Environmental modelling in the Gulf of Cadiz: Heavy metal distributions in water and sediments

  • Published on

  • View

  • Download


adkmectnumpichicrsithpplrianrt models have been compared with measurements in the GoC. In particular, theb-basiraltar,northetic coasScience of the Total Environment 407 (2009) 33923406Contents lists available at ScienceDirectScience of the Totl semechanisms of water exchange between the Atlantic Ocean and theMediterranean Sea; as well as the behaviour of the dense plume ofMediterranean water (Ambar and Howe, 1979; Garca-Lafuente et al.,2006; Criado-Aldeanueva et al., 2006; Machn et al., 2006; Garca-Lafuente and Ruz, 2007). The distribution of suspended matter andsediment transport in the GoC has also been investigated (Gonzalezet al., 2007; Freitas and Abrantes, 2002; Lobo et al., 2004; Cravo et al.,2006; Palanques et al., 19861987).The freshwater inputs of rivers discharging in the GoC are relativelyecosystem functioning. Thus, several papers concerning the distributionof metals in the GoC have been published in the last years (Sainz andRuiz, 2006; Morillo et al., 2004; Elbaz-Poulichet et al., 2001; Fernndez-Caliani et al.,1997). Indeed, studies on the distribution of tracemetals incoastal waters are frequently published in recent times (see for instanceLi et al., 2007; Suntornvongsagula et al., 2007; Cuong et al., 2008; Chenand Jiao, 2008; Valds et al., 2008; Marn-Guirao et al., 2008).The objective of this work consists of studying the dynamics ofheavy metals in the northern GoC by means of numerical modelling.small. However, the Guadalquivir, Guadiana athe southern Iberian Peninsula (Fig.1), presentmetal concentrations since they drain the IberRuiz, 2006), one of the most important mini Fax: +34 954486436.E-mail address: rperianez@us.es.0048-9697/$ see front matter 2009 Elsevier B.V. Adoi:10.1016/j.scitotenv.2009.01.023f oceanographic studiesed at understanding theecological interest. Consequently, it is relevant to study and understandthe geochemistry and dispersion patterns of heavy metals in the GoCsystem, since thiswill help assessing the potential inuence ofmetals onRecently, the GoC has been the subject odealing with surface and deep circulation, aimthe Strait of Gibraltar. The western bou9W meridian (Fig. 1).ndary is usually dened by the1. IntroductionThe Gulf of Cadiz (GoC) is the suwhich is nearest to the Strait Of GibOcean and the Mediterranean Sea. Itsboundaries are, respectively, the AtlanMediterranean Sea. 2009 Elsevier B.V. All rights reserved.n of the Atlantic Oceanconnecting the Atlanticrn, southern and easternts of Spain, Morocco andEurope. Mineral resources have been extracted in the last 5000 yearsduring two main periods: the Roman age and the last two centuries.During the last period, intensive exploitation has led to a relevantenvironmental impact, with vast surfaces coveredwithmining residuesand subjected to erosion (Sainz and Ruiz, 2006). The GoC is responsiblefor 510% of sh and shell-sh catches of Spain and Portugal (Beckerset al., 2007), holding important living resources of commercial andGulf of Cadiz plumes reach the Strait of Gibraltar, thus the three rivers constitute a source of pollutants into theSediment transportHeavy metalssediment and metal transpocontamination of sediments collected along the southern coast of Spain iswell reproduced by themodel. MetalEnvironmental modelling in the Gulf of Cand sedimentsR. Periez Dpto. Fsica Aplicada I, E.U. Ingeniera Tcnica Agrcola, Universidad de Sevilla. Ctra. Utreraa b s t r a c ta r t i c l e i n f oArticle history:Received 3 September 2008Received in revised form 16 December 2008Accepted 9 January 2009Available online 25 February 2009Keywords:Numerical modellingHydrodynamicsThe Gulf of Cadiz (GoC) connGoC is carried out throughcirculation and a 2D barotrosediment transport model wGoC. Then heavymetal dispemetal interactions and usestransport model has been athree rivers draining the Ibej ourna l homepage: www.end OdielTinto rivers, instronglyenhancedheavyian Pyrite Belt (Sainz andng areas in the south ofll rights reserved.iz: Heavy metal distributions in water1, 41013-Sevilla, Spains the Atlantic Ocean and theMediterranean Sea. An environmental study of theerical modelling. First, a 3D baroclinic model is used to obtain the residualmodel is applied to calculate tides. The results of thesemodels are used by a 3Dh provides suspended matter concentrations and sedimentation rates in theon patterns are investigated using a 3Dmodel which includes watersedimente outputs of the hydrodynamic and sediment transport models. The metalied to simulate the dispersion of Zn, Cu and Ni introduced into the GoC fromPyrite Belt, in the southern Iberian Peninsula. Results from the hydrodynamic,al Environmentv ie r.com/ locate /sc i totenvModels have been widely applied to simulate contaminant dispersionsince they may provide insights on the main environmental processesgoverning such dispersion and, consequently, may help to describeand characterize the environment (Scott, 2003). In particular, modelshave beenwidely applied to heavymetal (Tappin et al., 1997;Wu et al.,2005; Prandle et al., 1996; Liu et al., 1998) and radioactive element(Harms, 1997; Cetina et al., 2000; Periez, 2003; Monte et al., 2006)dispersion in coastal waters.Fig. 1. General localization of the study area and topography of the GoC (depths in m). The localization of Guadiana, OdielTinto and Guadalquivir rivers is also shown.3393R. Periez / Science of the Total Environment 407 (2009) 339234063394 R. Periez / Science of the Total Environment 407 (2009) 33923406Although some interesting modelling works describing watercirculation off Iberia and Morocco coasts have been published(Johnson and Stevens, 2000; Batteen et al., 2000), these models havea relatively low resolution, not providing detailed information aboutthe GoC basin circulation features. In the rst case, spatial resolution is10 min in both longitude and latitude (5 times the grid cell in thepresent model). In the second reference it is 10 km in longitude andlatitude, about three times larger than in this work. Other modellingworks are specically devoted to the study of Mediterranean waterspreading (Jungclaus and Mellor, 2000; Johnson et al., 2002; Serra etal., 2005). The excellent paper by Peliz et al. (2007) presents threenested model domains aimed at reproducing known features of theAzores current and of circulation inside the GoC.Published models describing trace metal dispersion in the GoCconsider metals as conservative tracers, no interacting with sedimentsand without any other sources and sinks (Elbaz-Poulichet et al., 2001;Beckers et al., 2007). The rst authors use a model to estimate thedilution of a conservative tracer released by the OdielTinto rivers.Beckers et al. (2007) apply a numerical model to reproduce observedmetal (again considering metals as conservative tracers) concentra-tions in surface waters of the northern GoC. Analysis of model resultsshowed that sources/sinks of metals due to interactions with sedi-ments (adsorption/desorption reactions as well as erosion and depo-sition processes) were apparent. Models which try to reproducemeasured levels of metals in bed sediments of the GoC have not beenpublished yet, to the author's knowledge.The model described in this paper consists of three sub-models:rstly, a hydrodynamic module which provides currents over thedomain. Two hydrodynamic models are used. A 2D barotropic modelis applied to calculate tides and 3D baroclinic model is used to obtainthe residual (mean) circulation. Tidal currents must be calculatedsince they may increase the bed stress and hence enhance sedimentresuspension and affect deposition of particles as well. Indeed, thesecond sub-model is a sediment transport model which providessuspended matter concentrations and sedimentation rates over thedomain. The third sub-model is the metal transport module, whichincludes advection/diffusion plus uptake/release reactions of metalsbetween the dissolved and solid (suspended matter in the watercolumn and bed sediments) phases.The threemodules, aswell as the numerical techniques used to solvethe involved equations are described in Section 2. Next, model resultsare discussed separately for water circulation, sediment transport andmetal distributions. Some sensitivity analysis are nally described.2. Model description2.1. Hydrodynamic modelsAs commented above, tides are required to calculate bed stress overthe domain, since it will affect sedimentation rates in the shallowerareas and/or where stronger tidal currents exist. A 2D depth-averagedmodel has been applied to calculate surface tides. This is a reasonableapproach that has already been successfully used in the Strait ofGibraltar (Tejedor et al.,1999; Periez and Pascual-Granged, 2008), theAlborn Sea (Periez, 2008) and even in the complete MediterraneanSea (Tsimplis et al.,1995). As has been shown, it is safe to neglect densitydifferences in tidal computations (Dyke, 2001; Yanagi, 1999).Equations are standard and may be seen for instance in Periez(2005a). The solution of these equations provides the water currentsat each point in themodel domain and for each time step. Currents aretreated through standard tidal analysis (Pugh, 1987) and tidalconstants are stored in les that will be read by the sedimenttransport model. The barotropic model includes the two main tidalconstituents,M2 and S2. Thus, the hydrodynamic equations are solvedfor each constituent and tidal analysis is also carried out for eachconstituent separately.The long-term circulation is obtained from a full 3D, primitiveequation, baroclinic hydrodynamic model. It is based upon thehydrostatic and Boussinesq approximations on a plane. The modelincludes equations for salinity and temperature evolution and waterdensity is calculated from them using a standard state equation(Batteen et al., 2000). The one-equation turbulencemodel described inPeriez (2005b) has been used to calculate the vertical eddy viscosity.Details on the 3D equationsmay be seen, for instance, in Kowalick andMurty (1993). A summary of the main equations involved in bothhydrodynamic models is presented in Appendix A.2.2. Sediment transportThe transport of sediments is described by a 3D advectiondiffusionequation to which some terms are added. These are external sources ofparticles (river supply), terms describing particle deposition on theseabed and erosion from the bed to the water column, and verticalsettling. The formulation of these processes is based upon standardformulae. Thus, the erodability constant is used for the erosion term.Particle settling and deposition are described using the settling velocity,which isobtained fromStoke's law.Critical erosionanddepositionstressesare used as usual. Details on the mathematical formulation may be seenelsewhere (Periez, 2005b; Liu et al., 2002a; Lumborg and Windelin,2003; Cancino and Neves,1999). Finally, it is also possible to calculate netsedimentation rates (SR) as the balance between the deposition anderosion terms. A summary of the equations may be seen in Appendix B.2.3. Metal transportNon conservative pollutants are those which do not remaindissolved in the water column, but have a certain afnity to be xedto particles (suspendedmatter and bed sediments). Thus, three phasesare included in the model: dissolved, suspended particles and bedsediments. Only ne sediments are considered (particles withdiameter b62.5 m) since metals are predominantly xed to these(Eisma, 1993). The exchanges between the dissolved and solid phasesmay be described in terms of kinetic transfer coefcients. Thus,assuming that adsorption/release reactions are governed by a singlereversible reaction, a coefcient k1 governs the transfer from the liquidto the solid phase and a coefcient k2 governs the inverse process. Also,the migration of metals to the deep sediment is included. Thus, metalsdeposited on the sediment surfacewill be buried byparticle depositionand will migrate below the mixed sediment layer which directlyinteractswith the dissolvedphase. This effectmaybe easily treated as adecay process with constant burial given by (Periez, 2008):burial =SRsL1where L is the sediment mixing depth (the distance to which thedissolved phase penetrates the sediment), s is the sediment bulkdensity (dry mass divided by wet volume) and SR is the sedimenta-tion rate provided by the sediment transport sub-model.It is known that adsorption depends on the surface of particles perwater volume unit at each point and time. This quantity has beendenoted as the exchange surface (Periez, 1999, 2000, 2002). Thus,the kinetic coefcient k1 is written as:k1 = Sm + Ssb = km1 + ks1 2where Sm and Ss are the exchange surfaces for suspended matter andbottom sediments respectively (dimensions [L]1) and is a parameterwith the dimensions of a velocity denoted as the exchange velocity(Periez, 1999, 2000, 2002). The delta function is introduced to takeinto account that only the deepest water layer interacts with the bedsediment. Thus b=1 for the deepest layer and b=0 elsewhere.3395R. Periez / Science of the Total Environment 407 (2009) 33923406Assuming spherical particles, the exchange surfaces are written as(see references cited above):Sm =3mR3andSs =3Lf 1 p R4where R and are particle radius and density respectively, m is thesuspended matter concentration, f is the fraction of ne particles inthe sediment, p is sediment porosity and is a correction factor thattakes into account that part of the sediment particle surface may behidden by other sediment particles. Finally, is the thickness of thedeepest water layer. This formulation has been successfully used in allmodelling works cited above. Real particles are not spheres, but withthis approach it is possible to obtain an analytical expression for theexchange surface (Duursma and Carroll, 1996). The kinetic coefcientk2 is considered to be constant.The equation that gives the time evolution of metal concentrationin the dissolved phase, Cd, is:@Cd@t+ u@Cd@x+ v@Cd@y+ w@Cd@z= A@2Cd@x2+@2Cd@y2 !+@@zK@Cd@z k1Cd+ k2mCs + bk2LsfAs5where Cs and As are, respectively, the concentrations of metals insuspendedmatter and bottom sediments. u, v andw arewater velocitiesalong the x, y and z axis and A and K are, respectively, the horizontal andvertical diffusion coefcients.The equation that gives the time evolution of metal concentrationin suspended matter is:@ mCs @t+ u@ mCs @x+ v@ mCs @y+ w ws @ mCs @z= A@2 mCs @x2+@2 mCs @y2 !+@@zK@ mCs @z + km1 Cd k2mCs bSED 6where ws is the particle settling velocity and SED is the deposition ofmetals from the deepest water layer to the sediment evaluatedaccording to:SED = SRCs b : 7Note that (b) means that the corresponding magnitude isevaluated at the deepest water layer.The equation for the temporal evolution of metal concentration inthe bottom sediment mixed layer is:@As@t= ks1Cd b Lsf k2As burialAs + SED 8where the deposition is now calculated as:SED = SRCs b sLf: 9The total metal content, Ap, in the sediment below the mixeddepth is given by the following equation:@Ap = burialLsfAs: 10@t2.4. Numerical solution and parametersThe topography of the GoC was obtained from the NOAA (USNational Ocean and Atmosphere Administration) GEODAS database,available on-line, with a resolution of 2 min in both longitude andlatitude. The grid consists of 64136 cells, extending from 8 to5.9 in longitude and from 33 to 37.5 in latitude. Fifty verticallevels are considered, with increasing thickness from the sea surfaceto the bottom.All the equations were solved using explicit nite differenceschemes. In particular, the scheme described in Periez (2005a) isused in the 2D barotropic hydrodynamic model. In the 3D baroclinicmodel the scheme in Kowalick and Murty (1993) is applied. Thesecond order accuracy MSOU (Monotonic Second Order Upstream)advection scheme (Vested et al.,1996) is applied for all advective termsand a second order scheme is used for diffusion as well (Kowalick andMurty, 1993).Time-step in the different sub-models is selected with care toassure that stability conditions are satised. A time-splitting proce-dure was used to treat the watersediment metal interactions. This isessentially the same as in hydrodynamic models, when time-splittingis used to solve the external and internal modes. The procedure isrequired because a very restrictive stability condition is introduced bysuch interaction terms (Periez, 2005a). Essentially, it consists ofusing a smaller time step to solve watersediment metal interactions.Thus, for each advection/diffusion time step, a number of smallersteps are needed to solve metal partition.Some open boundary conditions must be provided. In the baro-tropic model, tidal surface elevations and phases are specied alongthe open boundaries of the computational domain from Schwiderski(1980a,b). A radiation condition is applied to the water velocitycomponent that is normal to the open boundary (Jensen, 1998). Theequations are integrated, starting from rest, until stable oscillationsare obtained. Then tidal analysis is carried out and tidal constants arestored in les. In the case of the 3D baroclinic model, water velocitiesare specied along the Strait of Gibraltar from the average inow (intothe Mediterranean) of Atlantic Water and deep outow (from theMediterranean to the Atlantic) of Mediterranean water. Watertemperature and salinity along the open boundaries are speciedfrom the NOAA WOA05 (World Ocean Atlas 2005), available on-line.Summer values are used since sampling campaigns were carried outin this season. The 3D model is again started from rest and a windstress of 0.045 N/m2, directed to the south, is applied. It correspondsto the summer characteristic value (Machn et al., 2006). Residualcurrents are stored in les that will be later read by the sediment andmetal transport sub-models.A summary of parameters involved in the models is given inTable 1. Suspended particles (diameter b63 m) are characterized by amean size that is considered to be representative of suspended matterin this environment. The particle size controls, through Stoke's law,the settling velocity of particles and, as a consequence, affects thesedimentation rate as well. It also affects the adsorption of metalsfrom the dissolved phase (Eqs. (3) and (4)). The value used for particlediameter was 8 m. Freitas and Abrantes (2002) have found thatparticles with diameter b10 m are dominant in all water masses inthe Gulf of Cadiz. Moreover, this is the same value selected for a modelof the Alborn Sea (Periez, 2008) after a sensitivity analysis (theGoC reects the Alborn Sea since surface waters in the rst ow intothe second andMediterraneanwaters ow, in the deep layer, from theAlborn Sea into the GoC). Finally, the selected value is also consistentwith Liu et al. (2002b), who used a radius of 2.5 m to describe nesilts in their model. Of course, it would be convenient to have adetailed particle size spectra and then using several particle classes(Periez, 2005b), but it is not available in current literature.The critical deposition stress typically ranges between 0.04 and20.1 N/m (Tattersall et al., 2003), while the critical erosion stressThe correction factor has been xed as 0.1 according to previousmodelling works (Periez, 2000, 2003; Periez et al., 2005). Anintermediate value of 0.5was used for the sediment porosity. Indeed, itis 0.6 in the Alborn Sea model according to measurements (Periez,Table 1Summary of model parameters.Parameter description Value SourceWater kinematic viscosity =1.064106 m2/s Standard valueVertical diffusion coefcient K=variable Turbulence equationHorizontal diffusion A=2.0 m2/s Dick and Schonfeld (1996)Sediment mixing depth L=0.1 m Periez (2000, 2003)Particle density =2600 kg/m3 Standard valueParticle radius R=4.0 m Freitas and Abrantes (2002)3396 R. Periez / Science of the Total Environment 407 (2009) 33923406ranges 0.11.5 N/m2 (Tattersall et al., 2003). In the present applicationintermediate values of 0.06 and 1.0 N/m2 have been taken for thecritical deposition and erosion stresses respectively, as in the AlbornSea model described in Periez (2008). The erodability constant isxed as E=1.6103 kg/m2s. This parameter typically variesbetween 2104 and 3103 kg/m2s (Tattersall et al., 2003). Thefraction f of ne particles in bed sediments must be dened over themodel domain. This information has been compiled from Hernndez-Molina et al. (2006) and Gonzalez et al. (2007). Schematically, sandand sandy mud dominate to a depth of 30 m. The middle shelf, to adepth of 100 m, is characterized by an extensive mud belt. Finally,outer sediments below 100 m are dominated by mixtures of sand andclay.Suspended matter concentrations in the three main rivers (Fig. 1)have been dened. Of course, these magnitudes present seasonal (andat a shorter scale as well) variations. However, representative meanvalues have been used for the present application. In the OdielTintoestuary, the mean value measured in Periez et al. (1996) has beenused: 30 g/m3. The mean value measured by Gmez-Parra et al.(2000), 120 g/m3, was selected for the Gualdalquivir estuary. Finally,Fraction of small particlesin sedimentf=variable Gonzalez et al. (2007)Correction factor =0.1 Periez (2000, 2003);Periez et al. (2005)Sediment porosity p=0.5 Periez (2008) (similar value)Erodability E=1.6103 kg/m2s Tattersall et al. (2003)Critical erosion stress ce = 1.0 N/m2 Tattersall et al. (2003)Critical deposition stress cd=0.06 N/m2 Tattersall et al. (2003)Desorption kinetic coefcient k2=1.16105 s1 Nyffeler et al. (1984)Zn distribution coefcient kd=7.0104 IAEA (2004)Ni distribution coefcient kd=2.0104 IAEA (2004)Cu distribution coefcient kd=2.6104 CalibrationIn the cases of E and the critical stresses for erosion and deposition, the selected valuescorrespond to intermediate values within their range of variation commonly found inliterature. A reference (Tattersall et al., 2003) is given as an example.in the Guadiana estuary, a mean value of 30 g/m3 is in agreement withMachado et al. (2007) and with the empirical relation given for thisestuary by Wolanski et al. (2006).The sediment transport sub-model is run until steady particledistribution and SR are obtained. The model is forced by residualcurrents provided by the 3D baroclinic model. Instantaneous tidalcurrents are used only to calculate the bed stress. This approach hasbeenmade since themodel has been run to simulate the dispersion of apassive tracer with and without tides (of course residual circulation isconsidered in both cases), and differences in the tracer concentrationelds between both simulations could not be appreciated (thesimulated time was just a few months). Suspended particle concen-trations and sedimentation rates are stored in les that are read by themetal transport model.Table 2Values of river discharge metal concentrations (nM).Metal Guadiana Guadalquivir OdielTintoCu 42 58 214Ni 20 43 67Zn 49 33 129The vertical diffusion coefcient is taken equal to the vertical eddyviscosity, which is provided by the turbulence equation included inthe 3D baroclinic model. The horizontal diffusion coefcient dependson the horizontal grid spacing. Following Dick and Schonfeld (1996):A = 0:2055 103x1:15: 11The present grid resolution (2 min) gives A=2.0 m2/s.The model has been applied to simulate the dispersion of threemetals: Zn, Cu and Ni. Values for parameters describing metaltransport are summarized in Table 1 too.Fig. 2. Map of the GoC showing all locations mentioned in the text.2008). The sediment mixed depth is L=0.1 m (Periez, 2000, 2003).The so-called equilibrium distribution coefcient, kd, describes thepartition of a tracer between the dissolved and solid phases at equi-librium (Duursma and Carroll, 1996):kd =CsCd12where Cs and Cd are tracer concentrations in the solid and dissolvedphases respectively. This parameter is of course depending on theTable 3Established, index obs, (NOAA, 1982) and computed, index comp, amplitudes (A, cm)and phases (g, deg) of tidal elevations at several locations indicated in Fig. 2.Station M2 S2Aobs gobs Acomp gcomp Aobs gobs Acomp gcompFaro 92 94 99 68 32 125 36 91Chipiona 102 54 104 62 41 82 38 85Rota 105 52 103 62 37 78 38 85Cadiz 100 87 99 61 37 110 36 83Ayamonte 100 59 101 65 32 88 36 89Huelva 102 56 105 65 38 82 38 88Casablanca 99 56 92 53 35 81 36 77Rabat 88 59 98 57 35 83 36 78left) and current amplitude in m/s (right).3397R. Periez / Science of the Total Environment 407 (2009) 33923406tracer geochemical behaviour and on environmental conditions, thusit is a site-specic parameter.As described before (Nyffeler et al., 1984), k2 is very similar evenfor elements with a rather different geochemical behaviour, being (see Eq. (2)) the essential parameter describing the tracer geochem-ical behaviour. Thus, the same value is given to k2 for Ni, Zn and Cu,which is k2=1.16105 s1. This value has been successfully used inearlier simulations in the OdielTinto estuary (Periez et al., 2005)and in the Strait of GibraltarAlborn Sea (Periez, 2008) for severalelements (Ra, Cs, Pu). It was measured for Cs by Nyffeler et al. (1984).The exchange velocity for eachmetal is obtained from the equationFig. 3. Amplitude of the M2 tide in m (relating this parameter with k2 and its corresponding equilibriumdistribution coefcient, kd (Periez, 2005a):kd =k23R: 13The kd mean values recommended by the International AtomicEnergy Agency (IAEA, 2004) for coastal waters have been used. Thisprocedure has been widely used before (Periez 1999, 2000, 2003,2008).Distribution coefcients are given in Table 1. Nevertheless, they mayvary inmore than one order ofmagnitude depending on environmentalconditions. Consequently, the model sensitivity to this parameter hasbeen studied and results are presented below (Section 3.3.2). The IAEAdoes not provide kd values for Cu, thus in the case of this metal it had tobe obtained froma calibrationprocedure. The selected value, 2.6104 is,however, of the same order of magnitude of the measured Cu kd (Wenet al., 1999) in USA coastal waters (1.3104).Metal concentrations in the dissolved phase have been dened atthe three main sources, Guadalquivir, Guadiana and OdielTintoestuaries. Values, which are the same used by Beckers et al. (2007),are given in Table 2.The metal transport sub-model is started from backgrounddissolved concentrations, which correspond to the open AtlanticOcean metal concentrations reported in Elbaz-Poulichet et al. (2001).The corresponding background in the solid phase is obtained from themetal kd through Eq. (12). Equations are integrated until steady metaldistributions are obtained in all phases.All the modelling work has been carried out in a sequential way.Thus, the hydrodynamic sub-model is developed rst. Once thatacceptable tides and residual currents are obtained, these are used tosimulate sediment dynamics. Finally, metal transport is includedwhen the sediment sub-model model has been validated.3. Results and discussionA brief description of the hydrodynamic and sediment transportresults is given. Next, model results for metal distributions arediscussed.Fig. 4. Computed surface residual currents. Only one of each 4 vectors is drawn.3.1. HydrodynamicsA map showing the different locations and sampling points2006). The computed current at this position is 0.034m/s. The currentamplitude increases as approaching the Strait entrance, wherecurrents about 0.8 m/s are produced (the color scale in Fig. 3 islimited to 0.5 m/s for more clarity). A similar behaviour is observedfor the S2 tide.The residual surface circulation in the northern GoC is character-ized by a current directed to the SE (Criado-Aldeanueva et al., 2006;Garca-Lafuente and Ruiz, 2007) along the Spanish coast. Thiscirculation is a rather constant pattern during most of the year. Partof the ow enters the Strait of Gibraltar and part is deected to thesouth. The residual circulation computed with the baroclinic model atthe sea surface is presented in Fig. 4. The current is effectively directedto the SE over the continental shelf and part of this ow enters theStrait of Gibraltar. Maximum currents are of the order of 0.3 m/s, inagreement with Garca-Lafuente et al. (2006). The anticyclonic eddyat the east of Faro (see Fig. 2) has been described by Machn et al.(2006). Also, the cyclonic eddy in front of the Strait of Gibraltarappears clearly in the models of Beckers et al. (2007) and Peliz et al.(2007). These last authors have attributed it to the strong convergenceoccurring in this area.Below the surface, the Mediterraneanwaters ow into the Atlanticand mainly direct to the NW (Serra et al., 2005; Ambar and Howe,1979; Johnson et al., 2002). As an example, the computed circulation590 m below the surface is presented in Fig. 5. Only the northern partof the GoC is shown to appreciate details more clearly. These currentsare in agreement with the geostrophic velocities below 400 m,referenced to 300m, provided by Criado-Aldeanueva et al. (2006) andFig. 5. Computed residual currents 590 m below the surface in the northern GoC. Onlyone of each 4 vectors is drawn.3398 R. Periez / Science of the Total Environment 407 (2009) 33923406mentioned in the paper is presented in Fig. 2. Computed tidalconstants have been compared with established values for both M2and S2 tides. Results are shown in Table 3, and it may be seen that,generally, there is a good agreement between both set of data. As anexample, maps showing the M2 tide amplitude and the amplitude ofthe tidal current are presented in Fig. 3. The amplitude of the tide isabout 1 m over all the GoC, decreasing near the Strait of Gibraltar.Associated currents are weak, with amplitudes below 0.10 m/s overmost of the GoC. Indeed at the RAP (Red de Aguas Profundas, SpanishInstitute of Oceanography) buoy position (see Fig. 2), the M2barotropic tidal current is less than 0.03 m/s (Garca-Lafuente et al.,Fig. 6. Computed (lines) and measured (points) salinity awith the calculations by Peliz et al. (2007) for summer. Water velocityis higher close to the Strait and then slows to about 0.1 m/s, inagreement with Ambar and Howe (1979).Temperature and salinity proles in the water column calculatedby the model have been compared with those obtained fromobservations carried out under the TOROS project (Elbaz-Poulichetet al., 2001) in summer 1997. Examples for 4 points shown in Fig. 2may be seen in Fig. 6. There is an acceptable agreement betweenmodel results and experimentally obtained vertical proles of T and S.Finally, two vertical southnorth sections of water salinity arepresented in Fig. 7, at longitudes of 6.87 and 6.05. In the rstnd temperature proles in points indicated in Fig. 2.itude3399R. Periez / Science of the Total Environment 407 (2009) 33923406case, the core of more dense Mediterraneanwater is clearly seen, withthe salinity maximum below 600 m. This ow is aligned with theSpanish continental slope, in agreement with Jungclaus and Mellor(2000) and Peliz et al. (2007). Close to the Strait, the situation may becharacterized as a two-layer exchange ow, with the interface tilteddown southward (Jungclaus and Mellor, 2000).Although seasonally averaged values have been used for openboundary conditions and wind stress, the main features of watercirculation in the GoC are, generally speaking, reproduced by themodel for the considered season.3.2. Suspended matterAs expected, the river-discharged suspended matter is transportedto the southeast along the Spanish coast, which is the residual currentFig. 7. Vertical southnorth salinity sections at longdirection in this area of the GoC. As widely discussed (Gonzalez et al.,2007), there is a dominant eastward transport throughout the entirenorthern GoC. Some authors (Lobo et al., 2004) have postulated theexistence of some westward transport of sediments released by theGuadalquivir River in the inner shelf, although little indication of it hasbeen found in other works (Gonzalez et al., 2007).Maps of suspended matter concentrations at the surface and in thedeepest water layer are presented in Fig. 8. Logarithmic scale is used toappreciate differences. The suspended particle plumes produced bythe three rivers can be clearly seen in the maps. Indeed, concentra-tions of the order of 10 g/m3 are obtained near the river mouths.Concentrations of the order of 101 g/m3 are obtained in part of thenorthern GoC and much smaller values are apparent to the south.Computed surface particle concentrations are in agreement, by orderof magnitude, with measured surface concentrations. Indeed, con-centrations of the order of 101 g/m3 have been measured at thesurface in the northern GoC (Palanques et al., 19861987). Also,concentrations in the range 2045 g/m3 have been detected in theGuadiana River plume (Cravo et al., 2006) over a narrow area (b10 kmfrom the coast). In the deepest water layer concentrations are slightlyhigher than at the surface over most of the northern GoC, inagreement with Palanques et al. (19861987). In some areas as theStrait of Gibraltar and the Spanish continental slope, concentrationsare more signicantly enhanced in the deep water. This is probablydue to erosion produced by stronger tidal currents (in the area of theStrait) and also produced by the Mediterranean water current.A southnorth section of suspended matter concentrations at theAtlantic entrance of the Strait of Gibraltar (longitude6.10) is shownin Fig. 9. The plume of particles released by the rivers may be seen insurface waters at the northern area of the section. Maximumconcentrations, however, are obtained in Mediterranean waters, atdepth about 200 m and along the Spanish continental slope. Thisindicates that erosion is occurring in this area, as commented above.Computed net sedimentation rates (not shown) are larger, asexpected, in the vicinity of the river mouths, where they are of theorder of 103 g/cm2year. Much smaller values are obtained far fromthe river discharge inuence. In particular, there is a region where nosedimentation occurs at thewest of the Strait of Gibraltar. This is due tothe strong currents in theMediterranean outowwater (Fig. 5), whichkeep particles in suspension. Nevertheless, all these results have to beinterpreted with care since several approximations have implicitlys 6.87 and 6.05 (left and right respectively).been made in the model: there are variations in water exchangesthrough the Strait of Gibraltar, as already discussed in previousmodelling works (Periez, 2008). Of course, these variations willaffect the SR. Although constant suspended matter concentrationshave been dened in the three estuaries, there will also be seasonalvariations depending on pluviometry, for instance. Only ne particlesare considered in the model, and bed load transport is neglected. Thisapproximationwill lead to lower sedimentation rates. The same occursdue to the fact that organic particles are not included. Finally,atmospheric deposition events of particles coming from the SaharaDesert (Fabres et al., 2002) have not been considered in the modelsince they cannot be easily quantied (Periez, 2008). It must betaken into account that suspended particles will affect metal transportonly close to the coast, given the extremely low suspended matterconcentrations which are measured offshore (Palanques et al., 19861987), and river supply is themain sediment source in this coastal area.3.3. Heavy metal transport3.3.1. Metal distributionsAs commented before, the model has been applied to simulate thetransportof threemetals in theGoC:Zn,CuandNi. Pathwaysofmetals areessentially the sameas for suspendedparticles. Thus, theyare transportedto the east, along the Spanish coast, until the Strait of Gibraltar.The concentrations of several metals have been measured in thene sediment fraction (b63 m) along the Spanish coast from the3400 R. Periez / Science of the Total Environment 407 (2009) 33923406Guadiana to the Guadalquivir mouths (Sainz and Ruiz, 2006; Morilloet al., 2004) and also at some points closer to the Strait of Gibraltar(Riba et al., 2002). Samples were collected at an approximate distanceof 500 m from the shoreline. A comparison of measured (points) andcomputed (lines) metal concentrations along the northern coast ofthe model domain can be seen in Fig. 10. In general, there is a goodagreement between the measured and calculated concentrations,although in the case of Ni the model produces a concentration peak inthe area of the OdielTinto mouth that is not apparent in measure-ments. This discrepancy could be caused, at least partially, by adifferent nature of particles released from this river, which affects thepartition on Ni (although does not in the cases of Zn and Cu). Also, it isFig. 8. Computed surface suspended matter concentrations (g/m3) in logarithmic scaleat the surface (top) and bottom (down) of the water column.Fig. 9. Vertical southnorth section of suspended matter concentrations (g/m3) at theAtlantic entrance of the Strait (longitude 6.10).possible that the source term is overestimated in the case of Ni for theOdielTinto River. A clear conclusion cannot be obtained.Metal concentrations are very low westward from the GuadianaRiver. There is an increase in concentrations here since, as has alreadybeenmentioned, the three rivers drain the Iberian Pyrite Belt.Maximumconcentrations exist in the mouth of the OdielTinto rivers. Althoughriver ows are much smaller than those of the Guadiana andGuadalquivir, OdielTinto rivers are considerably more contaminated(Gonzalez et al., 2007) and, indeed, they have been recognized as themain source of metals along the coast (Morillo et al., 2004).There is not an appreciable metal enhancement in the area of theGuadalquivir mouth (except in the case of Ni where both measure-ments and calculations show such an enhancement). Metal concen-trations at the east of the OdielTinto rivers decrease much slowerthan at the west of the Guadiana River, which is due to the residualcurrents in the shelf, which ow towards the Strait of Gibraltar.Indeed, it has already been found (Elbaz-Poulichet et al., 2001) thatcoastal waters transport dissolved metals (which contaminate bedsediments as metals travel over them) from the OdielTinto rivers to aFig. 10. Measured (circles) and computed (lines) metal concentrations in sedimentsalong the northern coast of the model domain. The three arrows indicate the location ofthe Guadiana, OdielTinto and Guadalquivir mouths (from left to right).Fig. 11.Measured and computed (color scale) metal concentrations (nM) in surface waters of the northern GoC. A contour plot showing surface dissolved Zn concentrations (nM) inJune 1997 from experimental data (from Achterberg et al., 1999) is also shown (right of the bottom row).3401R. Periez / Science of the Total Environment 407 (2009) 33923406distribution coefcient kd through Eq. (13), as already described. Themodel sensitivity to changes in this last parameter has beeninvestigated since the kd is poorly dened and presents a high naturalvariability depending on environmental conditions (mainly pH,temperature and salinity, although other parameters as light intensitymay be relevant as well).Obviously, a higher kd means that the substance has a higherafnity to be xed to the solid phase and vice versa. Thus, surfacesediment metal concentrations are linearly correlated with the kd, asshould be expected from the model formulation.The distribution of metals in coastal sediments obtained withseveral kd values have been computed and are presented in Fig. 13a inthe case of Zn. The value 2.0105 has beenmeasured in coastal watersof the USA (Wen et al., 1999). A smaller value, by one order ofmagnitude, has also been tested (2.0104). Effectively, too high metalconcentrations, when compared with empirical data, are obtainedwith 2.0105. On the other hand, concentrations are too low if a kd of2.0104 is used. The best agreement between measured andcomputedmetal concentrations is obtainedwith the kd recommendedby the IAEA (2004), which is indicated in Fig. 13a as the nominalsimulation.The kd value in the case of Cu was obtained from a calibrationFig.12. Vertical southnorth sections of dissolvedmetal concentrations (nM) for Cu andNi at the Atlantic entrance of the Strait (longitude 6.10).3402 R. Periez / Science of the Total Environment 407 (2009) 33923406distance of more than 200 km. Moreover, it has been found that theserivers constitute a source of natural radionuclides into the Mediterra-nean through the Strait of Gibraltar (Gasc et al., 2002).Concentrations of dissolved metals have been measured in thenorthern GoC in the frame of the TOROS project (Elbaz-Poulichet et al.,2001) in summer 1997. A comparison between measured andcomputed surface concentrations may be seen in Fig. 11. In all cases,dissolved metal concentrations decrease quickly with distance fromthe coast. The highest concentrations are obtained in the mouth of theOdielTinto rivers, obviously as in the case of bed sediments. Theplume of dissolved metals reaches the Strait of Gibraltar, as has beenobtained frommeasurements and has been commented above (Elbaz-Poulichet et al., 2001; Gasc et al., 2002). Themaps presented in Fig.11are in good agreement with contour plots obtained from empiricaldata (Achterberg et al., 1999), which show essentially the samebanded structure. As an example, such contour plot is presented aswell in Fig. 11 for the case of Zn. It may be clearly seen that, effectively,the impacts from river outow are restricted to a narrow band alongthe shore. Samples were not collected in the coastal area from Cadiz tothe Strait, thus the impact is apparently restricted in the experimentalcontours of Fig. 11 to the zone located to the north of Cadiz.Dissolved metal concentrations deeper in the water column (notshown) are rather uniform (for instance in the range 35 nM for thecase of Zn) since deeperwaters (depth N50m) are not affected by riverdischarges. Indeed, it has been found that the main core of metalenriched water extends to about the 50 m isobath (Achterberg et al.,1999). Along Spain, this isobath is typically some20 km from the shore.Vertical southnorth sections of dissolved metal concentrations (CuandNi) at theAtlantic entrance of the Strait (longitude6.10) are shownin Fig. 12. In both cases, the metal plumes coming from the rivers areclearly visible above some 50m depth in the northern part of the section.Ni concentration in theMediterraneanwater is higher than inwater aboveit. The inverse situation is observed for Cu. These results are in agreementwith data presented in Gmez (2003), which show this effect.Computed background metal concentrations in suspended matterare of the order of 102 nmol/g,103 nmol/g and102 nmol/g for Cu, Zn andNi respectively. In Atlantic Shelf waters (English Channel) particulate Znconcentrations range 11604900 nmol/kg (Cobelo-Garca et al., 2005).In the case of Cu this range is 60380 nmol/kg. Information was notfound for Ni. Computed particulate backgrounds seem to be inreasonable agreement with measurements in the Atlantic Ocean.Model results, however, have to be interpreted with care (ascommented before for suspended matter transport) since river plumesof suspended particles and hence contaminants are sensitive to changesin wind speed and direction as well as they depend on pluviometry.Moreover, other meteorological conditions, such as atmosphericpressure differences between the Atlantic and Mediterranean, caninduce ow variations through the Strait of Gibraltar. East winds, whichmay be particularly strong in the Strait during spring and summer, caninhibit the water input from the GoC into the Mediterranean andconsequently retain pollutants in the GoC. On the other hand, organicparticles are not included and they are thought to play a specic role onsorption reactions on suspended matter since it might provide sorptionsites due either to their chemical structure or because of their polar (ornon-polar) character (Duursma and Carroll,1996). Finally, other sourcesofmetals in theGoChave not been included in this study, as such relatedto the intensive industrial activity in Cadiz Bay. These effects make thecomparison of computed metal concentrations in the water columnwith the corresponding measurements specially difcult (as alreadydiscussed by Dyke, 2001). However, bed sediments integrate all thisvariability and the generallygoodbehaviour of themodel in reproducingsediment contamination gives some condence on the model.3.3.2. Sensitivity analysisThemain parameter which describes the geochemical behaviour ofmetals is the exchange velocity , obtained from the equilibrium exercise. Thus, the kd value was changed, by trial and error, until a3403R. Periez / Science of the Total Environment 407 (2009) 33923406good agreement between model results and observation wasobtained. A summary of this exercise may be seen in Fig. 13b. Ingeneral, the best agreement was obtained with kd=2.6104. Thevalue measured by Wen et al. (1999), 1.3104, produces sediment Cuconcentrations lower than measured values, and, on the other hand,too high Cu concentrations are computed if 5.2104 is used. Thesensitivity analysis is not shown in the case of Ni, but similar resultsare obtained. The IAEA (2004) kd value, which is also similar to thevalue measured byWen et al. (1999), produces better results than if itis reduced or increased by a factor two.The sensitivity of the formulation of the watersediment metalinteraction processes to other parameters as particle size and density,sediment mixing depth, fraction of small particles and the correctionfactor has already been studied in detail (Periez, 2002, 2004) andwill not be repeated here.4. ConclusionsThe transport of heavy metals in the GoC has been studied bymeans of numerical modelling. Water circulation has been obtainedfrom two models working on the same domain. A 2D depth-averagedbarotropic model provides tides. The residual circulation is obtainedFig. 13. Model sensitivity to the distribution coefcient. (a): Zn. Results in Fig. 10 arepresented for three kd values. The nominal simulation corresponds to kd=7104.(b): Cu. In this case the nominal simulation corresponds to kd=2.6104.from a full 3D baroclinic model. This model is forced, at its openboundaries, by climatological data.Computed tides have been compared with observations, beingin good agreement with these. The residual circulation shows thewell-known southeast circulation along the Spanish coast and thedeeper Mediterranean water ow directed from the Strait ofGibraltar to the northwest. Computed vertical proles of tem-perature and salinity have been compared with observed prolesat some points in the GoC. Both set of data are, in general, in goodagreement.Results of the hydrodynamic models are used by a 3D sedimenttransport model which includes advection/diffusion of particles,settling, deposition and erosion of the sediment. The settlingvelocity is obtained from Stoke's law, and deposition and erosionare described using standard equations which include the erod-ability constant concept as well as erosion and deposition criticalstresses. The total bed stress is obtained from the tidal and residualcurrents at the bottom. Sources of particles are the three main riversdischarging in the GoC: Guadiana, OdielTinto and Guadalquivir.Only the lithogenic particle fraction is considered and bed-loadtransport is not included in the model. The model shows that theriver plumes are directed to the east, transported by the residualcurrents as should be expected. Suspended particle concentrationsprovided by the model are, generally speaking, in agreement withobservations.The three rivers mentioned above constitute a source of heavymetals to the GoC since they drain the Iberian Pyrite Belt. A modelwhich simulates metal transport has been developed. It usescomputed currents and suspended particle concentrations andsedimentation rates. Metal exchanges between the dissolved phase,suspended matter and bed sediments are included. These processesare formulated using kinetic transfer coefcients. The model has beenapplied to threemetals: Zn, Cu and Ni. Measuredmetal concentrationsin bed sediments collected along the Spanish coast are in agreementwith calculations. Levels of computed dissolved metal concentrationsin surface water are also in agreement with measurements. The pathfollowed by metals is the same as suspended particles, thus adominant southeast transport is observed. Indeed, metal plumesreach the Strait of Gibraltar and these rivers constitute a source ofheavy metals into the Mediterranean Sea. The behaviour of the threestudied metals is similar since so they are their correspondingdistribution coefcients, which vary within one order of magnitude.The described model, once it has been checked that provides realisticresults, may be used as a predictive tool to assess the effects ofaccidental or deliberate metal discharges into the GoC from rivers orother sources.With respect to future improvements in the model, it does notseem realistic, at this moment, to go deeper into biogeochemicalprocesses (effects of pH, redox conditions, bioturbation, organicparticles and biota etc) due to the lack of eld data which could beused to parameterize such processes and also to test model results.Nevertheless, it would of course be desirable to improve the model insuch directions, specially with respect to the inclusion of organicparticles. On the other hand, it would be interesting to modify themodel in such a way that could be run under non-steady conditions.The objective would be to study the effects in metal distributions ofseasonally changingwater current elds and of changes in particle andheavy metal discharges from rivers originating, for instance, fromheavy rain episodes.AcknowledgementsWork supported by the Research Project of Excellence RNM-419Tcnicas Ultrasensibles para la Determinacin de Radionucleidos enMuestras Ambientales, Junta de Andaluca, Spain. The author isindebted to the Spanish Ministerio de Educacin y Ciencia for a3404 R. Periez / Science of the Total Environment 407 (2009) 33923406fellowship to stay during three months at the University of Wales,Bangor, where part of this work was carried out.Appendix A. HydrodynamicsThe depth-averaged hydrodynamic equations, used for tidal calcu-lations, are:@@t+@@xHu + @@yHv = 0 14@u@t+ u@u@x+ v@u@y+ g@@x v + uH= A@2u@x2+@2u@y2 !15@v@t+ u@v@x+ v@v@y+ g@@y+ Xu +vH= A@2v@x2+@2v@y2 !16where u and v are the depth averaged water velocities along the xand y axis, h is the depth of water below the mean sea level, is thedisplacement of the water surface above the mean sea levelmeasured upwards, H=h+ is the total water depth, is theCoriolis parameter (=2wsin , where w is the Earth rotationalangular velocity and is latitude), g is acceleration due to gravity, is a mean value of water density and A is the horizontal eddyviscosity. u and v are friction stresses that have been written interms of a quadratic law:u = kuu2 + v2pv = kvu2 + v2p 17where k is the bed friction coefcient.The full 3D hydrodynamic equations including the terms corre-sponding to density gradients are written in the hydrostatic andBoussinesq approximations as:@@t+@@xh + Zhudz +@@yh + Zhvdz = 0 18@u@t+ u@u@x+ v@u@y v + g @@x+g0Zz@w@xdz =@@zK@u@z + A@2u@x2+@2u@y2 !19@v@t+ u@v@x+ v@v@y+ Xu + g@@y+g0Zz@w@ydz =@@zK@v@z + A@2v@x2+@2v@y2 !20where w is water density, 0 is a reference density, and K and A arethe vertical and horizontal eddy viscosities respectively.The vertical component of the water velocity, w, is obtained fromthe continuity equation:@u@x+@v@y+@w@z= 0: 21The water density is derived from an equation of state relatingdensity to salinity and temperature:w = 0 1 T T0 + S S0 22where S is salinity, T is temperature, =2.14104 and =7.45104.The reference salinity is taken as 0=999.7 kg/m3 at S0=0 andT0=10 C.Water salinity is determined fromanadvectiondiffusion equation:@S@t+ u@S@x+ v@S@y+ w@S@z= A@2S@x2+@2S@y2 !+@@zK@S@z 23and a similar equation is used for temperature:@T@t+ u@T@x+ v@T@y+ w@T@z= A@2T@x2+@2T@y2 !+@@zK@T@z : 24Vertical eddy viscosity is determined from a 1-equation turbulencemodel. The equation for the turbulent kinetic energy E is:@E@t= K@u@z 2+@v@z 2 + 0@@zK@E@z e + g0K@@z: 25The rst term in the right side of the equation representsgeneration of turbulence by the vertical shear, the second term isdiffusion of turbulence and the last term is loss of turbulence bybuoyancy (conversion of kinetic energy into potential energy). represents dissipation of turbulence, that is written as:e = C1E3=2 26where is a mixing length and C1 a numeric coefcient. The verticalviscosity is nally written as a function of energy as:K = C0E1=2 + t 27whereC0 is a numeric coefcient andt is a background value of viscosity,that is the minimum possible value that it may have. The values of thenumeric constants appearing above are: 0=0.73, C0=C1/4, C1=C03 andC=0.046. The background viscosity is xed as t=104 m2/s.The mixing length is derived from an algebraic expression:=11=1 + 1=228with1 = z + z0 + h e1z + hh 292 = zs z 30where =0.4 is the von Karman's constant, 1=2.0 and zs and z0are the roughness lengths of the sea surface and bottom respectively.Appendix B. Suspended matter transportThe equationwhich provides the suspended matter concentration,m, is:@m@t+ u@m@x+ v@m@y+ w ws @m@z= A@2m@x2+@2m@y2 !+@@zK@m@z31where ws is the settling velocity of suspended particles. The depo-sition and erosion terms are incorporated into the sea bed boundarycondition of the equation. The deposition rate is written as:DP = wsm b 1bcd 32wherem(b) is particle concentration evaluated at the sea bottom, b isbottom stress due to tides and the residual current, and cd is a critical3405R. Periez / Science of the Total Environment 407 (2009) 33923406deposition stress above which no deposition occurs since particles aremaintained in suspension by water turbulence.The settling velocity is determined from Stokes's law:ws = wwgD218m33where and D are suspended particle density and diameter res-pectively and is kinematic viscosity of water. The erosion rate iswritten in terms of the erodability constant:ER = Efbce 1 34where E is the erodability constant, f gives the fraction of ne particlesin the bed sediment and ce is a critical erosion stress below which noerosion occurs. The model can also calculate sedimentation rates (SR)as the balance between the deposition and erosion terms.ReferencesAchterberg EP, Braungardt C, Morley NH, Elbaz-Poulichet F, Leblanc M. Impact of LosFrailes mine spill on riverine, estuarine and coastal waters in southern Spain. WaterRes 1999;33:338794.Ambar I, Howe MR. Observations of the Mediterranean outow. 2, the deep circulationin the vicinity of the Gulf of Cadiz. Deep-Sea Res 1979;26A:55568.Batteen ML, Martinez JR, Bryan DW, Buch EJ. A modelling study of the coastal easternboundary current system off Iberia and Morocco. J Geophys Res 2000;105(C6):1417395.Beckers JM, Achterberg EP, Braungardt Ch. Comparison of high spatial resolution tracemetal distributions with model simulations for surface waters of the Gulf of Cadiz.Estuar Coast Shelf Sci 2007;74:599609.Cancino L, Neves R. Hydrodynamic and sediment suspension modelling in estuarinesystems. Part I: description of the numerical models. J Mar Syst, 1999;22:10516.Cetina M, Rajar R, Povinec P. Modelling of circulation and dispersion of radioactivepollutants in the Japan Sea. Oceanol Acta 2000;23:81936.Chen K, Jiao JJ. Metal concentrations and mobility in marine sediment and groundwaterin coastal reclamation areas: a case study in Shenzhen, China. Environ Pollut2008;151:57684.Cobelo-Garca A, Prego R, De CastroM.Metal distributions and their uxes at the coastalboundary of a semi-enclosed ria. Mar Chem 2005;97:27792.Cravo A, Madureira M, Felicia H, Rita F, Bebianno MJ. Impact of outow from theGuadiana River on the distribution of suspended particulatematter and nutrients inthe adjacent coastal zone. Estuar Coast Shelf Sci 2006;70:6375.Criado-Aldeanueva F, Garca-Lafuente J, Vargas JM, del Ro J, Vzquez A, Reul A, et al.Distribution and circulation of water masses in the Gulf of Cadiz from in situobservations. Deep-Sea Res II 2006;53:114460.Cuong DT, Karuppiah S, Obbard JP. Distribution of heavy metals in the dissolved andsuspended phase of the sea-surface microlayer, seawater column and in sedimentsof Singapore's coastal environment. Environ Model Assess 2008;138:25572.Dick S, Schonfeld W. Water transport and mixing in the North Frisian Wadden Sea.Results of numerical investigations. Ger J Hydrogr 1996;48:2748.Duursma EK, Carroll J. Environmental compartments. Berlin: Springer; 1996.Dyke PPG. Coastal and Shelf Sea modelling. The Netherlands: Kluwer; 2001.Eisma D. Suspended matter in the aquatic environment. Berlin: Springer-Verlag; 1993.Elbaz-Poulichet F, Morley NH, Beckers JM, Nomerange P. Metal uxes through the Straitof Gibraltar: the inuence of the Odiel and Tinto Rivers (SW Spain). Mar Chem2001;73:193213.Fabres J, Calafat A, Sanchez-Vidal A, Canals M, Heussner S. Composition and spatio-temporal variability of particle uxes in the western Alborn gyre, MediterraneanSea. J Mar Syst 2002;33/34:43156.Fernndez-Caliani JC, Ruiz-Muoz F, Galn E. Clay mineral and heavy metaldistributions in the lower estuary of Huelva and adjacent Atlantic shelf, SWSpain. Sci Total Environ 1997;198:181200.Freitas PS, Abrantes F. Suspended particulate matter in the Mediterranean water at theGulf of Cadiz and off the southwest coast of the Iberian Peninsula. Deep-Sea Res II2002;49:424561.Garca-Lafuente J, Ruz J. The Gulf of Cadiz pelagic ecosystem: a review. Prog Oceanogr2007;74:22851.Garca-Lafuente J, Delgado J, Criado-Aldeanueva F, BrunoM, del Ro J, Vargas JM.Water masscirculationon thecontinental shelf of theGulf of Cadiz.Deep-SeaRes II 2006;53:118297.Gasc C, Antn MP, Delfanti R, Gonzlez AM, Meral J, Pappuci C. Variation of the activityconcentrations anduxes of natural (210Po, 210Pb) and anthropogenic (239,240Pu, 137Cs)radionuclides in the Strait of Gibraltar (Spain). J Environ Radioact 2002;62:24162.Gmez F. The role of the exchanges through the Strait of Gibraltar on the budget ofelements in the western Mediterranean Sea: consequences of human-inducedmodications. Mar Pollut Bull 2003;46:68594.Gmez-Parra A, Forta JM, Delvalls TA, Senz I, Riba I. Early contamination by heavymetals of the Guadalquivir estuary after the Aznalcollar Mining Spill (SW Spain).Mar Pollut Bull 2000;40:111523.Gonzalez R, Araujo MF, Burdloff D, Cachao M, Cascalho J, Corredeira C, et al. Sedimentand pollutant transport in the northern Gulf of Cadiz: a multi-proxy approach. J MarSyst 2007;68:1-23.Harms IH. Modelling the dispersion of 137Cs and 239Pu released from dumped waste inthe Kara Sea. J Mar Syst 1997;13:1-19.Hernndez-Molina FJ, Llave E, Stow DAW, Garca M, Somoza L, Vzquez JT, et al. Thecontourite depositional system of the Gulf of Cadiz: a sedimentary model related tothe bottom current activity of the Mediterranean outow water and its interactionwith the continental margin. Deep-Sea Res II 2006;53:142063.IAEA. Sediment distribution coefcients and concentration factors for biota in themarine environment. Technical Reports Series 2004;422 (Vienna).Jensen TG. Open boundary conditions in stratied ocean models. J Mar Syst1998;16:297322.Johnson J, Stevens I. A ne resolution model of the eastern North Atlantic between theAzores, the Canary Islands and the Gibraltar Strait. Deep-Sea Res I 2000;47:87599.Johnson J, Ambar I, Serra N, Stevens I. Comparative studies of the spreading ofMediterranean water through the Gulf of Cadiz. Deep-Sea Res II 2002;49:417993.Jungclaus JH, Mellor GL. A three-dimensional model study of the Mediterraneanoutow. J Mar Syst 2000;24:4166.Kowalick Z, Murty TS. Numerical modelling of ocean dynamics. Singapore: WorldScientic; 1993.Li Q, Wub Z, Chua B, Zhanga N, Caia S, Fanga J. Heavy metals in coastal wetlandsediments of the Pearl River Estuary, China. Environ Pollut 2007;149:15864.Liu YP, Millward GE, Harris JRW. Modelling the distributions of dissolved Zn and Ni inthe Tamar Estuary using hydrodynamics coupled with chemical kinetics. EstuarCoast Shelf Sci 1998;47:53546.LiuWC, Hsu MH, Kuo AY. Modelling of hydrodynamics and cohesive sediment transportin Tanshui River estuarine system, Taiwan. Mar Pollut Bull, 2002a;44:107688.Liu JT, Chao S, Hsu RT. Numerical modeling study of sediment dispersal by a river plume.Cont Shelf Res 2002b;22:174573.Lobo FJ, Snchez R, Gonzlez R, Dias JMA, Hernndez-Molina FJ, Fernndez-Salas LM, et al.Contrasting styles of the Holocene highstand sedimentation and sediment dispersalsystems in the northern shelf of the Gulf of Cadiz. Cont Shelf Res 2004;24:46182.Lumborg U, Windelin A. Hydrography and cohesive sediment modelling: application tothe Romo Dyb tidal area. J Mar Syst, 2003;38:287303.Machado A, Rocha F, Gomes C, Dias J. Distribution and composition of suspendedparticulate matter in Guadiana estuary (southwestern Iberian Peninsula). J CoastRes 2007;50:10405.Machn F, Pelegr JL, Marrero-Daz A, Laiz I, Ratsimandresy AW. Near-surface circulationin the southern Gulf of Cadiz. Deep-Sea Res II 2006;53:116181.Marn-Guirao L, Lloret J, Marin A. Carbon and nitrogen stable isotopes and metalconcentration in food webs from a mining-impacted coastal lagoon. Sci TotalEnviron 2008;393:11830.Monte L, Hakason L, Periez R, Laptev G, Zheleznyak M, Maderich V, et al. Experiencesfrom a case study of multi-model application to assess the behaviour of pollutantsin the DnieperBug estuary. Ecol Model 2006;195:24763.Morillo J, Usero J, Gracia I. Heavy metal distribution in marine sediments from thesouthwest coast of Spain. Chemosphere 2004;55:43142.Nyffeler UP, Li YH, Santschi PH. A kinetic approach to describe trace elementdistribution between particles and solution in natural aquatic systems. GeochimCosmochim Acta 1984;48:151322.NOAA. Computer Applications to Tides in the National Ocean Survey. Supplement toManual of Harmonic Analysis and Prediction of Tides (Special Publication No. 98).National Ocean Service, National Oceanic and Atmospheric Administration. U.S.Department of Commerce; 1982. January 1982.Palanques A, Plana F, Maldonado A. Estudio de la materia en suspensin en el Golfo deCdiz. Acta Geol Hisp 19861987;2122:4917 (in Spanish).Peliz A, Dubert J, Marchesielo P, Teles-Machado A. Surface circulation in the Gulf ofCadiz: model and mean ow structure. J Geophys Res 2007;112:C11015 (20 pages).Periez R. Three dimensional modelling of the tidal dispersion of non conservativeradionuclides in the marine environment. Application to 239,240Pu dispersion in theeastern Irish Sea. J Mar Syst 1999;22:3751.Periez R. Modelling the tidal dispersion of 137Cs and 239,240Pu in the English Channel.J Environ Radioact 2000;49:25977.Periez R. The enhancement of 226Ra in a tidal estuary due to the operation of fertilizerfactories and redissolution from sediments: experimental results and a modellingstudy. Estuar Coast Shelf Sci 2002;54:80919.Periez R. Kinetic modelling of the dispersion of plutonium in the eastern Irish Sea:two approaches. J Mar Syst 2003;38:25975.Periez R. On the sensitivity of a marine dispersion model to parameters describing thetransfers of radionuclides between the liquid and solid phases. J Environ Radioact2004;73:10115.Periez R. Modelling the dispersion of radionuclides in the marine environment.Heidelberg: Springer-Verlag; 2005a.Periez R. Modelling the transport of suspended particulate matter by the Rhone Riverplume (France). Implications for pollutant dispersion. EnvironPollut 2005b;133:35164.Periez R. A modelling study on 137Cs and 239,240Pu behaviour in the Alborn Sea,western Mediterranean. J Environ Radioact 2008;99:694715.Periez R, Pascual-Granged A. Modelling surface radioactive, chemical and oil spills inthe Strait of Gibraltar. Comput Geosci 2008;34:16380.Periez R, Abril JM, Garca-Len M. Modelling the suspended matter distribution in anestuarine system: application to the Odiel river in southwest Spain. Ecol Model1996;87:16979.Periez R, Absi A, Villa M, Moreno HP, Manjn G. Self-cleaning in an estuarine areaformerly affected by 226Ra anthropogenic enhancements: numerical simulations.Sci Total Environ 2005;339:20718.Prandle D, Ballard G, Flatt D, Harrison J, Jones SE, Knight PJ, et al. Combining modellingand monitoring to determine uxes of water, dissolved and particulate metalstrough the Dover Strait. Cont Shelf Res 1996;16:23757.Pugh DT. Tides, surges and mean sea level. Chichester: Wiley; 1987.Riba I, Del Valls TA, Forja JM, Gmez-Parra A. Inuence of the Aznalcllar minning spillon the vertical distribution of heavy metals in sediments from the Guadalquivirestuary (SW Spain). Mar Pollut Bull 2002;44:3947.Sainz A, Ruiz F. Inuence of the very polluted inputs of the TintoOdiel system on theadjacent littoral sediments of southwestern Spain: a statistical approach. Chemo-sphere 2006;62:161222.Schwiderski EW.Ocean tides, part1: globalocean tidal equations.MarGeod1980a;3:161217.Schwiderski EW. Ocean tides, part 2: a hydrodynamical interpolation model. Mar Geod1980b;3:21955.Scott EM. Modelling radioactivity in the environment. UK: Elsevier; 2003.Serra N, Ambar I, Kase RH. Observations and numerical modelling of the Mediterraneanoutow splitting and eddy generation. Deep-Sea Res II 2005;52:383408.Suntornvongsagula K, Burkeb DJ, Hamerlynckb EP, Hahna D. Fate and effects of heavymetals in salt marsh sediments. Environ Pollut 2007;149:7991.Tappin A, Burton JD, Millward GE, Statham PJ. A numerical transport model for predictingthe distributions of Cd, Cu, Ni, Pb and Zn in the southern North Sea: the sensitivity ofmodel results to the uncertainties in the magnitudes of metal inputs. J Mar Syst1997;13:173204.Tattersall GR, Elliott AJ, Lynn NM. Suspended sediment concentrations in the Tamarestuary. Estuar Coast Shelf Sci 2003;57:67988.Tejedor L, Izquierdo A, Kagan BA, Sein DV. Simulation of the semidiurnal tides in theStrait of Gibraltar. J Geophys Res 1999;104:1354157.Tsimplis MN, Proctor R, Flather RA. A two dimensional tidal model for theMediterranean Sea. J Geophys Res 1995;100:1622339.Valds J, RommD, Alvarez G, Ortlieb l, GuiezM. Metals content in surfacewaters of anupwelling system of the northern Humboldt Current (Mejillones Bay, Chile). J MarSyst 2008;71:1830.Vested HJ, Baretta JW, Ekebjrg LC, Labrosse A. Coupling of hydrodynamical transportand ecological models for 2D horizontal ow. J Mar Syst 1996;8:25567.Wen LS, Santschi P, Gill G, Paternostro C. Estuarine trace metal distributions inGalvelstone Bay: importance of colloidal forms in the speciation of the dissolvedphase. Mar Chem 1999;63:185212.Wolanski E, Chicharo L, Chicharo MA, Morais P. An ecohydrology model of the Guadianaestuary (South Portugal). Estuar Coast Shelf Sci 2006;70:13243.Wu Y, Falconer R, Lin B. Modelling trace metal concentration distributions in estuarinewaters. Estuar Coast Shelf Sci 2005;64:699709.Yanagi T. Coastal oceanography. The Netherlands: Kluwer; 1999.3406 R. Periez / Science of the Total Environment 407 (2009) 33923406Environmental modelling in the Gulf of Cadiz: Heavy metal distributions in water and sedimentsIntroductionModel descriptionHydrodynamic modelsSediment transportMetal transportNumerical solution and parametersResults and discussionHydrodynamicsSuspended matterHeavy metal transportMetal distributionsSensitivity analysisConclusionsAcknowledgementsHydrodynamicsSuspended matter transportReferences


View more >