The influence of hydrodynamic processes on zooplankton ...doglioli/Qiu_et...processes using bio-physical models provides a rational to understand the in-fluence of such physical

The influence of hydrodynamic processes on

zooplankton transport and distributions in the

North Western Mediterranean: estimates from

a Lagrangian model

Z.F.Qiu a,b, A.M.Doglioli a, Z.Y.Hu a, P.Marsaleix c,F.Carlotti a,∗

aLaboratoire d’Oceanographie Physique et Biogeochimie, Aix Marseille Universite,CNRS, UMR 6535, Marseille, France

bKey Laboratory of Ocean Circulation and Waves, Chinese Academic of Sciences,Qingdao, China

cLaboratoire d’Aerologie, 14 Avenue Edouard Belin, Toulouse, France

Abstract

A Lagrangian module has been developed and coupled with the 3D circulationmodel Symphonie to study the influence of hydrodynamic processes on zooplank-ton transport and distributions in the North Western Mediterranean (NWM). Theindividuals are released every 3 days from March to August 2001 and tracked for40 days in two initial areas: the surrounding of the DYFAMED sampling station inthe central Ligurian sea and the Rhone river plume, which are either considered aspassive particles or able to perform diel vertical migrations (DVM) with a simpleswimming pattern. The simulations suggest strong seasonal patterns in the distribu-tions of the individuals released at DYFAMED sampling station. Individuals couldbe spread all over the NWM basin after 40 days but different pattern s occurs fol-lowing the season, the initial released depth of the individuals, and the capacity ofDVM or not. An offshore-shelf transport only occurs in April and May with particlesending in the Gulf of Lions (GoL) in low concentrations. For the other months, thenorth current can be properly considered as a barrier for particles entering into theGoL from the offshore sea. At the end of the 40 days, passive individuals released inthe Rhone river plume are spread on the GoL shelf or in the Catalan sea. Followingthe season between a quarter to a half of the initial released individuals stay in theGoL. Simple DVM behavior does not increase the retention on the shelf.

Key words: Lagrangian, zooplankton, North Western Mediterranean, Gulf ofLions

Preprint submitted to Elsevier 30 September 2008

1 Introduction

In the last decades, it becomes clear that physical processes are importantdrivers of population dynamics in the oceans (Miller et al., 1998; Robinsonet al., 2005). Zooplankton organisms are critically dependent on their physi-cal environments but they are not passive particles as their definition states.Indeed, they are able of vertical swimming behavior which could influencetheir spatio-temporal distribution. At each phase of their developments, theyhave to find prey and avoid predators, and at the adult stage they have to findmate and reproduce. Zooplankton transport in a variety of physical conditionscan be considered as the combination of two closely linked aspects. The firstinvolves zooplankton transport by non-stationary flow field and the secondinvolves the behavioral response of zooplankton organisms, mainly swimming,to the changes of environmental conditions.

Investigating the relationships between particle distribution and physicalprocesses using bio-physical models provides a rational to understand the in-fluence of such physical structures on the dynamics of local ecosystems (Levyet al., 2000; Oschlies, 2002).

Lagrangian particle-tracking models coupled with hydrodynamic models areparticularly efficient tools to examine the role played by various physical pro-cesses, to study transport processes over an entire basin and to simulate com-plex and interactive processes acting at different scales (e.g. Miller et al., 1998;Blanke et al., 1999; Falco et al., 2000; Guizien et al., 2006; Speirs et al., 2006;Lett et al., 2007). Recent developments of lagrangian modeling have high-lighted the linkages between physical structures, zooplankton behavior andmarine productivity in region such as the Benguela ecosystem (Mullon et al.,2003; Lett et al., 2007) or the Georges bank (Miller et al., 1998). Cianelli et al.(2007) simulated for particle exchange in the Gulf of Lions (hereinafter GoL)using a Lagrangian approach coupled with the three-dimensional (3D) circu-lation Symphonie model. They found that particle distributions are stronglyrelated to the mesoscale and sub-mesoscale hydrodynamic structures on theshelf and to the offshore circulation associated with the North Current (here-inafter NC).

With the particle-tracking approach, one can attain trajectories and dis-tributions of zooplankton under different physical conditions, and add thezooplankton behavior to transport in comparison with the simple transport ofpassive zooplankton individuals.

Diel vertical migration (hereinafter DVM) is the most common swimming

∗ Corresponding author.Email address: [email protected] (F.Carlotti).

2

behavior of zooplankton. The potential effect of DVM on larval transport hasbeen clearly demonstrated (Batchelder et al., 2002; Alexei et al., 2007; Carret al., 2007) or not (Jenkins et al., 1999).

The North Western Mediterranean (hereinafter NWM) and particularly theGoL, is a favorable area to study the influence of water circulations and estu-arine inputs on biological activity and distributions.

In the NWM, large scale circulation is dominated by the NC that forms inthe Ligurian sea where the Western Corsica Current merges with the EasternCorsica Current. From the Ligurian sea, the NC flows along the continentalshelf to the Catalan sea. Sometimes a branch of the NC can intrude into thecontinental shelf of the GoL (Millot and Wald, 1980; Estournel et al., 2003;Petrenko, 2003; Petrenko et al., 2005), from the northeastern side splittinginto two branches when it flows across the 200m-deep isobath. A southern onefollows the continental slope, and a northern one occasionally enters the shelfzone following a north-western direction. The main hydrodynamic featuresof the GoL have been previously researched, by experimental and numericalways (e.g. Millot, 1999; Estournel et al., 2003; Petrenko, 2003; Andre et al.,2005; Petrenko et al., 2005). The shelf circulation in the GoL is complex dueto the combined effects of various forcings, which are mainly the strong windsblowing from the north (Mistral) and from the northwest (Tramoutand), theNC, the Rhone river discharges and the complex topography, characterized byseveral canyons.

The NWM sub-basin is one of the most productive areas in the Mediter-ranean owing to important river discharges from the Rhone and Ebre riversand strong wind mixing on the shelf and in the open sea, the central divergencezones of the Liguro-Provenal and Catalano-Balearic seas, gyres, upwellingsand vertical mixing. Production and stock of phytoplancton occurs in favor-able zones (fronts, whirlwinds, river plumes) inducing production and stockof zooplankton in relation dominated by Calanus helgolandicus, Centropagestypicus, and Pseudocalanus and Paracalanus sp. (Champalbert, 1996). Smallpelagics such as sardine and anchovy, and medium size-pelagics such as mack-erel and bonite are the main contributors to total landings (about 50%). Sar-dine and anchovy, which have a schooling behaviour, are the main pelagicspecies landed in the Western Mediterranean.The GoL is a major anchovyspawning area in the NWM (Garcia et al., 1996; Palomera et al., 2007), owingto its relative high primary production over the year (Diaz et al., 2001).

In this study, we investigate the influence of the hydrodynamic processes onzooplankton transport and distributions in the NWM. We simulate the trajec-tories of passive particles and vertically migrating organisms in the region byusing a lagrangian-particle algorithm driven by velocity fields from a 3D hy-drodynamic model. Our goal is to present a first step to better understand the

3

mechanisms guiding the biological production of the NWM and particularly inthe GoL. An outline of the Lagrangian approach is given in the next section.In section 3 we present the results of numerical simulation of the period fromMarch to August 2001 with taking zooplankton either as passive particles orwith active DVM. In the following text, the term “particles” will be used todescribe these simple zooplankton individuals. Forthcoming simulations arediscussed in section 4.

2 Materials and methods

2.1 Hydrodynamic model

We use the 3D numerical hydrodynamic model Symphonie to determinecurrent velocity fields within the model domain shown in Fig. 1. A detaileddescription of the model is given by Marsaleix et al. (2008) and referencestherein. During last ten years the model has been successfully used for severalstudies in the NWM and the GoL: the NC intrusions on the continental shelf(Auclair et al., 2001; Petrenko et al., 2005; Gatti et al., 2006), wind-forcedcirculation (Auclair et al., 2003; Estournel et al., 2003; Petrenko et al., 2008)and the Rhone river plume dynamics (Estournel et al., 2001).

The model uses the Arakawa-C grid with horizontal mesh of 3 km and ahybrid (z-σ) coordinate system (Estournel et al., 2007) with 41 vertical levels.A minimum depth of 3 m is imposed in nearshore areas. Turbulence closure inthe vertical direction is achieved through the scheme proposed by Gaspar etal. (1990). An upwind type advection scheme is described in Hu et al. (2008)and references therein.

(Fig. 1)

The real fresh-water inputs from the Rhone river are provided by the “Com-pagnie Nationale du Rhone” every day and meteorological forcings are givenfrom the Meteo-France model Aladin at high frequency (3h). The restoringterms of the open boundary scheme allow forcing the model with the featuresof the general circulation given by the regional-scale model MOM.

Modelling results were recently validated against the satellite measurementsby Bouffard et al. (2008).

4

2.2 Particle tracking model

We use a Lagrangian particle-tracking code based on the ROMS OfflineFloats (ROFF, introduced in details by Carr et al., 2007). This code uses anadvanced fourth-order accurate Adams-Bashford-Moulton predictor-correctorscheme to integrate d~x/dt = [~u(~x, t)] over time given the initial condition~x(t0) = ~x0. The right-hand side ~u(~x, t) = ~usym + ~udvm is comprised by a seriesof stored 3D Symphonie velocity fields and zooplankton swimming velocitieswithin their DVM behaviors.

The item ~usym is estimated using linear interpolation in time and space ofthe discrete velocity fields, i.e. the velocity values are linearly interpolatedfrom the velocity values of the eight nearest grid cells. The velocity fields aredaily averaged simulations of the Symphonie model, which are provided every24 hours.

Two types of numerical experiments are performed about ~udvm: (i) zoo-plankton is considered as passive drifters for which the transport processesare determined uniquely by the velocity fields of the model circulation; (ii)DVM behaviors has been considered as follow (Fig. 2): if a particle is in alayer upper than -50m on 06:00, it will swim down with the velocity 50 m/hfrom 06:00 to 08:00, otherwise the zooplankton transport processes are onlydetermined by the velocity fields of the model circulation; from 18:00 to 20:00,all particles will swim up from deeper depth to near-surface with the velocity50 m/h. No limits are fixed for depth reached by zooplankton. The value 50m/h is suggested by field observations (Mauchline, 1998).

For the vertical direction we modified the original ROFF code and we usethe z coordinate system. Moreover, a particle reflection condition is imple-mented at the model rigid boundaries (the coastal boundaries and the bottomboundaries) while the particles leaving the model domain through the openboundaries are assumed to be lost.

To estimate the exchange between the shelf and the open sea, we choose twolocations to release the particles. The first one is located in the Rhone riverplume (position R in Fig. 1) and considered as a shelf area. The second one isaround the oceanographic sampling station DYFAMED (position D in Fig. 1)and considered as an offshore area. The 200 particles are released at positionR in a rectangle area of 60×20 km with the center 4.8◦E, 43.2◦N, and at twodifferent depths, with 100 particles at 5m and 100 particles at 20m. Moreover,the 200 particles are released at position D in a square area of 30×30 km withthe center 7.87◦E, 43.42◦N, and at two different depths, with 100 particles at5m and 100 particles at 100m.

To estimate the influence of the time variability, each particle is started

5

every 3 days from March 1st to August 31, 2001 and tracked for 40 days. Thelife spans of different species vary considerably. We use 40-day trajectoriesbecause they would represent one life duration of many species (Mauchline,1998).

(Fig. 2)

After accurate sensitivity tests and on the basis of computing time con-straints we decide to run the particle-tracking model with a time-step of 300seconds.

2.3 Simulation analysis

The modeled domain, extends between longitude 1.75◦W and 10.90◦E andlatitude 38.28◦N and 45.61◦N (Fig. 1). In order to classify the different zonesof the NWM as aggregative or dispersive, we divide the modeled domain into 9sectors. Sector 1 and 2 correspond to the shelf area delimited by the isobath of200m in the GoL and in the Catalan sea respectively; sector 3 marks shelf areaaround the Balearic islands in the Catalan sea; sector 4, 5 and 6 represent theLigurian sea (Here sector 5 and 6 represent different ecosystems in the Liguriansea); sector 7 bounds the center of the NWM gyre; sector 8 is the area wherethe main branch of the NC passes; sector 9 represents the offshore zone in theCatalan sea.

3 Results

3.1 Passive particles

Particles are released at position D and position R every 3 days from March1st to August 31, 2001. During the simulations, the particles are simulated aspassive zooplankton individuals. The final distributions are shown in Fig. 3and Fig. 5. The particles released in one month are plotted in one figure. HereFig. 3 shows the final locations of the particles released at position D andFig. 5 at position R. The release locations are also included. The percentagesof the particles distributions in different sectors are shown in Table 1 andTable 2. The particles released in one month have been considered as a whole.

6

3.1.1 Particles released at position D

After 40 days being released at position D, the particles could almost spreadanywhere in the NWM (Table 1 and Fig. 3). After being released at position D,the particles are divided into two parts resulting from the currents. One partfollows the anticlockwise circulation, firstly drifts southern and then easternalong the WCC. After reaching the area northeast to the Corsica Island, theparticles flow northern along the continental slopes, then return in the NC andgo back at position D again. At the end of simulations, this part remains in thesector 4 and 5, i.e. the Ligurian sea. Another part of the particles follows theNC westward. From the sector 6, the main particles trap in the NC (the sector8) along the continental slope and then flow into the Catalan sea (the sector 2and 9), a part moves more offshore into the sector 7 and a few particles enterin the sector 1 with the intrusion of the NC at the eastern side of the GoL.In the Catalan sea, the particles reach as far as the Channel of Ibiza whereit splits in two parts. The main part re-turn into the Northwestern Basin andsome particles enter in the sector 3 (Fig. 3).

Strong seasonal patterns are observed in the transport and distributions ofthe particles (Table 1, Fig. 3). Only 17% of the particles, which are releasedin May and June, maintain in the Ligurian Sea (the sectors 4-6) while themaximum, over 60% in March. On the contrary, only 9% enter in the CatalanSea (the sectors 2-3 and 9) in March while 46% in May (Table 1). The distri-butions of the particles in the path areas of the NC are more offshore in Julyand August than in other months (Fig. 3). At the end of the simulations, theparticles scatter in the northeastern part of the Catalan Sea, which released inMarch and April. The main particles released in July and August concentratein the path of the NC. And the particles released in May and June distributemainly similar with the cases in July. However, comparing to only a few par-ticles locate in the northeastward re-turn branch in July, more particles are inthe path of the re-turn branch in May and June, even some flow to the areaseast to the Balearic Island in May.

Moreover, different patterns appear in the final distributions of the particleswith different initial release depths. At the end of the simulations, the mainparts of the particles locate in the path of the NC, which are released at -100m.The particles released at -5m distribute more complex than those released at-100m, which scatter in the whole areas of the NWM, and the patterns aredifferent in different months. Moreover, along the path of the NC, the particlesreleased at -5m locate further than the particles released at -100m, and onlya few of the latter particles reach the open boundary.

Under specific wind and stratification conditions, surface waters of the NCtend to penetrate onto the shelf at the eastern entrance of the gulf (Millotand Wald, 1980; Auclair et al., 2003; Echevin et al., 2003; Petrenko, 2003;

7

Petrenko et al., 2005). Consequently, the penetration will carry the particlesinto the shelf. For an instance, the trajectories of the particles released at May15, 2001 are shown in Fig. 4. In Fig. 4, when the particles are transported atthe east edge of the GoL, a part of them enters in the shelf along the shore.However, during our simulations from March to August, only a few particleslaunch into the GoL and finally staying there(Fig. 3 and Table 1), with themaximum percentage 3%. Most of the time, the NC flow strongly along thecontinental slope of the GoL, and the particles drift within the NC along theslope instead of entering in the GoL. Thus we can deduce that the NC can beconsidered as a barrier for particles entering into the GoL, at least during theperiod of our simulations.

(Fig. 3)

(Table 1)

(Fig. 4)

3.1.2 Particles released at position R

At the end of 40 days simulations, the distributions of the particles releasedat position R can be divided into two parts (Fig. 5). One remains in the GoLand the other one scatters in the path of the NC and in the Catalan sea.About a quarter to one half of the particles remains in the GoL (Table 2) andno one particle enter into the Ligurian sea during the simulations.

After released, one part of particles flows northwestward along the ashorethen moves westward in the north part of the GoL. After launching the westernside of the GoL, some particles move northeastward towards the central partof the shelf, with a speed of order of 10 cm/s, and the others go southward.Another part of the particles, is derived by the shelf currents in the Rhoneriver region, flow southwestward out of the shelf to be trapped in the path ofthe NC directly or first follow the Rhone river plume looping anticyclonically.The particles entering the offshore branch of the NC are not able to reenter theshelf. At the end of the simulations, the particles are found in the southwesternpart of the sector 7,8 and the southeastern part of sector 9 covering the wholespan of the regional cyclonic circulation.

Weak seasonal patterns appear in the final distributions of the particlesoriginated from position R. The final distributions in the GoL are almostthe same in different months, with the particles scattering in the whole areas(Fig. 5). The percentage is a little different, with 24% in March while 47% inApril and 56% in May, and about one third of the particles in other months(Table 2).

8

The final distributions of the particles in the Catalan sea vary larger thanthose in the GoL. The seasonal variation is the same as that of the particlesreleased at position D, which have been described before. More particles locatein the continental shelf in the Catalan sea from position R than position D,with the maximum percentage 7% versus 1% (Table 1 and Table 2).

The different distribution patterns are also weak of the particles with differ-ent released depths. At the end of simulations, the distribution patterns of theparticles released at -5m are similar as those at -20m. Concerning the particlesflow out of the GoL, the trajectories show that they can move out throughalmost anywhere of the south boundary to the offshore sea (data not shown).Therefore, in contrast to the exchange of the particles from the offshore sea tothe shelf area, which is only a few under a certain circumstances and throughthe specific place where the NC penetrate, the exchange from the shelf areato the offshore is large and frequent.

(Fig. 5)

(Table 2)

3.2 Particles with DVM

At the end of the simulations with DVM, the distributions characteristics ofthe particles released at position D are mainly similar with the passive parti-cles. However, several differences exist in the two situations. The percentagesof the particles in different sectors are shown in Table 3 and the final distri-butions are plotted in Fig. 6 (here for instance, only the particles released inMarch are shown). In Table 3, the percentages in the sector 5A, 7A are lowerthan those shown in Table 1, on the contrary, the percentages in the sectors5B, 8A are higher. These can be described as the final distributions of theparticles with DVM concentrate more in the areas of the NC path than thepassive particles, which are also shown in Fig. 6. Still very few particles re-leased at position D can enter in the GoL. Moreover, few or none particles canlaunch the continental shelf in the Catalan sea at the end of the simulations.Strong seasonal patterns also appear in the final distributions of the particles.However, the differences of the particles with different released depths becomeweak or nearly disappear (Fig. 4).

The final distribution patterns of the particles simulated with DVM aredifferent compared to the cases of passive particles, with the percentages indifferent sectors are shown in Table 4, which released at position R. Theparticles, which remain in the GoL at the end of the simulations, are less over10% than the passive particles. Moreover, unlike the passive particles scatterevery areas in the GoL, the particles with DVM locate mainly in the center

9

and western parts of the shelf. More particles go into the NC and enter inthe Catalan sea. From Table 4 one can see that about 40% of the particlesdistribute in the sector 8 except for April, while about one third of the particleslocate in the sector 9 except for May.

(Table 3)

(Fig. 6)

(Table 4)

4 Discussion

4.1 Fate of particles released in the central Ligurian sea

4.1.1 General description

The distributions of the passive particles depend on the direction and thevelocity of the currents which vary seasonally.

After 40 days, in all simulations there is an overall spatial distribution cov-ering the whole system of the NWM, with a clear role of the NC as vector fromthe Ligurian sea to the Catalan sea, highlighting a potential high connectivitybetween the different regions.

Firstly, a cyclonic circulation in the Ligurian sea can trap the particles andmaintain many particles between the north of Corsica and the Gulf of Genoa.In the simulation starting in March 1st, April 1st, July 1st and August 1st

(Fig. 3A, B, E and F), a large part of the particles (≥15%) released at positionD are maintained in this region (see Table 1) and in other simulations, a fewparticles (≥2%) end in this region.

A second feature in the region between the Azure Coast and the North Westof Corsica (sector 6) can move particles from the NC to offshore and trap themsuch as simulations starting in March 1st and July 1st (Fig. 3E). Poulain (2008)presented a recent example of drifter measurements in the Ligurian sea andan update on the MedArgo float program. And he showed these two featuresclearly in the drifter trajectories in the Liguro-Provencal basin.

Most of the particle trapped in the NC can pursuit their way along the shelfbreak of the GoL down to the Catalan sea. In the Catalan Sea the NC reachesas far as the Channel of Ibiza where it splits in two parts and the main part re-circulates into the Northwestern Basin to generate the Balearic Front (Andre

10

et al., 2005). The circulation in the Catalan sea can aggregate many particlesreleased in position D, such as the simulations starting in April 1st, May 1st,July 1st. As written before, in all simulations, a part of the particles arrivein the North of the Balears where a branch of the North Current produce acyclonic gyre and an offshore dome (Salat, 1996). This cyclonic circulation isknown for higher concentration of zooplankton (Alcaraz et al., 2007).

During our simulations, an offshore-shelf transport is few and the particlesending in the GoL is in low concentrations. One reason is that the NC cir-culates southwestwards along the shelf break delimiting the GoL, it can actas a barrier and separates the shelf circulation from the regional circulation.Moreover, the southwestern branch of the current at the Rhone river plumetake the particles reenter in the NC flow, which occasionally move into thesoutheastern part of the GoL advected by the northern branch of the NC.Therefore, the exchange from the offshore areas to the shelf is few.

4.1.2 Initial position

In our simulations, the initial release positions (5m or 100m depths) stronglyinfluence the final distributions of the passive particles. The main reason isthat the currents in the surface layer are more complex than those in thedeeper layer due to the wind forcing. Moreover, the velocities of the NC arelarger in the upper layer than in the deeper layer. However, the final distribu-tions discrepancy with the different initial positions has been reduced whenDVM considered. This is a consequence of the particles, which are released atdifferent initial depths, have nearly the same vertical migrations after beingsimulated a few steps with DVM scheme. Compared to passive transport atthe surface, the particles with DVM will reside longer at deeper depths. On thecontrary, the particles will maintain longer in surface or near-surface watersrelative to passive transport at deep depths.

4.1.3 Seasonal variability

The different distribution patterns of particles released for 40 days fromposition D highlight the influence of the seasonal patterns of the current fieldsin the NWM. According to the modeled current flows by the Symphonie model(data not shown), the southeastern part of the currents at position D turnsweak or even disappears in May and June compared with other months. Thiscan explain why more particles were advected to the Catalan sea and theGoL. Furthermore, according to several campaigns performed in the NWM(Alberola et al., 1995; Petrenko, 2003), the NC flux varies throughout theyear, with a maximum flux (1.5-2 Sv down to 700 dbar) during winter andspring seasons (roughly from December to May). The velocities of the NC are

11

larger in May and June than those in July and August, thus the particles willflow faster accordingly.

4.2 Fate of particles released in the Rhone river plume

4.2.1 General description

As the passive particles released at position R, the distribution is a conse-quence of the shelf circulation and the open-coastal water exchanges. The GoLcan be properly considered as a retention area for the particles transport anddistributions. One reason is the shelf circulations are weak in the main areasof the GoL, thus the speed of the particles transport is small. Another reasonis probably due to the gyres and eddies on the shelf. When particles enterin, they are confined in so that the particles are prevented from or delayedescaping the GoL.

During the simulations, no one passive particles released at position R enterinto the Ligurian sea. These could be deduced that at least under the currentconditions as we simulate, the zooplankton in the GoL will not take an effecton the ecosystems in the Ligurian sea.

4.2.2 Effect of DVM

In a upwelling region, there are often offshore currents at the surface andonshore currents at deeper depths. Consequently DVM reduces the transportof the particles away from the region, compared to the particles transportpassively at the surface (Botsford et al., 1994; Batchelder et al., 2002). Ourresults provide the evidence. The trajectories of the particles released at po-sition R on March 1st are shown in Fig. 7 as passive tracking and with DVMscheme. Some passive particles flow out of the GoL through the south side,and then move southwestward to the sector 7 and 8. However, under the sameconditions of the currents, the particles with DVM remain in the GoL after40 days simulations.

However, affected by the complex circulations, the particles with DVM donot transport in a simply way as Fig. 7 show. And the results show thatour simulations with DVM do not increase the retention on the shelf. Thetransport and distributions of the particles are determined by both the shelfcirculations and the DVM behaviors.

The shelf circulations of the GoL are so complex, due to various forc-ings, which include strong northwestern (tramoutand) and northern (mistral)winds, the western Mediterranean mesoscale circulations (the NC), and the

12

fresh water input (the Rhone river plume). The hydrodynamic characteris-tics of the waters in the GoL are much variable under different environmentalconditions, even during the same wind event. The NC shows a clear seasonalvariability, and the direction and extent of the Rhone river plume are greatlyaffected by the winds (Marsaleix et al., 1998; Estournel et al., 2001; Naudin etal., 2001). Furthermore, inertial oscillations can be generated under a certainconditions (Petrenko et al., 2005).

In view of our scheme for DVM is so primary, the present analysis representsthe first attempt to characterise particles transport with DVM in the GoL. Sothe results should be considered as suggestive rather than conclusive. ActuallyDVM is a very complex behavior of zooplankton, which is concerned withtemperature, nutrition, prey-predator relationships and life habits and so on.At the next step, we will advance an complicated scheme to analyze the DVMbehaviors by considering more factors.

(Fig. 7)

5 Conclusions

Up to now, we have developed a Lagrangian tool to simulate the transportand distributions of particles coupling with the 3D circulation model Sym-phonie. The particles could be zooplankton, sediments or other passive sus-pended matters. And a primary DVM scheme has been successfully consideredfor zooplankton. The Lagrangian tool has been used to estimate the influenceof hydrodynamic processes on zooplankton transport and distributions in theNWM, with the particles released around the DYFAMED sampling stationand the Rhone river plume every 3 days from March to August 2001 for 40days.

Our results suggest the particles transport and distributions are stronglyrelated to the hydrodynamic structures on shelf and the offshore circulationsassociated with the NC. In the regional scale, the particles could launch al-most anywhere in the NWM after being transported for 40 days, which re-leased around the DYFAMED station. Strong seasonal patterns appear in thedistributions of the particles in the offshore while weak patterns in the GoL.Large differences also exist in the distributions of the passively drifted parti-cles with different released depths, although which become weak or disappearwith DVM. Our results also suggest that at least in spring and summer condi-tions, the current fields in the NWM favor a shelf-offshore particle exchange,whilst an offshore-shelf transport is nearly inhibited. The NC can be properlyconsidered as a barrier for particles entering into the GoL from the offshoresea. The GoL can be defined as a retention area, because about a quarter

13

to one half particles remain in the shelf after 40 days, which released at theRhone river plume. Another part of the particles scatter in the path of theNC and in the Catalan sea.

DVM behaviors also have an impact on the zooplankton transport anddistributions combined with the hydrodynamic processes. The discrepancy re-duces in the final distributions of the particles released at different depths,compared to those passive particles. The simple DVM scheme does not in-crease the retention on the shelf. More complicated scheme will be advancedfor DVM in the future to better understand the particles transport and distri-butions, and the mechanisms guiding the biological production of the NWM,particularly in the GoL.

Acknowledgements

We thank Dr. Capet for kindly providing the ROFF codes. This researchwas supported by CNRS.

References

Alberola, C., Millot, C., Font, J., 1995. On the seasonal and mesoscale vari-abilities of the Northern Current during the PRIMO-0 experiment in thewestern Mediterranean Sea. Ocean. Acta. 18 (2), 163-192.

Alcaraz, M., Calbet, A., Estrada, M., Marrase, C., Saiz, E., Trepat, I., 2007.Physical control of zooplankton communities in the Catalan Sea. Progr.Oceanogr. 74, 294-312.

Alexei, S., Konstantin, K., 2007. Modelling distribution of flounder larvae inthe eastern English Channel: sensitivity to physical forcing and biologicalbehaviour. Mar. Ecol., Prog. Ser. 347, 233-245.

Andre, G., Garreau, P., Garnier, V., Fraunie, P., 2005. Modelled variability ofthe sea surface circulation in the North-western Mediterranean Sea and inthe Gulf of Lions. Ocean. Dyn. 55, 294-308.

Auclair, F., Marsaleix, P., Estournel, C., 2001. The penetration of the northerncurrent over the Gulf of Lion (western Mediterranean Sea) as a downscalingproblem. Ocean. Acta. 24, 529-544.

Auclair, F., Marsaleix, P., De Mey P., 2003. Space-time structure and dynam-ics of the forecast error in a coastal circulation model of the Gulf of Lions.Dyn. Atmos. Oceans. 36, 309-346.

Batchelder, H., Edwards, C., Powell, T., 2002. Individual-based models ofcopepod population in coastal upwelling regions: implications of physi-

14

ologically influenced diel vertical migration on demographic success andnearshore retention. Progr. Oceanogr. 53, 307-333.

Blanke, B., Arhan, M., Madec, G., Roche, S., 1999. Warm water paths in theequatorial Atlantic as diagnosed with a general circulation model. J.Phys.Oceanogr. 29, 2753-2768.

Botsford, L.W., Moloney, C.L., Hastings, A., et al., 1994. The inuence of spa-tially and temporally varying oceanographic conditions on meroplanktonicmetapopulations. Deep-Sea. Res. II. 41,107145.

Bouffard, J., Vignudelli, S., Herrmann, M., Lyard, F., Marsaleix, P., Menard,Y., Cipollini, P., 2008. Comparison of ocean dynamics with a regional circu-lation model and improved altimetry in the North-western Mediterranean.Terr. Atmos. Ocean. Sci. 19, 117-133.

Carr, S.D., Capet, X.J., Mcwilliams, J.C., Pennington, J.T., Chavez, F.P.,2007. The influence of diel vertical migration on zooplankton transport andrecruitment in an upwelling region: estimates from a coupled behavioral-physical model. Fish. Oceanogr.

Champalbert, G., 1996. Characteristics of zooplankton standing stock andcommunities in the western Mediterranean Sea: relations to hydrology. Sci.Mar. 60 (Suppl. 2), 97-113.

Cianelli, D., Diaz, F., Leredde, Y., Marsaleix, P., Carlotti, F., 2007. Particleexchange and residence times in the North Western Mediterranean. Nuovo.Cimento. C. 30, 138-149.

Diaz, F., Raimbault, P., Boudjellal, B., Garcia, N., Moutin, T., 2001. Earlyphosphorus limitation during spring in the Gulf of Lions. Mar. Ecol., Prog.Ser. 211, 51-62.

Echevin, V., Crepon, M., Mortier, L., 2003. Interaction of a coastal currentwith a gulf:application to the shelf circulation of the Gulf of Lions in theMediterranean Sea. J. Phys. Oceanogr. 33, 188-206.

Estournel, C., Broche, P., Marsaleix, P., Devenon, J.L., Auclair, F., Vehil,R., 2001. The Rhone river plume in unsteady conditions: numerical andexperimental results. Est. Coast. and Shelf Sci. 53, 25-38.

Estournel, C., Durrieu, X., Marsaleix, P., Auclair, F., Julliand, C., Vehil, R.,2003. Observation and modelisation of the winter coastal oceanic circulationin the Gulf of Lions under wind conditions influenced by the continentalorography (FETCH experiment). J. Geophys. Res. 108 (C3), pages 7-1 to7-18.

Estournel, C., Auclair, F., Lux, M., Nguyen, C., Marsaleix, P., 2007. “Scaleoriented” embedded modeling of the North-Western Mediterranean in theframe of MFSTEP. Ocean Sci. Discuss. 4, 145-187.

Falco, P., Griffa, A., Poulain, P.M., Zambianchi, E., 2000. Transport propertiesin the Adriatic Sea as deduced from drifter data. J. Phys. Oceanogr. 30,2055-2071.

Franks, P.J.S., 1992. Sink or swim: accumulation of biomass at fronts. Mar.Ecol., Prog. Ser. 82, 1-12.

Garcia, L.E., Castellon, A., Font, J., Tintore, J., 1996. The Balearic current

15

and volume transports in the Balearic basin. Ocean. Acta. 19, 489-497.Gaspar, P., Gregoris, Y., Lefevre, J.M., 1990. A simple eddy kinetic energy

model for simulations of the oceanic vertical mixing:Testsat station Papaand long-term upper ocean study site. J. Geophys. Res.95,16179-16193.

Gatti, J., Petrenko, A., Devenon, J.L., Leredde, Y., Ulses, C., 2006. The Rhoneriver dilution zone present in the northeastern shelf of the Gulf of Lion inDecember 2003. Continent. Shelf. Res. 26, 1794-1805.

Guizien, K., Brochier, T., Duchne, J.C., Koh, B.S., Marsaleix, P., 2006. Dis-persal of owenia fusiformis larvae by wind-driven currents: turbulence, swim-ming behaviour and mortality in a three-dimensional stochastic model. Mar.Ecol., Prog. Ser. 311, 47-66.

Hu, Z.Y., Doglioli, A.M., Petrenko, A.A., Marsaleix, P., Dekeyser, I., 2008.Numerical simulations of eddies in the Gulf of Lion. (Submit to Geophys.Res. Lett.)

James, I., 1996. Advection schemes for shelf sea models. J. Mar. Syst.8,237-254.

Jenkins, G.P., Black, K.P., Keough, M.J., 1999. The role of passive transportand the influence of vertical migration on the pre-settlement distributionof a temperate, demersal fish: numerical model predictions compared withfield sampling. Mar. Ecol., Prog. Ser. 184, 259-271.

Lett, C., Veitch, J., van der Lingen, C.D., Hutchings, L., 2007. Assessment ofan environmental barrier to transport of ichthyoplankton from the southernto the northern Benguela ecosystems. Mar. Ecol., Prog. Ser. 347, 247-259

Levy, M.,Memery, L., Madec, G., 2000. Combined effects of mesoscale pro-cesses and atmospheric high-frequency variability on the spring bloom inthe MEDOC area - a mathematical model. Deep-Sea Res. I. 47, 27-53.

Marsaleix, P., Estournel, C., Kondrachoff, V., Vehil, R., 1998. A numericalstudy of the formation of the Rhone River plume. J. Mar. Syst. 14 (1-2),99-115.

Marsaleix, P., Auclair, F., Floor, J.W., Herrmann, M.J., Estournel, C.,Pairaud, I., Ulses, C., 2008. Energy conservation issues in sigma-coordinatefree-surface ocean models. Ocean. Mod. 20 (1),61-89.

Mauchline, J., 1998. The biology of calanoid copepods. Adv. Mar. Biol 33, 710pp.

Miller, C.B., Lynch, D.R., Carlotti, F., Gentleman, W., Lewis, C.V.W., 1998.Coupling of an Individual-Based Population Dynamic Model of Calanusfinmarchicus to a Circulation Model for the Georges Bank Region. Fish.Oceanogr. 7, 219-234.

Millot, C., Wald, L., 1980. The effect of Mistral wind on the Ligurian currentnear Provence. Ocean. Acta. 3, 399-402.

Millot, C., 1999. Circulation in the Western Mediterranean Sea. J. Mar. Syst.20, 423-442.

Mullon, C., Freon, P., Parada, C., van der lingen, C., Huggett, J., 2003. Fromparticles to individuals: modelling the early stages of anchovy (Engrauliscapensislencrasicolus) in the southern Benguela. Fish. Oceanogr. 12(4-5),

16

396-406.Naudin, J., Cauwet, G., Fajon, C., Oriol, L., Terzic, S., Devenon, J.L., Broche,

P., 2001. Effect of mixing on microbial communities in the Rhone RiverPlume. J. Mar. Syst. 28, 203-227.

Oschlies, A., 2002. Nutrient supply to the surface waters of the North Atlantic:A model study. J. Geophys. Res. 107 (C5), 3046.

Palomera, I., Olivar, M.P., Salat, J., Sabates, A., Coll, M., Garca A., Morales-Nin, B., 2007. Small pelagic fish in the NW Mediterranean Sea: An ecologicalreview. Progr. Oceanogr. 74 (2-3), 377-396.

Petrenko, A., 2003. Variability of the circulation features in the Gulf of LionNW Mediterranean Sea. Importance of inertial currents. Ocean. Acta. 26,323-338.

Petrenko, A., Leredde, Y., Marsaleix, P., 2005. Circulation in a stratified andwind-forced Gulf of Lions, NW Mediterranean Sea : in-situ and modelingdata. Continent. Shelf. Res. 25, 7-27.

Petrenko, A., Dufau, C., Estournel, C., 2008. Barotropic eastward currents inthe western Gulf of Lion, north-western Mediterranean Sea, during stratifiedconditions. J. Mar. Syst. doi:10.1016/j.jmarsys.2008.03.004.

Poulain, P.M., 2008. Integration of Lagrangian observations into a Mediter-ranean Marine Observatory, Towards an integrated system of Mediterraneanmarine observatories. No 34 in CIESM Workshop Monographs [F. Briand,Ed.], 144 pages, Monaco.

Robinson, A.R., Brink, K.H., 2005. The Global Coastal Ocean: Interdis-ciplinary Multiscale Processes, Regional Dynamics and Syntheses. TheOceanography Society Meeting, Paris, France, 9 June 2005.

Salat, J., 1996. Review of hydrographic environmental factors that may influ-ence anchovy habitats in northwestern Mediterranean. Sci. Mar., 60(Supl.2),21-32.

Speirs, D.C., Gurney, W.S.C., Heath, M.R., Horbelt, W., Wood S.N., Cuevas,B.A., 2006. Ocean-scale modelling of the distribution, abundance, and sea-sonal dynamics of the copepod Calanus finmarchicus. Mar. Ecol., Prog. Ser.313, 173-192.

17

Fig. 1. Model domain (dashed rectangle) and the corresponding bathymetry (thincontours). Black filled rectangles represent the particles release locations (see text).The model domain is divided into 9 sectors (thick lines) for the analysis of particledistribution. Sectors indicated with A/B considered the layer upper 200m depth(A) and the layer below 200m depth (B).

Fig. 2. Schematic representation of the modeled zooplankton swimming behaviorswith diel vertical migration (DVM).

18

Fig. 3. The final locations after 40 days of passive particles released at position D(blue square) for different months: (A) March, (B) April, (C) May, (D) June, (E)July and (F) August. Empty red circles: final positions of particles released at -5m;Full green lozenges: final positions of particles released at -100m.

19

Fig. 4. Trajectories of 32 passive particles released on May 15, 2001 at -5m atposition D (green circles): Red circles: final positions after 40 days transport. Forgraphical purposes only a subset of particles is shown.

20

Fig. 5. The final locations after 40 days of passive particles released at position R(blue square) for different months: (A) March, (B) April, (C) May, (D) June, (E)July and (F) August. Empty red circles: final positions of particles released at -5m;Full green lozenges: final positions of particles released at -100m.

21

Fig. 6. The final locations after 40 days of particles released with DVM at position D(blue square) for March 2001. Empty red circles: final positions of particles releasedat -5m; Full green lozenges: final positions of particles released at -100m.

Fig. 7. Trajectories of 10 particles released at position R on March 1st as passive par-ticles (A) and with DVM (B). Red circles: release positions. For graphical purposesonly a subset of particles is shown.

22

Table 1Distributions (%) of passive particles after 40 days released at position D in every3 days from March 1st to August 31, 2001

1 2 3 4 5A 5B 6A 6B 7A 7B 8A 8B 9A 9B

March ≤1 ≤1 1 10 41 1 12 ≤1 4 0 21 2 6 2

April 2 ≤1 ≤1 4 15 2 5 1 11 0 25 14 19 2

May 2 1 2 3 7 1 5 1 6 0 20 8 32 11

June 3 1 3 ≤1 3 ≤1 9 3 13 0 24 7 32 3

July ≤1 0 ≤1 2 27 ≤1 17 0 2 0 28 5 20 ≤1

August ≤1 ≤1 ≤1 3 18 0 14 0 8 0 16 1 30 9

Table 2Distributions (%) of passive particles after 40 days released at position R in every3 days from March 1st to August 31, 2001

1 2 3 4 5A 5B 6A 6B 7A 7B 8A 8B 9A 9B

March 24 3 5 0 0 0 0 0 16 0 38 ≤1 14 ≤1

April 47 1 1 0 0 0 0 0 12 0 26 0 12 ≤1

May 56 4 ≤1 0 0 0 0 0 2 0 29 0 10 0

June 35 2 2 0 0 0 0 0 7 0 27 0 28 0

July 34 1 3 0 0 0 0 0 5 0 25 0 32 0

August 36 7 3 0 0 0 0 0 ≤1 0 23 0 30 0

Table 3Distributions (%) of particles after 40 days released with DVM at position D inevery 3 days from March 1st to August 31, 2001

1 2 3 4 5A 5B 6A 6B 7A 7B 8A 8B 9A 9B

March ≤1 0 ≤1 2 38 4 13 2 3 0 28 5 4 ≤1

April 2 0 0 1 10 5 5 2 2 0 39 13 19 2

May ≤1 ≤1 2 3 2 4 3 3 2 0 26 14 30 10

June 2 0 3 0 ≤1 1 4 6 1 0 35 6 39 1

July ≤1 0 0 0 26 ≤1 15 ≤1 2 0 40 7 10 ≤1

August ≤1 0 0 ≤1 9 ≤1 7 ≤1 7 0 37 11 27 ≤1

23

Table 4Distributions (%) of particles after 40 days released with DVM at position R inevery 3 days from March 1st to August 31, 2001

1 2 3 4 5A 5B 6A 6B 7A 7B 8A 8B 9A 9B

March 10 3 7 0 0 0 0 0 ≤1 0 46 ≤1 32 ≤1

April 32 2 3 0 0 0 0 0 6 0 22 ≤1 32 2

May 42 2 ≤1 0 0 0 0 0 ≤1 0 51 ≤1 4 0

June 16 ≤1 ≤1 0 0 0 0 0 0 0 47 ≤1 36 ≤1

July 23 ≤1 ≤1 0 0 0 0 0 ≤1 0 49 ≤1 28 0

August 20 ≤1 ≤1 0 0 0 0 0 0 0 39 ≤1 39 ≤1

24