Water formation in winter 2010, Journal of Marine Systems, 105–108(0), 52-59, doi:http://dx.doi.org/10.1016/j.jmarsys.2012.05.010. Warn-Varnas, A., A. Gangopadhyay, J. A. Hawkins, and A. R. Robinson (2005), Wilkinson Basin area water masses: a revisit with EOFs, Continental Shelf Research, 25(2), 277-296, doi:http://dx.doi.org/10.1016/j.csr.2004.09.005. Wesson, J. C., and M. C. Gregg (1994), Mixing at Camarinal Sill in the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 99(C5), 9847-9878, doi:10.1029/94JC00256. Whitehead, J. A. (1985), A laboratory study of gyres and uplift near the Strait of Gibraltar, Journal of Geophysical Research: Oceans, 90(C4), 70457060, doi:10.1029/JC090iC04p07045. Yan, M. (2005), Methods of Determining the Number of Clusters in a Data Set and a New Clustering Criterion, Virginia, Blacksburg.
VII. GeneralDiscussion 109 VII. GENERAL DISCUSSION Brieflysummarized,theresearchofthisdissertationhasbeencarriedout asfollows:Thestartingpointwasthestudyoftheinfluenceofregional‐scale structuresonthedrainageofdeepMediterraneanwatertowardtheAtlantic, whichwascompletedbytheevaluationoftheconsequencesthattidalforcing intheStraitproducesboth,ontheventilationofthedeep/bottomlayersof theAlboranSeaandintheformationofdeepwaterintheGulfofLion.The nextstepclarifiedtheroleofthedynamicsoftheStraitinfuelingthe productivityofthesurfacewatersoftheAlboranSea,particularlyinits northwesternsector.Finally,anexhaustivecharacterizationofthewater massesinvolvedintheexchangewaspresented.Overall,thisresearch providessatisfactoryanswerstoquestionslikehowandfromwherethedeep waterisuplifted,whatisthetidalcontribution,howitaffectsatthephysical‐ ecologicalsystemandwhicharethewatermassesparticipatinginthedeep waterventilation.Amoredetaileddiscussionofthemainresultsispresented below. ChapterIIIstudiedtheroleoftheWAGinupliftingdeepwaterand elucidatedwhetherornotthegyrefacilitatesitsdrainagetowardtheAtlantic Ocean.AscommentedinsectionII.d,theexperimentalsupporttothisstudy wasprovidedbyatimeseriesofθobtainedbyaCTprobeplacednearthesea bedinESthatsampledthedensestwaterflowingouttheMediterranean. SimultaneousaltimetrydatawereemployedtodeviseanindexoftheWAG intensity,whichwascomparedwiththeθseries.Thecomparisonrevealsthat theyarehighlycorrelatedwithθlaggingbytwoweekstheWAGindex.This resultstronglysuggeststhattheWAGhelpstoraisetheWMDWpoolresiding inthedeeplayerofthewesternAlboranSea. Thedepthfromwhichthiswaterisaspiratedwasalsoaddressed.García Lafuenteetal.[2011]showedthattheaveragedminimumθobservedinES correspondstoθ=12.98CinCS.Thisisothermisfoundbetween300and600 mdepthintheeasternboundaryoftheStraitand,accordingtothehistorical MEDARdatabase[MEDAR‐Group,2002],atameandepthof520minthe southernhalfofthewesternAlboranSea.Ifthesamereasoningisappliedto theextrememinimaofθobservedinES,thentheentireMediterraneanlayer intheeasternapproachoftheStraitcouldbeforcedtooverflowCSandleave theMediterranean.
VII. GeneralDiscussion 110 AfurtherproofoftheeffectoftheStraitofGibraltaronthedynamicsof thedeeplayerintheAlboranSeawasprovidedbythecomparisonofthe outputsofthesamenumericalmodelappliedtotwodifferentdomains:the firstdomainincludedtheStraitbutitwasclosedinthesecondone.Inthefirst case,theisotherm12.85C,whichisagoodproxyoftheWMDWinthe Alboranbasin,wasshalloweranditwasevenshalloweracrosstheeastern mouthoftheStrait,thusreinforcingtheBernouillisuctionasthemechanism responsiblefortheupliftoftheWMDW. Aslongastheaspirationisinducedbythehighvelocitiesoccurringinthe Strait,thetidalcurrentsmustbeconsidered.Theywerenotincludedinthe previousstudyandnumericalmodelsthatignorethetidalforcingwould underestimatetheMediterraneanventilation.Theeffectoftideswas assessedinChapterIVbyemployingtwonumericalmodelsofdifferentspatial extension.Inbothcases,thesameapproachofcomparingtheoutputs producedbyrunswithandwithouttidesintheforcingtermswasthebasic toolofanalysis.Theregionalmodel,whosedomainincludedtheGulfofCadiz, StraitofGibraltarandAlboranSea,reproducedthewell‐knowntidally inducedincreaseofthelong‐termexchangeandagreatermixinginthecase ofthetidalrun.Furthermore,theinterfacelayerwasshallowerandthickerin theeasternpartoftheStrait,aresultofthetidallyenhancedmixinginthe TangierBasin,fromwherethewaterislateradvectedtotheeast.The shoalingandthickeningoftheinterfacemeansthatmoredeepwaterhas beenbroughttowardthesurfacelayerthaninthecaseofnon‐tidalrun.This inturncausesacolderandsaltierAtlanticJet,whichwas0.37°Ccolderand 0.47unitssaltieracrosstheeasternboundaryoftheStraitinthetidalrun. DespitetheweakeningofthissignatureastheJetproceedseastwards,itis stillnoticeableintheAlmeria‐Oranfrontintheeasternboundaryofthe AlboranSea. Thebasin‐scalemodel,whichincludedthewholeMediterraneanbasinand theGulfofCadiz,providedtwoworthnotingresults.Thefirstonedealswith thevolumeofwatercolderthan13°C(agoodproxyoftheWMDWinthe AlboranSea)flowingwestwardsthroughtheeasternboundaryoftheStrait (TACinFigure1),whichwas28%higherinthetidalrun,thussuggestingthat thedrainageofWMDWisparticularlyfavoredbytides.AlthoughtheWMDW volumeflowingouttheMediterraneanhashighinterannualvariability,itis
VII. GeneralDiscussion 111 noteworthythatthetidalrunalwaysprovidedgreaterflowsthanthenon‐tidal run.Thesecondisrelatedtotheeffectsthatthementioneddifferencesofthe AtlanticJetdensityobservedbetweentidalandnon‐tidalrunsintheregional modelcouldhaveontheformationofWMDWintheGulfofLion.Three proxiesofthedeepwaterformationinthisarea,namely,theMixedLayer Depth(MLD),thesurfaceareasusceptibletoparticipateinthedeep convection(DCA),andthedeepwaterformationrate,wereestimatedand comparedinthetworuns.Exceptforthefirstyear,thethreevariablesare alwaysgreaterinthetidalrun,whichistheexpectedoutcomeforaless buoyantincomingAtlanticJet(colderandsaltier)foundinthetidalrunofthe regionalmodel.Thesevariablesalsodisplayagreatyear‐to‐yearvariability,as itdoesthevolumeofWMDWflowingoutthroughtheeasternboundaryofthe Straitmentionedabove.Thesetwoprocessesseemtoberelated,reachingthe outflowofWMDWitsmaximumayearafterthegreatestWMDWformation event. OneofthelocaleffectsoftidesintheStraitdiscussedinChapterIVisthe enhancedmixingandconsequentthickeningoftheinterfaciallayer. ConsideringthattheMediterraneanisadeep,nutrientrichwater,itsmixing withtheoverlyingAtlanticwaterfertilizestheAtlanticJet,whichwillbemore enrichedthanintheexperimentwithouttidalforcing.Oncemore,inChapter Vtheoutputsoftwoidenticalbiogeochemicalmodelsrunwithandwithout tidesarecompared.Thecomparisonindicatesthattidesincreasethe productivityoftheAlboranSeainabout40%,andthattheincreaseiseven greaterbetweentheStraitand3°W(thewesternAlboranSea)whereit reachesa60%.ThehypothesisoftidalflowsintheStraitfertilizingthe Atlanticinflowisexaminedbycomputingthetransportsofnutrienttoward theAlboranSeathroughtheeasternboundaryoftheStrait.Thenutrientflow ishighlytime‐dependentinthetidalsimulation:nitratetransportoscillates between350‐550mmolNO3m‐2inthiscase,whileitismuchlessvariablein thenon‐tidalexperiment(200‐250mmolNO3m‐2).Theimportantpointis, however,thatthemeannutrientfluxisnearly75%greaterinthetidalcase, whichexplainsthelargeincreaseofproductivityobtainedinthetidal simulation. Tidescouldalsofertilizetheinflowingwateralsobydiapycnalmixing drivenbytheeastwardspropagationoftheinternaltidalboregeneratedat
VII. GeneralDiscussion 112 thesill,whichisreproducedbythetidalsimulation.Thetopicwas investigatedconductinganewexperimentinwhichthephysicswasasinthe non‐tidallyforcedsimulationbutwithamodifiedsetofecosystemequations inwhichbiochemicaltracersarenudgedtowardstheirvaluesinthetidalrun withintheStrait.Resultsfromthisexperimentdonotshowsignificant differenceswithrespecttothetidalsimulation,thusindicatingthatthe increasedproductivityobservedintheAlboranSeaisdrivenbythe intensificationofnutrienttransportthroughtheStrait,whiletidallydriven flowsintheAlboranSeawouldnothavemuchrelevance. AdetaileddescriptionofthewatermassesexchangedthroughtheStrait wasthesubjectoftheChapterVIofthisdissertation.InadditiontoWMDW andLIW,TDWandWIWhavebeenrecentlyproposedtobepartofthe outflow.TogetherwiththetwoAtlanticwatersNACWandSAW,theysumup tosixwatermassesinvolvedintheexchange.Therelativecontributionof eachofthiswatermassestotheexchangehasbeenaddressedbyacluster analysisoftheirhydrologicalproperties.TheobservationsfromtheGibraltar InternationalCampaign(seeChapterVIfordetails)indicatethatthewater masseseastofCShavearatherclearspatialdistribution,withtheWMDW stackedupinthesouthandtheTDWandLIWoccupyingpreferablythe northernarea,althoughthelatterisfoundatmiddepthatalllatitudes.On theotherhand,westwardsofCSthespatialdifferentiationofthe Mediterraneanwatersisgreatlyreducedor,even,vanishesduetothestrong mixingtakingplaceintheTangierbasin.Heretheclusteralgorithmrevealeda mixedwaterwithsimilarproportionsofthethreeprevailingMediterranean watermasses(WMDW,LIWandTDW;inthiscaseWIWwasnotpresent), whichpreventustoestablishreliabledistinctionsbetweenthem.Therefore, fromthemainsillofCamarinaltothewest,itseemsreasonabletospeakofa uniqueMediterraneanWater.Theanalysisofthezonalevolutionofthe densestwaterindicatesthatboththepotentialtemperatureandthesalinity decreaseasweproceedtowardthewest,andthatthesameappliestothe NACWwhenwemovefromwesttoeastinthiscase.Bothofthemunderwent themainchangesinthesurroundingsofCSwhencrossingthesill,thus confirmingthegreatmixingoccurringinthisarea. ThementionederosionofNACWsignalandthefadingofthespatial differentiationbetweentheMediterraneanWatersarealsovisibleinthe
VII. GeneralDiscussion 113 repeatedCTDtransectscollectedinthetwoboundingsectionsoftheStrait withintheframeoftheINGRESprojects.Thiscollectionofdatasupportsthe analysisofGICdataexplainedaboveandconfirmsthepresenceoffour MediterraneanWatermassesintheeasternpartoftheStraitthatreducesto auniqueMediterraneanwaterinthewesternpart.Timevariabilityhasbeen analyzedwiththisdatasetandratherclearseasonalfluctuationswerefound forSAWandNACWinthewesternsection,whichwererelatedtotheannual solarcycleandwiththeupwellingseasonintheGulfofCádiz,respectively. ThefourMediterraneanwatersarewelldistinguishedintheeasternsection, althoughthedataalsohighlightstheintermittencyoftheWIW,whichwas notdetectedinyear2012,justtheyearoftheGICexperiment.Coinciding withtheabsenceofWIW,aslightlywarmerandfresherLIWwasdetected andtheWMDWshowedarelativeθmaximum,whichwasalsoinsinuatedin themixedMediterraneanwateratthewesternsection.Thesefeaturespoint atamild2010‐2011winterwithoutWIWproduction.Ontheotherhand,the WMDWshoweditsabsoluteminimumofpotentialtemperatureinyear2009, whichwouldbeanindicatoroftheextraordinaryWMDWformationinthe GulfofLionthisyear[Salatetal.,2010].Overall,theanalysisofthetemporal fluctuationsillustratestherelativelyhighinterannualvariabilityofthe presenceandlocationoftheMediterraneanwatermassesintheStraitof Gibraltaratitseasternapproach,avariabilitythatissmoothedoutinthe westernexitduetothestrongmixingintheTangierbasin.
IX. ResumenenEspañol 120 Porúltimoseabordaelanálisisylaidentificacióndelasmasasdeaguaque formanpartedelosflujosentranteysaliente.AésterespectoMillotetal. [2006]sugierequeademásdelasdosmaasdeaguaMediterránea tradicionalmenteconsideradasenelflujodesalida,elAguaIntermedia Levantina(LIWenadelante)ylaWMDW,otrasmassasdeaguaparticipanen ésteflujo.EnesteestudioMillotponedemanifiestolapresenciadelAgua DensadelTirreno(TDW),formadaporlamezclaentreWMDWresidenteenel TirrenoconLIWprocedentedelacuencaEstedelMediterráneo.Segúnlos resultadosdeMillotlaTDWhabríasidodetectadaenelflujodesalidaenlos años2003y2004.EnlamismalíneaMillot[2009]sugierelaparticipaciónde unacuartaaguaMediterránea,elaguaintermediadeInvierno(WIW),quese formaporlapérdidadetemperaturadelAguaAtlánticaModificada(MAW) duranteelinviernoenlaplataformacontinentaldelaregiónnoroccidentaldel MarMediterráneo.LaWIWseidentificaenundiagramaθ‐Sporunmínimode salinidadenelintervalodedensidad28<σθ<29,yestaríaposicionadaentrela NACWylasotrastresaguasMediterráneas(WMDW,LIWyTDW).Más recientemente[Millot,2014]reconsideralaideaincidiendoenlahipótesisde unflujosalienteformadopordiferentesproporcionesdelascuatroaguas Mediterráneas,queseríanaundistinguiblesinclusodespuésdesobrepasarel umbraldeCamarinalensurecorridohaciaelOcéanoAtlántico.Elautorbasa susconclusionesenelanálisisdelapendientedelaslíneasdemezcladefinidas enundiagramaθ‐S.Sinembargolaidentificacióndelasdiferentesmasasde aguaMediterráneasmásalládelumbraldeCamarinalesunresultado controvertido[GarcíaLafuenteetal.,2011].Lacuestiónesampliamente tratadaenelCapítuloVI(MediterraneanwatersalongandacrosstheStraitof Gibraltar,characterizationandzonalmodification)enelcualseexaminanlas masasdeaguaquefluyenatravésdelEstrechodeGibraltar,cuálessu distribuciónyvariabilidadtemporal. LascaracterísticastermohalinasdelasaguasMediterráneasquefluyena travésdelEstrechosediferenciansutilmente,yademássupresenciaoposición enunaseccióndeterminadaesmuyvariableydependeengranmedidadela fasedelamarea.LacampañaoceanográficaGibraltarInternationalCampaign, (verelcapítuloVIparamásdetalles)sellevóacaboconelobjetivodetomar nuevasmedidasenelEstrechomedianteunMVP(MovingVesselProfiler),este instrumentopermiteunmuestreodealtaresoluciónespacial,ademásde favorecerlarealizacióndelostranseptosenuncortoperiododetiempo,ypor
IX. ResumenenEspañol 121 lotantominimizalamencionadavariabilidadenelconjuntodedatostomados. Lasmedidasseprocesaronusandounanálisisdeclústerespecíficamente desarrolladoparadichoconjuntodedatos.Elmétodoclasificacadamedidadel CTDenunclusterenfuncióndeladistanciaaunoscentroidespreviamente definidosporunascaracterísticasθ‐S‐σquesecorrespondenconcadaunade lasposiblesmasasdeaguaquepuedenformarparteenalgúnmomentodel flujoatravésdelEstrecho.Lavariabilidadtemporalsehaplanteadopormedio deunacoleccióndemedidasdeCTDenlosextremosEsteyOestedelEstrecho deGibraltar(ESyTACdenlaFigure1),dichasseccionessehanvenido repitiendodesde2005enelmarcodelosproyectosINGRES. Losdetallesespecíficossobrelametodologíaencadacasopueden encontrarseenloscapítuloscorrespondientes. IX.c. Resultados más significativos Demaneraresumida,lainvestigacióndesarrolladaalolargodeesta memoriasehallevadoacabodelsiguientemodo;elpuntodepartidahasido elestudiodelainfluenciaquelasestructurasdeescalaregionalpuedentener enlaevacuacióndelaguaMediterráneaprofundaatravésdelEstrechode GibraltaryhastaelOcéanoAtlántico.Seguidamenteelpuntoanteriorse completaconlaevaluacióndelasconsecuenciasquelainclusióndel forzamientomarealenelEstrechodeGibraltartienetantoenlaventilación delascapasprofundasdelMardeAlboráncomoenlaformacióndeagua profundaenelGolfodeLeón.Elsiguientepasoconsisteendilucidarcuáles elpapelqueladinámicadelEstrechotieneenlaproductividadprimariadelas aguassuperficialesenelMardeAlborán,particularmenteensuparte noroccidental.Porúltimosepresentaunacaracterizacióndetalladadelas masasdeaguaqueparticipanenlosflujosintercambiados.Ensuconjunto estamemoriaresuelvesatisfactoriamentecuestionescomodedóndeycómo esaspiradaelaguaprofunda,cualeslacontribucióndelamareaental proceso,comoafectaesteforzamientomarealalsistemafísico‐ecológicoy cuálessonlasmasasdeaguaqueparticipanenlaventilacióndelagua profunda.Losprincipalesresultadossediscutenacontinuación.
IX. ResumenenEspañol 122 EnelCapítuloIIIsehaestudiadoelpapelquedesempeñaelWAGenla aspiracióndeaguaprofundaysehaaclaradoqueestegiroanticiclónico facilitasudrenajehaciaelOcéanoAtlántico.Paraellosehanestudiadolas seriesdetiempodeθobtenidasconunsensordeconductividad‐temperatura desplegadocercadelfondoenEspartelquemonitorizaelaguamásdensaque fluyehaciafueradelMediterráneo.Simultáneamentemediantedatosde altimetríasehaideadouníndiceparaseguirlaevolucióndelaintensidaddel WAG,elcualsehacomparadoconlaseriedeθ.Elresultadoesunaalta correlaciónentreambasvariables,conundesfasededossemanasentreel índicedelWAGylaθmedidaenEspartel.EsteresultadosugierequeelWAG facilitaelascensodelaWMDWqueresideenlapartemásprofundadelMar deAlboránoccidental. Ademássehaestimadolaprofundidaddesdelacualestasaguasson aspiradas.SiguiendoaGarcíaLafuenteetal.[2011],laθmínimapromedio observadaenelumbraldeEspartelsecorresponderíaconunaθ=12.98Cen elumbraldeCamarinal.Dichaisotermaseencuentraentre300y600mde profundidadenelextremoEstedelEstrecho(secciónTACenlaFigure1)y,de acuerdoconlabasededatoshistóricosdeMEDATLAS[MEDAR‐Group,2002], aunos520mdeprofundidadenlamitadsurdelMardeAlboránoccidental. Siestemismoprocedimientoseaplicaalosextremosdelamínimaθ observadaenelumbraldeEspartelsepuedededucirquetodalacapa MediterráneacercadelasecciónEstedelEstrechopodríasersuccionada, superarelumbraldeCamarinalyabandonarelMediterráneo. PorotroladosehaestudiadocómoinfluyeladinámicadelEstrechode GibraltarenlacapaprofundadelMardeAlboránpormediodela comparacióndelassalidasdelmismomodelonuméricoaplicadoados dominiosdiferentes,enelprimercasoseincluyeelEstrechomientrasqueen elsegundoéstesecerró.Enelprimercasolaisotermade12.85C,queesun buenarepresentacióndelWMDWenlacuencadeAlborán,resultósermás somerasobretodoenlazonamáscercanaalextremoEstedelEstrecho.Dicho resultadoindicaquelasucciónporefectoBernouillieselmecanismo responsabledelaaspiracióndelWMDW. Debidoaqueelprocesodeaspiracióndeaguasprofundasestá claramenteinducidoporlasaltasvelocidadesquetienenlugarenelEstrecho deGibraltar,lascorrientesmarealesdebenserconsideradas.Estascorrientes
IX. ResumenenEspañol 123 nosehanincluidoenlosestudiosanteriores,sinembargolosmodelos numéricosquenotienenencuentaelforzamientomarealpodríanestar subestimandolaventilaciónenelMarMediterráneo.Elefectodelasmareas sedesarrollaenelCapítuloIVhaciendousodedosmodelosnuméricoscon diferenteescalaespacial,enamboscasosseprocedecomparandolassalidas deunasimulaciónconmareasconrespectoaotraquenoincluyeel forzamientomareal.ElmodeloregionalincluyeelGolfodeCádiz,elEstrecho deGibraltaryelMardeAlborán,estemodeloreproduceelincrementodel intercambioalargotérminoinducidoporlasmareasasícomounamayor mezclaenelcasodelasimulaciónconmareas.Además,lacapainterfaciales mássomeraydemayorespesorenlaparteEstedelEstrecho,comoresultado delamayormezclaenlacuencadeTángerenpresenciademareas.Desde aquíestaaguaseráposteriormenteadvectadahaciaelEste.El asomeramientoyelmayoresperordelainterfaseimplicaquemásagua profundasehatransportadoalascapassuperficialesqueenelcasodeno incluirmareas.ElloasuvezcausaunchorroAtlánticomásfríoymássalino, 0.37Cmásfríoy0.47unidadesmássaladoalolargodelasecciónEstedel Estrecho(secciónTACenlaFigure1)enlasimulaciónqueincluyemareas. EstaseñalpuedereconocerseinclusoenelfrenteAlmería‐Oránenelextremo orientaldelMardeAlboránapesardesuerosiónamedidaqueelchorro AtlánticofluyehaciaelEste. ElmodeloaescaladecuencaincluyeelMarMediterráneoyelGolfode Cádiz.Apartirdeéstemodelosehanobtenidodosresultadossignificativos. Elprimerodeellostienequeverconelvolumendeaguaqueseacumulapor debajodelos13C(unlímitequerepresentabienelWMDWenelMarde Alborán)quefluyehaciaelOesteatravésdelextremoorientaldelEstrecho deGibraltar(TACenlaFiguraFigure1),dichovolumenesun28%mayorenla simulaciónconmareas,locualsugierequeeldrenajedeWMDWestá favorecidoporlaaccióndelasmareas.ApesardequeelvolumendeWMDW quefluyehaciafueradelMediterráneotieneunagranvariabilidadinteranual, hayquedestacarquelasimulaciónconmareasarrojasiempremayoresflujos queenelcasodenoincluirmareas.Elsegundoresultadorelacionalosefectos mencionadosanteriormenteenelchorroAtlántico(másdensoenla simulacióndemareas)conlaformacióndeWMDWenelGolfodeLeón.Se handefinidotresvariablesqueevalúanlaformacióndeaguaprofundaen ésteárea,laprofundidaddelacapademezcla(MLDdesussiglaseningles),el
IX. ResumenenEspañol 124 áreasuperficialquepuedeparticiparenelprocesodeconvecciónprofunda (DCAdesussiglaseningles),yelratiodeformacióndeaguaprofunda.Estas variablessecalcularonycompararonenambassimulaciones.Elresultado indicaque,exceptoenelprimeraño,lastresvariablesmencionadasson siempremayoresenlasimulaciónqueincluyeelforzamientomareal,comose esperaqueseaparaunchorroAtlánticoconmenosflotabilidad(másfríoy mássalado)vistoenelmodeloregional.Ademáslasvariablesmencionadas muestranfluctuacionesanualesimportantes,aligualquelohaceelvolumen deWMDWquefluyeatravésdelasecciónEstedelEstrechocomentado anteriormente.Estosdosprocesosparecenestarrelacionados,talesasíque elflujosalientedeWMDWesmáximounañodespuésdelmayoreventode formacióndeWMDW. UnodelosefectoslocalesdelasmareasenelEstrechodiscutidaenel CapítuloIVeselincrementoenlamezclayelconsecuentemayorespesorde lacapainterfacial.ConsiderandoqueelaguaMediterráneaesunagua profunda,ricaennutrientes,sumezclaconelaguaAtlánticasituadasobre ellafertilizaríaelchorroAtlántico,queestaríamásenriquecidoqueenel experimentosinforzamientomareal.Denuevo,enelCapítuloVsecomparan lassalidasdedosmodelosbiogeoquímicosejecutadosconysinmareas.Dicha comparaciónindicaquelasmareasprovocanunaumentodelaproductividad delMardeAlborándealrededordeun40%,yqueesteincrementoesincluso mayoreneláreaquequedaentreelEstrechoy3W(elMardeAlborán Occidental)dondeesteaumentopuedealcanzarel60%.Lahipótesisdeque losflujosmarealesenelEstrechopuedenfertilizarelflujoAtlánticode entradaseestudiamedianteelcálculodeltransportedenutrienteshaciael MardeAlboránatravésdelasecciónEstedelEstrecho(TACenlaFigura1).El flujodenutrientesvaríaampliamenteconeltiempoenlasimulacióncon mareas,eltransportedenitratooscilaentre350‐550mmolNO3m‐2endicho caso,mientrasqueesmuchomenosvariableenelexperimentoqueno incluyeforzamientomareal(200‐250mmolNO3m‐2).Elresultadomás interesante,sinembargo,esqueelflujomediodenutrienteses aproximadamenteun75%mayorenelcasodeconsiderarlasmareas,locual explicaríaelgranincrementodeproductividadqueseobtieneenla simulaciónconmareas.
IX. ResumenenEspañol 125 LamareatambiénpodríafertilizarelaguaquefluyehaciaelMediterráneo mediantemezcladiapicna,pormediodelapropagaciónhaciaelEstede ondasinternasgeneradasenelumbral,lascualesreproduceelmodelocon mareassatisfactoriamente.Estaposibilidadsehainvestigadomedianteel desarrollodeunnuevoexperimentoenelcuallafísicasemantienecomoen lasimulaciónnoforzadaconmareasmientrasquelostrazadores biogeoquímicossemodificanparaquetiendanhaciasusvaloresdentrodel Estrechoenelmodeloconmareas.Losresultadosdeesteexperimentono muestrandiferenciassignificativasconrespectoalasimulacióninicialcon mareas,conlocualseconcluyequeelincrementodeproductividadquese observaenelMardeAlboránsedebealaintensificacióndeltransportede nutrientesatravésdelEstrecho,mientrasqueotrosflujosmarealesenelMar deAlboráncomolapropagacióndeondasinternasnosonsignificativos. Unadetalladacaracterizacióndelasmasasdeaguaqueseintercambiana travésdelEstrechoeselobjetivodelCapítuloVIdeéstamemoria.Ademásde laWMDWylaLIW,recientementesehapropuestoquetambiénelTDWyel WIWformanpartedelflujodesalida.JuntoaestasaguasMediterráneas fluyenensentidoopuestodosaguasAtlánticas,laSAWylaNACW,entotal sumanhastaseismasasdeaguaparticipandoenelintercambio.La contribuciónrelativadecadaunadeestasseismasasdeaguaalintercambio sehaabordadopormediodeunanálisisdeclusterdesuspropiedades hidrológicas.LasobservacionesobtenidasenlaGibraltarInternational Campaign(verelCapítuloVIparamásdetalles),muestranquelasmasasde aguaalEstedelumbraldeCamarinalsedistribuyenespacialmentedeuna formamuchomásclara.LaWMDWseapilacontralaparedSurylaTDWyla LIWocupanpreferiblementeeláreanorte,aunquelaLIWseencuentraa profundidadesmediasentodaslaslatitudes.Porotrolado,alOestedel umbraldeCamarinalladiferenciaciónespacialdelasAguasMediterráneas dejadesertanclara,einclusodesaparecedebidoalamezclaquetienelugar enlacuencadeTánger.Enestazonaelalgoritmoúnicamentedetectaun aguamuymezcladaconproporcionessimilaresdelastresmasasdeagua Mediterráneaspredominantes(WMDW,LIWyTDW,enestacampañalaWIW noseencontrabapresenteenningunadelassecciones),locualimpide distinguiradecuadamenteentreellas.Porlotanto,desdeelumbralde CamarinalhaciaelOeste,parecerazonablereferirseaunaúnicaagua
IX. ResumenenEspañol 126 Mediterránea.Elanálisisdelaevoluciónzonaldelaguamásdensaindicaque ambas,laθylasalinidaddisminuyenamedidaquenosdesplazamoshaciael Oeste,ylomismoocurreconelNACWamedidaquefluyedeOesteaEste. Ambasexperimentanloscambiosmássignificantesenlosalrededoresdel umbraldeCamarinal,locualdacuentadelaimportantemezclaquetiene lugarenlazona. LaerosióndelaseñaldelaNACWyladesaparicióndeladiferenciación espacialentrelasaguasMediterráneassedetectatambiénenlasrepeticiones delasseccionesCTDllevadasacaboenlasdosseccionesdelosextremosEste yOestedelEstrechoenelmarcodelosproyectosINGRES(Figure1,TACyTES respectivamente).Estosdatosapoyanlosresultadosobtenidosenelanálisis delosdatosdelaGibraltarInternationalCampaigndetalladosenelpárrafo anterior,yconfirmanlapresenciadehastacuatromasasdeagua MediterráneaenlasecciónEste,quesereducenaunaúnicamasadeagua MediterráneaenlaparteOeste.Lavariabilidadtemporalanalizadaconeste conjuntodedatosmuestraquetantolaSAWcomolaNACWpresentan variabilidadestacionalenlasecciónOeste,debidaalciclosolaranualyla estacióndeafloramientoenelGolfodeCádizrespectivamente.Lascuatro aguasMediterráneasdefinidasanteriormentesepuedendistinguir claramenteenlasecciónEste,noobstantelosdatosresaltanlaintermitencia delaWIW,quenosedetectóen2012,talycomoseobservóelmismoañoen losdatosdelaGibraltarInternationalCampaign.Coincidinendoconla ausenciadeWIWlaLIWquesemidiófueligeramentemáscálidaymásdulce, ylaWMDWsecaracterizóporunmáximorelativo,lamismaseñalseinsinúa tambiénenlamezcladeaguasMediterráneasdelasecciónOeste.Estas característicasapuntanauninviernosuaveen2010‐2011enelque probablementenoseformóWIW.Porelcontrario,laWMDWpresentaun mínimoabsolutodetemperaturapotencialenelaño2009,locualpodríaser unefectodelaextraordinariaformacióndeWMDWquetuvolugarenel GolfodeLeóntalaño[Salatetal.,2010].Laserietemporalhamostradoque, sobretodo,lasaguasMediterráneastienenunaaltavariabilidadinteranual queafectatantoasupresenciacomoasulocalizaciónenlasecciónEstedel EstrechodeGibraltar,unavariabilidadquesesuavizaenelcontornoOeste debidoalaintensamezclaqueocurreenlacuencadeTánger.
IX. ResumenenEspañol 127 IX.d. Principales conclusiones de esta memoria LadinámicadelEstrechodeGibraltaresfundamentalenlaventilaciónde lasaguasprofundasdelMardeAlborán,lacualasuvezesfacilitadaporla presenciacasipermanentedelgiroanticiclónicodelMardeAlborán Occidental(CapítuloIII). EnelcapítuloIIInosetieneencuentaladinámicamareal,sinembargoen elcapítuloIVseconcluyequeestadebeserresultaenlosmodelosaescala regionalydecuencaquepretendenreproduciralgunascaracterísticas relevantesdelintercambioalargotérminoy,muyprobablemente,otros importantesprocesosquetienenlugarmásalládelEstrecho,comola formacióndeaguaprofundaenelGolfodeLeón. ElmismorazonamientoseaplicaalaproductividadprimariaenelMarde Alboránlacualsereproducesatisfactoriamentecuandoelforzamiento marealseincluyeenunmodeloregionalfísico‐ecológico(CapítuloV).Más específicamenteseencuentraqueladiferenciaentreteneronoencuentalas mareaspuedesuponerhastaun60%másdeproducciónprimariaenla cuencaOestedeAlborán. ConrespectoaladistribuciónespacialdelasmasasdeaguaenelEstrecho (CapítuloVI),hastacuatromasasdeaguaMediterráneassehandetectadoen elflujosalienteenlaseccióndelcontornoEstedelEstrecho,apesardeque unadeellas,laWIW,puedenoestarpresenteendeterminadosaños.Laseñal deestasaguassesuavizalentamenteamedidaquefluyenhaciaelOeste,ylo hacenabruptamenteunavezsobrepasanelumbraldeCamarinal.Segúnlo anterior,alOestedelumbralseríamásapropiadohablardeunaúnicaagua Mediterránea.Conrespectoalavariabilidadtemporalseobservaquelas señalesestacionalessonmásevidentesenlasaguasAtlánticas,mientrasque lavariabilidadinteranualesmásnotableenlasaguasMediterráneas.
X.Listofacronyms 128 X. List of acronyms AC: Algeciras‐Ceuta AS: AlboranSea AWs: Atlanticwaters CS: CamarinalSill CT: Conductivity‐Temperature CTD: Conductivity‐Temperature‐Depth DCA: DeepConvectionArea ES: EspartelSill GIC: GibraltarInternationalCampaing GMSM:GlobalMediterraneanSeaModel LIW: LevantineIntermediateWater MAW: ModifiedAtlanticWater MLD: MixedLayerDepth MVP: Movingvesselprofiler MWs: Mediterraneanwaters NACW: NorthAtlanticCentralWater RSGM: RegionalStraitofGibraltarModel SAW: SurfaceAtlanticWater SoG: StraitofGibraltar TDW: TyrrhenianDenseWater
X.Listofacronyms 129 WAG: WesternAlboranGyre WIW: WinterIntermediateWater WMDW:WesternMediterraneanDeepWater θ: Potentialtemperature σθ: Potentialdensity
EstaTesisDoctoralseharealizadograciasalaconcesióndeunabecade FormacióndePersonalInvestigador(BES‐2011‐043421)delaSubdirección GeneraldeFormacióneIncorporacióndeInvestigadoresdelMinisteriode CienciaeInnovación,enelmarcodelProyectoINGRES‐3(CTM2010_21229) “FlujoMediterráneoenGibraltar,influenciaenlaventilacióndelasaguas Mediterráneasyprimeraevoluciónyacoplamientoconlasaguascentralesen elGolfodeCádiz”. EltrabajoseharealizadoenelDepartamentodeFísicaAplicadaIIdela UniversidaddeMálagayparcialmente,graciasalasayudasparala realizacióndeestanciasbrevesconcedidasporlamencionadasubdirección, eneldepartamentodeClimateModelling,delEnteperlaNuoveTecnologie, l’Energiael’Ambiente(ENEA),centrodeinvestigaciónenRoma(Italia).