0.

THIS PAPEn NOT TO BE CITED WITIlOUTPRIOR ltEFERENCE '1'0 TIIE AU'l'llOR

ICES C.1\1. 19% PAPERS:7 - Thcme Session S

~..,>~.......-----....~ .• , ...

I~>" .,!\ : ; c. ! ;..., _. .

" . _. "". I·~

"'<-:..~"/

A NUl\IERICAL l\!I0DEL SYSTE1\I OF THE REGION AHOUND TIIEIUERIAN PENINSULA: 1\IODEL VALIDATION AND APPLlCATION

'1'0 HAKE LAHVAE DRIFT IN TIIE BAY OF 13ISCAY

Joachim 13artsch, Alicia LavEn and Lorenzo l\Iotos

AßSTHACT

\\'ithin the EU fuudcd SEFOS project (Shclf Edge Fisheries and Oceauography Studies) anumerical model systcm cOllsisting of CI circulatiou <lud t.r<lnsporl 1l1Odel !JilS \)('('11 IIsed 10

model the dispersiou of hake eggs and lan'c\(' (Jltrll/ccil/.~ /l/crIILCcill.~). Tlw iln'" of 111<'

model systcm rangcs fwm Nortll\\'cst Africa. to t he 1I0rthern Ba)" of Biscay aud co\'crs 1h('shclf edge and adjacent waters with Cl horizontal rcsolution of 10' x 10'. The circulationmodel was run using tidal forcing, c1imatological density fieIds a., H'cll as realistic six-hourlywind stress fields 1'01' 1994 amI 19%. Current mcter data collected in the Cantabrian Seain the mixed layer and the upper part of the North Atlantic ('entral Water is used formodel validation. Simulated currents from the circulation model H'cre found to be in goodagreemcnt with observed current data on the shdf amI in thc vicinity 01' the shc1f cdge.Simulated currcnts are used as input to the transport model. Egg and larval data used inthc dispersion simulation was collected on two cruiscs carried out in February (lnd ;\Iarch1995. Horizontal distributions of stage I eggs were uscd as lohe initi<ll distribution for thedispersion scenado. Tracers were disperscd in the depth range 50 - 150m for 39 day:;. Basedon estimated egg development times and larval growth rates, stage I eggs will have growllinto 6mm larvae in approximately 39 days. The tracer distribution after dispersion from 17February to 27 March is compared to the distribution of Gmm lan'ae observcd betH'ccn 22 ­27 March 1995. Agreements amI discrepancies bet,,"cen observations and model results arediscussed.

Kcywords: NumcricaI model system, model validation, larvae dispersion

Joachim Barlsch: Biologische Anstalt He/goland, Notkestl'. ;]1, J:J6'07 Ilambll/'!J, (,'u'/wwy

(lel: .{9 40 89U'9.1177, leu: 494089693 115. e-mail: bartsch@:ifm.llni-Iwmbu/'!J.dt).'Alicia LavEn: Ccntl'O Oceanografico de Santal/der, Apdo. ~240, 39080 Sal/lander, Spain (tel:34 42275033, fa:t: 34 42 275072, e-mail: lavilla@ccai.r3.11llican.es).Lorenzo 1\lolos: AZTI, Food and Fish Technological Inslitute, A vda. Satrllsteglli 8, 20008San Sebastian, Basque Counll'y, Spain '(tel: :J.f 4:3 214124, fax: 3443212162. e-mail:[ol'cnzo@I'p.azli.es).

1

1. INTRODUCTION

One of the specific objeet.ives of the SEFOS (Shelf Edge Fisheries ami OccanographyStudies) project is to gaiu information on the transport. patterns or the eggs a1ld lan'ae01' thc target species in thc SEFOS area. In this invcstigation J/t'/'lllccilt8 111( rlllccilt ...is considered. '1'0 stu<.ly thc drift 01' eggs and larvae o\'er wecks (lud months lIumcricalmodel systems are nce<.led. Only thcy can providc the \'ariable circ1\lation fleld O\'crthese time spaus with the required ::;patial and temporal resolution. CurrclIt clatafrom field sun'eys and moorings are too sparsc to accomplish this. Examplt's of suchmodclling drift studies are those carried out in the Gulf 01' Alaslm Shelf arca on \ValleycPollock (Stabcuo et al., 1995), in thc Georges Bank area on eod and Iladdock (Wernerct. al., 1!.JD:3) and in the German I3ight on Sprat (I3artsch, 199·la and b).

The occanographic and biological data collected during SEFOS in tlw Bay 01' ßiscayi::; us~d to \'alidat.e t.hc circlllat.iou model and to test tlH.' prl'dicl in)lilldcilSl ability ur •tlll" model system.

2. THE CIHCULATION AND TIL-\NSPORT MODEL SYSTEM

A thrce <.limensionalnonlinear baroclinic lllllilericalmodel is u:,;e<.l to :,;inlldatc thc cir­culation in t1le region around the Ibcrian Peninsula, the shelf edgp area and adjacenloceanic l'egions.TI}(' cil'culationmodel is hased on 1l:\~lSO;\I (llAl\lhurg Slwlf Oceanl\lodel) which was deve\oped at thc Institut für !\IccrcskulHk. lIamburg and W<I:'; trilll:>­fel'red to thc sout1lern SEFOS area. No details of thc model will Iw givcn hen~. <I:'; tlwsehase been publishcd clsewhere (13ackha us, 1985).

The model area reaches from 33°N to .!i°N aud frotll 15°\\' to lOW (fig. 1). Thehorizontal grid resolution is 10'x 10', i.e. 1:3 km (east-west) by 18..) km (norl h-south)in thc I3ay of Uiscay. The vertical resolution consists or IG layers, with layer <.lepthsranging from 10 m 1.0 ·10 m in the upper 100 m aud increasing steadily in the lo\\'er •layers. The model is forced by the l\h-tidc, six hOlldy wind stress fields from ECl\lWF(European Centre 1'01' l\Iedium-Range Weather Forecasting) as weil as monthly clima­tological density fields which are trcilted prognostically hut. ",hieh are relaxed 10\\'ardsthe climatological Illonthly mean.

The time step 01' the circulation model is ·15 min. Duc to the high data storagerequirements, the current data, as weil as the variance of the current was only outputon a daily basis, i.e. the currents we1'e integrated ove1' two l\12 - tidal cycles. The <.lailythree dimensional cur1'ent anel variance fielcls serve as input to t\ie transport model.To eluciclate the configuration of the model system, a schematic diagram i::; given infig. 2.

The simulations 01' the drift routes are perfol'med by mcans 01' tracers, in esscnce"lllarkeel" water pa1'ticles which are introduced int.o the model area and are pllrsued inthe space and time dOlllain. A simple lllethod with \vhich the transport of substances

2

••

•

(particIes) can be simulated is the r..lontc - Carlo method (Bork anel I\laier - Heimer,1978). It is assumed that the current fieIel can be split into a large scale mean current(rr) anel a small scale random fluctuation (-n'), i.e. u = U + u/. The small scalefluctuations are parameterised using the current variance (Backhaus, 1~8~; Bartsch,19!Ha). Thus, at cach timestep, the particle experiences 0. <.lirectional transport by tlleme~n current (advection) as weIl as 0. ran<.lom transport, in magnitude anel direction,by the small scale fluctuations (diffusion).

The boundary conditions in the transport model prescribe a. 110 - flux cOlldition atclosed boundaries, whereas the tracers may leave the model area at open boundaries.The model area; grid size and the vertical resolution are in accordance with that ofthe circulation model, while the time step of the transport model is tllrec 1I0urs.

3. I\IATERIALS AND I\lETIIODS

(a) CURHENT ~IETER DATA

Within tlle SEFOS projcct current data from three moorings were collected for theduration 01' one year. Data from the upper layers and from the first deploYlllclltperiod from 19 February to 15 r..lay 1995 \\"crc uscd 1'01' nlOdel validation. DuC'to thc fishing effort in the area and based on previous experience, moorings \\"creplaced in the locations shown in fig. 3. The current meter on the shclf wasdeployed at 75 m depth, which is below the depth of nets used to catch peIagiefish. At the shelf break (mooring 2) the shallow eurrent meter was eleployeel tostudy the mixed layer and the second at 180 m to study the upper part 01' theNorth Atlantie Central \Vater.

Moorings were deployed from the BIO Francisco Na,·arro. eTD profiles wereeollected during tllc deploymcnt to observe thc hydrographie cOllditions. 01' thearea with thc aim of studying thc influcncc 01' thc wind strc.'ss on tht' stratificationof the upper layer. In March-April al\(ll\lay-Junc Hl95 two oceanographic SEFOScruises were earried out in the Cantabrian Sea eovcring thc area in thc vicinityof the eun'ent meters (Porteiro et al, 1996).

The original eurrent meter time serics were obtaincd with a sampliilg interval of30 min. 1'0 obtain a preliminary smoothing 0. Godin A2 A2 A3 filter was used,whieh effeetively eliminated the high frequeney noise. The data was filtered againto eliminate the effeet 01' the l\12-tide. This filtering was carried out. with a GodillA24 A2·1 A25 filter with a. cut off period 01' 25 hours (Godin, 1991; Alonso anelDiaz dei Rio, 1996). After filtering the data. was ayailahle onee an hOIlr.

3

(b) HAKE EGG AND LARVAL SURVEYS

Data on distribution and abundance of hake eggs anti larvae \\ere collected duringtwo consecutive plankton cruises carried out with RV Investigador in tlte timepcriods 15-2-1 February and 22-30 1'1arch 1!J95, respectiv<,ly. Tlte cruises werespecifically clesigned to establish the horizontal and "ertical distributions of ltakeeggs and larval' togetlter with complementary hydrographical observations at twoconsecutive times Tl and 1'2.

Bot.h cruises covered similar areas of the 8ay 01' Biscay (fig. ·1), i.t'. frolll t Ill'Spanish coast in the south up to -170 30' N and from the 100 III isobatll in theeast to 20-30 nauticalmiles beyond the shelf edge in the \'lest. 80th areal coverageami timing of the cruises were dctennined according to previous work in the areaami a lit.erature review (Casarino and l'[otos, 1!J!).l) with thc aim of cO\'ering amain area of adult hake distribution at thc peak of thc spawning period. •

A detailcd report on the metltods used during the cruises is gin'n in l'[otos et a!.(l!J%). The horizontal distribut.ion of hake eggs and lan·at· was dl'tel'lllillt'd I'I'OIlI

sampies colleckd hy oblique BONGO (GO CIlI diatllt'l<-r) Iwl hauls (Stlli,,11 <lIldnicllilrdson, J!)T7), These sampies n-ere colleckd het.n-ceu l(; - ~~ February and22 - 27 l\Iarch respecti\'cly. Thc vertical distribution was derived frolll sömplescollcctcd at sclccted stations hy 1 m'l mouth aperture ring net hauls. This uel wasfumished with a double aperturcjclosurc mcchanism allO\ving for thc sampling ofdiscrcte layers.

Plankton sampies were fixcd inmcdiately after the complction of thc haul in abuffered cOl11merci<d formaldehyde solution on tap water (10%). All fish eggs an<!lan'<w were sorted frol11 thc plankton sal1lples in the laborator)'. Hake eggs amilan'~e wcre idcntificd (Porehsky, U)75; Coomhs, EH).!; Marrale ct al.. U)!)G; Hus­seil, lU(8), rounted and dassified by de\'(~loplll('l\tal stagt' (('()olllhs <llld \litcllt'11.1082). Lan'ae were measureu (SL) to the lower 0.1 1Il111. No correctioll was madeto allow for shrinkage of larvae due to fixation procedures. 8ailey (1981) gave ashrinkage figure of 9-22% for small Pacific hake larvae « <1 mm) fixed 5-10 mi- •nutes after collectioll, and he assulIled smaller shrinkage rates for bigger .larvae.Egg andlarval abundances per haul were transformed to densities (numbers per10 1112 ) using standard procedures (Smith amI nichardson, UH7).

Temperature, time ami depth profiles were recorded by a. l\IINILOG VEl\ICOdata logger in all planktoll 10ws. CTD SEABInn ~!) profiles 01' tClll(lcrature,condllc\.ivity. tlll'bidity ami clowphyll a \\'ere ("ollt,(:\('<1 CI! se!e('(ed st.ill iOllS (fig.'1). Unfort ullatcly, thc CTD was not deployed d uri IIg thc FebruCl ry cru i:->e d ue torough sea cOllditions.

•

•

(c) EGG DEVELOPl\lENT AND GROWT1I RATES

Data on temperature relateu developmental rates of hake eggs amI yolk sae larvaeare a\'ailable in the literature (Coombs and ~litcheJl, 1982; ~Iarrale et al., 1906).1Iowe\'er, no data is available on growth rates of post-larvae 01' tbe Europeall hake,ncither frolll our cruise Bor frolll thc literature. Alteruati\-el)', we used tlte onl)'availahle larval growth data for t he gelllls ~ Icrlucci us, i.e. da ta frolll ;\ lerlucci usproductus, a hake species inhabiting waters in thc Northcastern Pacific Occan(Bailey, 1981; Sailey ct al.,UJ82).

Aecording to Coombs and !vIitchell (1982) the time necessary 1'01' eggs to reachthc end of stage IV is identical to the incubation time (duration from spawningto hatching). This ranges from 6.82 to 5.27 days at 10 °C and 12°C, respectivcly.

Data on growth rates of yolk-sac larval' is also available. Fig. 2 of Coombsand l\litchell (1982) showed pictures 01' different stages 01' yolk-sac lalTa rearedat different temperatures. An early stage at 8.1°e, C\. lllid stage at 10.0oe, a.

late stage at 15.6°C, ami a very la te one at 15.GoC. They assigll to thesl' stages.agl's from hatching to ferti1ization 01' 200, 191, 1Ti amI 1!)1 hours, respectively.1\larra1e et al., (1996) ineubated hake eggs and yo1k-sae larval' from an artificalfertilization experiment. They observed that the dl'velopml'nt of yolk-sac larval'from hatching to the cxhautioll of thc yolk lasted around 84 hours at 17°e.

In order to have some idea of the actual growth rates 01' post-larval' of 1\1. merlue­eius, the a\'ailable information regarding other 1\lerlueeius speeies was reviewed.Bai1ey (1981) presented developlllellt and growth rates of eggs, yolk-sac larvaeami post-larval' of the Paeific hake, 1\lerluccius productus. The results are SUlll­marized in tab. 1. Egg development rates are slower in tlle European hake thanin the Pacific hake. These slower uevcloplllent rates lllay a.lso hold during the lar­val stages. Assuming similar growth patterns 1'01' yolk-sac larval' 01' both species,.the end of this stage at 12°C would be reached one da)' later in M. lllerluccius,when they are'" 12 days old.

Using the (post) larval growth rates given in tab. 1 1'01' M. productus and thelengths at the beginning of the postlarval stage (1.72 mm for the Gompertz fitand 2.75 mm for the lineal fit), it would take about 31 days (G0111pertz fit) 01' 32days (Linear fit) to reach a length of 6 111111. Similarly, it \\'ould take between 34(Gompertz fit) to 38 days (Linear fit) to reach 7 nll11. Assullling further, that \1.merluccius larvae have slower growth rates than their paeifie counterparts, it isprobable that the stage I eggs on cruise 1 (around 17 February) dc\'clopcd into6 mm lar\'ae by cruise 2 (around 27 1\larcl1) at a tcmpera.ture of '" 12°C, i.e.uuring 39 days.

.5

.1. RESULTS

(a) cunnENT OBSERVATIONS

The metcorological COlHlitions over the Ba,)' of Bisca,)' during the sampled periodare eharacterised oy a. eontinuous track of high and 100v pressure eells (InstitutoNacional de l\leteorologia., 1995). These conditions praduee NW, SW and NEwinds as weil as relaxed periods. There were two periods with a high pi'essurc ecllremaining o\'er the area 1'01' longer periods, one in late ~Iarch anti thc othcr duringthc first. half 01' April. In these periods NE winds domillated tlw meteorologiealeonditions o\"(~r the l3ay of Biseay.

Current meter mooring :3 lies on the shelf in water ~O m deep. The obscl'\:c<1 timescries for this eurrcnt meter, deployed at '·lm, is given in rig.. 5a. Ovcr the shd!",currents were oscillatory NW/SE with periods around 8 - 10 da,)'s. ;\Ieiln values01' the cast and north compollents \\'ere 0.-12 em/s and -0..17 em/s, ;Illd maximumvalues were l:l cm/s and - ·1.5 cm/s rcspeetively 1'01' the pcriod.

1\Iooring 2 was deploycd at the shclf cdge in a water depth 01' .568 nl. The topcurrent meter at mooring 2 was deployed at 75m and tllC' tilllC' ;,;cries is given infig. Ga. At the bcginning 01' the sampling period, currents "'ere mainly eastwanl,eorresponding to the winter eirculation in the southern Bay of Biscay. :\roundthe middle of 1\Iarch, the mean direction of the currellt revcrsed and now showedwestward flow, whieh is usually ooserved during spring amI summer (Pingrce andLe Cann, 1~90; Diaz del Rio et al. , 1992; Porteiro ct al., 19~G). l\Iean valuesfor the cast and north eomponents were -:1.2 em/s and -:3..5 cm/s, and maximumvalues were -18 cm/s and -12 em/s respectively for thc pcriod.

The seeond current meter a! mooring 2 was deployed at 180m depth alld the timeseries is gi\'en in fig. 7a. Thc current was predominantly eastward throughout theperiod and only a! the end of 1\larch ami at the end of April west.ward direetedcurrents were ooserved. 1\Iean values for the east and lIorth componellts were 1.7C111/S and -2.5 cm/s, ami maximum \'alues were - 12 cm/s and -~ cm/s respectively1'01' the period.

In this area of the Bay of Biscay, dellsity gradients are very small anu winu stressis the l11ain forcing factor of cii'clilation in the area. The x-component of thewind stress sccms to oe corrclated to cast component of thc current in thc mixedlayer. O\"cr the continental shclf salinity stratification was ooservcd in winterand during spring salinity ami temperature stratification was obsen'ed above theupper current meter. Intcnsification of the currcnts in the central period couldbc duc to a lack of stratification.

At the shelf break at the beginning of the samplillg period the upper :WO m ofthe water column was nearly homogeneous, out in 1\Iay watci's in thc vicinityof 11l00rings 2 and 3 wcre stratified. \Vestward currents are forced by a high

6

•

•

•

•

• pressurc cell over thc I3ay of I3iscay. If thc situation is not persistcnt, as was thecase aroulld February 28 amI !\Iarch 12, a pulse 01' wcstward current 01' 1:2 cm/s01' 14 cm/s at 75 m depth and weaker currents at ISO 1Il depth were observed.Ir the high pressure cell ispersistent, westward Oow persists over thc period at75 m dcpth as was the case iil thc sccond half of ~Iarch anti April with someOuctuations in intensity, and reduced wcstward Oow is found at ISO m depth.

(b) ~10DEL VALIDATION

Thc horizontal model grid resolution is 13 x 18.5 km in the Bay of Biscay. Silllu­latecl CUlTent data is taken from that model grid box which cncompasscs the areain which the current meter mooring was deployed. The appropriate layer is cho­sen in accordance with thc dcpth of t11e currcnt meters. r..10del data is availablconce a day alHlwas extracted at the specific grid points and layer depths. Timeseries were constructed for the period February to May 1995 amI were comparedto thc obscrved currcnt time series.

The modelled time series 1'01' mooring 3 is given in fig. 5b. It ('all Iw seen thatmodelIed currcnts speeds orten underestimate observcd LUlTeHt speeds. C'urreHtOuctuations in thc u - componcnt (cast - west.) are reprodliceJ by tbc model.although peaks do lt~t always coincidc. SilHulated ClI1Tcnt \'ariability in tlw v ­component (north - sOllth) is rl'duced comparl'd to the observations.

Thc corresponding model time series (model layer 5) 1'01' mooring 2, top currentmeter is shown in fig. Gb. Again current speeds are generally unJerestimated bythc model bot11 in thc u- and cspecially thc v - components. Current fluctuati­ons are reproduced by the model aud usually peaks in current speeds betweenobsen'ations and simulations coincidc.The model timc series (model layer 7) for mooring 2, bottOlll current meter isshown in fig. 7b. 1\lodelled and observed time series show many similaritiesfor this currcnt meter. r..la:ximum observe'd currcn t speeds are I"v 12 cm/s whill'ma.ximum modelled current speeds are I"v 10 cm/s. Current variability is similarfor observation and simulation both for the u- and v-component and usuallycurrent speed peaks coincide.

(c) HAKE EGG AND LARVAE DISTRIBUTIONS

Figs. Sa an<! Sb show the distribution of st.age I hake eggs Oll the first crllis(' alldG nun larvae on thc second cruise respectively. Uoth larval <llIll egg abulldallceswere an order of Illa~nitudehigher ill ~larch titan they were ill February sltowillgthat thc secoild cruise \vas eloser to the seasonal spawning peak.

In February, stage I eggs werc found in two patches (Fig. 8a). A bigger patchextenclecl from 45° N to 4Go N having 10wer abundances on the central transect.Eggs \vere more abundant over the shelf edge both in the northern amI in thesouthern tips of the patch and extended over the outer platform (100-200 m)

7

in the northern and southern parts. Another, smalleI' patch \vas located in thesoutheastern corner of thc Bay of ßiscay (eastern Cantabriall Seal.

The second cruise started on 22 1\larch, 29 days after the end of thc first one, andwas completed on 30 March. Larvae bigger than G lllm SI. appeal' in t\\'o areas(Fig. Sb). A more extensive one in the northern part of thc sampling area, from·16° to .l/o 15' N, with main abundances around ·17° N over the shelf edge and20-40 miles off the shelf cdgc. This patch cxtcndcd further to the south down to·16° N over shelf waters 100-200 m deep. A second. snwller patch was observcdin the southeastern corner of the Ba)' of Biscay.

(d) TRACER SIJ\lULATIONS

Currents in the region around the Iberian Peninsula. were simulatcd for 19!H and1995. The simulated eurrents as \\'ell as the daily variance of the CUlTents for •February ami 1\Iarch 1995 served as input to the t.ransport model. :\n examplcof the eireulation field on 10. ~l. 19!)5 in model layers 5 anel G is given in figs. !JaamI 9b.

A simplified scheme of vertical distributions of hake eggs amI larvae in the Bayof Biseay base<.! on the field data eollected in the February emd l'darch cruiseswas used in the transport model. As most eggs and larval' were found in thedepth range of 50 - 150m (1\lotos et al., 1996), the tracers migrated randomly inthis depth range, i.e. tracers spend an equal amount of time in thc modellayersbetween 60 - 150m. The tracers thus [eeI the combined effeet of thc currents inthis depth range. This random migration is not light depcndcnt as the \'crticaldistribution of hake eggs and larval' was similar during night and da,\".

The distrihution of stage I eggs mound 1i February (fig. Si\.) was illtcrpolatcd Olltothe model grid (fig. 10a) and servcd as the initial distribution for the dispersionsimulation. Thc simulation was begun on li Februar)' CIS practically all stage Ieggs \vere caught on 17 and 18 Februar)'. The G mlll hake larvCle collectcd on •the second eruise bewteen 22 - 27 1\larch (fig. Sb) was interpolated onto themodel grid (fig. lOb) and served as the final distribution for comparison with thedispersion simulation. The dispersion simulation was terlllinated on 27 1\Iarchas most 6 n1111 larval' werc caught 011 26 and 27 March. except for those in thcsoutheastern corner of thc ßay of Biscay which \\'ere caught on 2~ ~darell.

This cmise sun'ey design enab!cd the testillg of the trallsport llIodd hy llsing t!H'distribution of stage I eggs at time Tl as initial distribution, sillllliating for :39days and cOlllparing the resulting tracer distribution at timc T2 with thc 6 mmlarvae distribution actually found during the second cruise.

Tracers were introduced into thc model domain at the appropriatc points givenby the interpolated egg distribution in fig. 10a ill the depth range GO - 150m on17 February. Low concentrations are in yellow and the highest concentrations arein \"iolet. Three distinct areas of high stage I egg concelltrations ean be secn. Thenorthel'll and centra! (at '15°) arcas lie elosc to thc shelf cdgc. Thc ccntret! patch

8

•

•

with the highest concentrations lies further off shelf than the northcrn patch.Another patch \vith lower concentrations than thc other t\\'o is evident in thcsouthcastern corner of thc Day of Biscay.

Thc interpolated larvae distribution of thc 6m111 length dass given in fig. lObshows six areas of highcr concentrations w.r.t. the surroullding areas. The hig­hest eoneentration was found in the north at ·16 0 JO' N and :) 0 \\1 and wassituated olT-shdf in deeper water (2000 - :noo 111). :\nother high concl'ntrationpatch was found in the southeastern corner of the Uay of Uiseay. Four sllwller,lower eOllcentration pntehes werc found along thc shelf edgc, sitllated in a bandstretching from ·16 0 50' N southeast.ward to ·15 0 ·10' N Oll the shelf side in wateroetween 100 - 200m deep.

Thc tracer distribution after 3!J days dispersion is givell in fig. 10c. Five areas ofhigher concentration W.r.t. the surrounding areas can be seen in thc distribution.Thc highest concentrations ",ere found at the shelf cdgc at .15 0 20' N (md in the:'southcasteru part of thc Day of Biscay. Three iIl'e<lS of high!..'r COIlccntratiolL caubc seen aligned aloug t.he shdf edge on t.lle sllelf side l"rolll ./GO :W' .\l \,0 _1;")0 ·10' N.

COlllparing the si mula ted distri bu tiolL (Iig. 1Oe) to t.lw obsen'cd la l'\'ae distri bu­tion (fig. lOb) it is evident that there are two mailL discr<~pancies. The obsel'\'(~d

high conccntration of l,u'\'ae in the north at ·lG 0 :30' N and ;j 0 \V is not evidentin thc modelIed distribution. Secondl)' t.he high simulated tracer eOllcentrationat thc shelf cdgc (at .150 20' N) eanllot be secn in the observed distribution. Onthc other hand, simLilatcd and ohsen'ed distributions are in accordance in thesoutheastern Bay of Discay and in the threc distinct higher coneentratioll areasalong the shelf edge from ·16 0 30' N to ·15 0 -10' N. .

5. DISCUSSION

As current meters measure at a specific point and llumericalmollels give clIrrents re­presentative for an area, i.c. a model grid box, it is commOll to comparc transportrates through a section of a lltuuber of moorings to the appropriate scctioll of a numberof model points. This was not possiblc in this case as moorings \\'cre located such thatno transport rates throllgh a section perpendicular to the shdf edge. i.c. pcrpcndicularto the dominant current direction could bc ealclllated. Thus vclocity time series ofcularian mcasuremcnts (currellt meter) are COlll pared to veloei ty ti mc series represcn­tative for an area of 240.5 km2 (model). This is one reason for model - mcasurementdisercpalleies.

A seconcl important eause for discrcpancies between obsen"ations and si;nulations is thchorizontal resolution of the windstress fields from ECi\lWF usecl in the simulations.These dnta have aresolution of 1x 10 ",hieh is subsequently interpolated onto themodel grid. This horizontal resoltition causes a spatial homogeneity of thc windstrcssfields which is not necessarily ohserved in reality. Sma11 scale fronts, which are linkedto strongly fiuctuating momentum Buxe::; at the sea slll-facc (\"OU Storclt, 198-1) 01'

9

mesoscale vortiees whieh givc ris~ to exccptional strong winds (Paulus, 1~)8:n cancause strong currents. These phenomcna arc not rcsoh'cd Ly thc windstress field use<!for the foreing of the eirculation model. Thus clIl'rent peaks in the observations causedLy such small scale phenomcnii will bc laeking in the simul'lted time serics.

The third major reason for diserepancies is to be found in thc densilY field u"ed inthe simulations. As no quasi - synoptic density fields for tltt' lllOdel area and thetime under eonsideration are existent, lllonthly climatological dcnsity fields are used(Levitus, 1!)S2). These da ta have aresolution of 1x 1° which is intcrpolakd onto thcmode! grid, Although these data give a good approximation of the 1Il0nthly densityfieltl for the area. under consideration, they wi 11 neyert heless bc different t0 I hc realdcnsity field for the time pcriod eonccl'lled. Different horizontal density gradi(~nts andl1lcsoscale eddies me to he expccted in thc real dcnsity field and these will inf!uenc<'the CUlTent obse1'\'ations.

Bcaring in mind the abo\'clllcntioncd eOlllmcnts, the agreC\llent betwe('n o1>s('1'\'('d ,mdmodelIed currenls is good and thc disercpaneics are mainly caused by local, \W.'soscaleeffects whieh are not resoh'ed by thc input data to thc circulation model.

The cgg/larvae dispersion exercise assumed that 6mlll lar\'(w oLsen'cd dlll'ing the se­cond cruise deriw'd frolll the stage I eggs ob8el"\'cd dlll'ing the first cruisc, w!lieh de\'('­loped at incuLation tcnlJwmtlll'es of arou!ld 12°C (~Iotos ct aL 1!J!J6). The a\'ailablcdata on hake egg dC\'e!opment amI hake laiTal growth rates was revised and rorlllS t.hebasis of this assumpt.ion (see section :k). This assulllption is also supporkd hy tlwSmallllllll1ber of larval' I'l.l'ger than Gmm found during t.he sccolld cruisc at :~ stations,These stations coincidcd with stations where (j nHll lan'ac were found. 11. lllUSt. benoted hcre that the catchability 01' plankton uds depellds, aillollg olher fadors, onthe sizc of fish larvac to be caught.. Large lan'ac are more mobile (lnd are aGle tocseapc the nds and t.hus thc abundances 01' large lan',1e obselTed during field sur\'cysare gellerally lInderestimated (Sherlllau and 1I0ney, lU'1). IIo\\'c\"cr, differellt authorshave reportl'd catches of rather large lan'ae ",hen using BONGO ncts. For iustance,Olivar et al. (1!JS8) reported ·10 C111 BONGO catches of C(\pe hake lar\'ae, ~Ierluecius

capensis, up to a Icugth of 16 llUll, with a. unifol'll1ely decrcasiug frequeney histogralllfrom 3nll11 to 11.5m111 and discontinllous catches aftcl'wanls. Taking illto account l!Ieuse of 60 C111 nONGO nets iu the current in\'estigation, it cau Ge cxpected that lar\'alabundances in the eateh are an cstimation 01' the field ahundances of larval' llntil thl'Yreach about 10 m111 length. Furtherlllore, small changes in gl'owth rates (cg. 0.0·1mm/day) resldts in a larvae length diffcrcnce of 1 111m after 25 days. These factorsneed to be considered when comparing the simlllated tracer distributions with theobserved G lllm larval' distributions,

During the second cruise high llumbers of 6 mm lan'ae were found on thc two northernsections (fig. Sb). Duc to bad wealher these two sections, exc('pt for foul' statiolls,

.were not salllpled on thc first crllise. This could imply t1wt CI sizl'abl(' portioll 01'

egg production in this area was missed, Furthermore, egg abum[anccs 01' stages ülderthan stage I peakcd in thc llorthcl'lllost station 01' thc Februar)' cruise, showiug thatthere were high egg abundanees in the area. These zero stage I egg aGundanees in

10

•

•

•

•

the far north, which are used in the initial'distribution in the tracer simulation could,at least in part, account for the cliscrcpancy bctween larvae observations and tracerdistributions in the far nm,th. '

The high simulated tracer conccntration at the shelf edge (at ,15° ~O' N) origiuates frolllthe high egg concentration at 45°N (fig. 10a) which has been dispersed northwestwardsalong the shelf edge. This high concentration is not evident in the larvae obsen'ationsand may be duc to either enhanced advection onto the shelf in reality, high mortalitywhich is not considered in the transport model 01' stronger advection along the shelfedge than given by the model system.

The European hake has several known juvenile nursery areas (Anon., 1996). In the Bayof Biscay, a main nursery area is 10cated off the northern coast (Southern ßritauny),where hake juveniles (0 group) concentrate from 120m water depth to the coastlinefrom the end of thc spring to the autulllll (Bez ct al., lU95). llo\\'c\'er. bakt' eggs in tbi~

area in Februar)' aud ~darch are distribuled wcll offsbore lllaillly o\'t'r tb(' sbelf \'dg<'(l\lotos et al., 19%). Eggs and larvae with tll('ir lilllited acti\,(, borizoulal S\\'iUllllillgabilities, must therefore be transported towards the coast o\'er the 5helf into the llurseryareas. Prevalent residual currents in the area and diffusion processes will determineif the eggs and larvae migrate towards the coast, remain in the spawning areas 01'

. are dispersed offshore. Data on recruitment in the Northern ßiscay area show thatsettlement of juvenile hake starts in the outer part of muddy bottoms and that aprogressive displa.cement proceeds towards the inner basin (Guichet, 1988). At anyrate, larvae must appeal' over the platform to have chances 01' settling onto lllUtidybottoms and a general transport across the shelf would enable the dispersiou into tbenursery areas.

The results 01' the tracer simulation show that northern transport Was prevalent in thearea in the relevant layers where the eggs and larvae developed (Fig. 9). In addition,it also shows drift onto the shelf (Fig. 10e). These conditions, both northern transportand drift onto the shelf, favour the arrival of hake larvae in coastal areas. In the fielddata, a patch of 6mm larvae extended over the platform between .!G0 and ·n° 15'N atwater depths from 120m to 160m. These larvae are probably approaching the nurseryareas which start at 120m water depth (Guichet, 1988).

1I0wever, the main patch 01' Gmm hake larvae is found along <lnd off tl1(~ shelf edge at47°N. This larval patch roughly Coillcides \Vi th a me::;o::;cale gyre ::;tructl1l'e lcyclonic,70 km diameter) identified in the arca (fig. 11). The gyre was located off shell' andvclocities were of the order of 10-15 clll/s (l\lotos ct al., 199G). This gyre can entrainouter shelf and shelf edge waters and conscquently rctain biological material, c.g.fish eggs and larvae 01' it can advect the biological material towards offshore areas.This mechanism would act against the onshore transport of hake eggs and larvae intonursery areas and can potentially be detrimcntal for thc success 01' the recruitment.

Anothcr known nursery area is the southcastern corner of the Bay 01' ßiscay at '" 2°W(Sanchez, 1993). There, settlement to thc bottolll starts in water dcpllts dos<' to ~OOlll

an<! juveniles progressively Illove tO~\'<nds the 100m depth contour. Afteni:ards, they

11

disperse over the nursery area located between 100 anel 200m water depth (Farinaand Abaunza, 1991; Sanchez, 1993). In this area, a patch of stage I eggs anel anotherpatch of 6mm larvae w"ere found in the first and in the second cruise, respectively. Thissuggests that eggs spawned in February were retained in the arca and de\"clopcd to6mm larvae by 22 1\larch.

Alternatively, the occurrence 01' 6111111larvae cau result frolll thc (lCCUlllulation 01" larvaecoming from tbe coastal areas of thc Cantabriall Sea.. The gClleral circulation ill thcarea often shows a.n castwarcl transport elose to the coast amI a recurring cyclonic gyrein the soutbeastern part of the Bay of Biscay. This can result in the acculllulationof biological material at the eastern tip of thc Cantabrian Sea, where the directiol1pattern of the currents is variable and the speeds diminish. The general castwardtransport and northwestcrly wind regime may explain the entrainmcnt of hake lar\"aeon the Cantahrian shelf during winter months (Sanchez, 1993; Cabanas, 1993). Thisretention proccss in the southeastcrn part 01' the Bay 01' ßiscay was correctly silludated •by the model system.

ACKNOWLEDGEIVIENTS

This research is fundcd by the Europeall Union under grant NI'. AIR2 - CT93 - 1105. 1'hanksgoes to ECl\lWF (European Centre for l\Iedium-Hallge WeatlH'r Forecasting) in HC'ading anelthe DKRZ (Deutsches I~lirna Rechenzelltrum) 1'01' lIlaking c\.\'ailable th(> lllt't('orological dalafor the circulation model. We are grateful to the Spanish Instituto Nacional de l\letcorologiafor providing us with meteorological time series at Cabo Penas. Adolfo Uriarte amI PaulaAlvarez collaborated in the processing anel analysis of the hake egg allel larval cruise data.Finally we would like to thank aH SEFOS colleagues who participated directly 01' indirectlyin this investigation for their contribution. •

12

•

•

REFERENCES

ANONYl\10US, 1996: \Vorking Group on thc Assesslllcnt of Southern Shelf DClllersalStocks. ICES C.t-.r. 96/ Assess:5.

ALONSO J. amI G. DIAZ deI RIO, 1996: Currcnt mcasurcmcnts in Cantabrian sea. Ananalysis of tidal currcnts (submitted to Oceanologica Acta)

ßACKIIAUS, J. 0., UlS5: A tlm~c-dimensionallllodel for the silllulation of shdf sei\. dYlIa­mies. Deutsche Hydrographische Zeitschrift, 38, lieft -1, 165 - 187.

BACKllAUS, J. 0., 19S9: On the atmosphcrically induced variability of the circulation ofthe Northwest Eüropean shelfsea and related phenolllcna. In: Davies, A. ~1. (cd.): ?-.lodclingt-.larine Systems, Vol. I. eRe Press, Inc. Boca. Haton, Florida, ~n - l:n.

BAILEY, K.t-.r., 1981: An analysis of thc spa.wning, early life history and rccruitlllent of tlH'Pacific hake, Merluccius productus. PhO Thesis. Univcrsity of Washington, t981.

BAILEY, K.t-.l., FItANCIS, R.C. amI P.R. STEVENS, 198:2: The life history and lishcry ofPa.cific \Vhiting, t-.lerluccius productus. CaleOFI Rep., Vol XXIII.

BARTSCH, J., and R. KNUST, 199·1a: Simulating the dispersion of vertically llligratingsprat larvae (8]wattus sprattus (L.)) in the Gcrman Bight with a cireulation and transportmodel system. Fisheries Oceanography, Vo1. 3, NI'. 2, pp. 92 - 105.

ßARTSCH, J., an<! R. KNUST,I!.J9.tb: Prcdicting the dispersion of 8prat 1<U'vac (S']JI'attlt::;s]H'attus (1.)) in the German Bight. Fishcries Oceanography, Vol. 3, NI'. -1, pp. :292 - :296.

BEZ, N., J. RIVOIRAHD and J. Ch. POULAHD, 19!)5: Approchc transitive ct densites dcpoissons. Cahiers de Geostatistiquc 5. '

BORK, 1. anel E. 1\1AIER - REIl\lER, 1978: On the spreading of power plant cooling waterin a tidal river applied to the river EIbe. Advances in \Vater Resourees, 1, 161 - 1GS.

CABANAS, J.I\1., 1993: Caracteristicas oceanograficas de la plataforma continental atlan­tica de la Peninsula Iberica y su pobislbre relaccion con la distribucion de 10. mcrluza. InGonzalez-Garces, A. y Pereiro, F.J. eds. 1994. Jornadas sobre el cstado actual delos eono­cimientos de las poblaciones de merluza que habitan la platafo1'1na continerital Atlantica yMediterranea de la Union Europca con especial atencion a 10. peninsula Iberica. Publicacion

privada, pp. 255-279.

CASARINO B. and L.MOTOS, 1994: Identification and distribution 01' hake, M erlucciusmerluccius, eggs and larvae in Bay of ßiscay waters. Annex to the 1st SEFOS Annual

Report. Aberdeen, 1994.

COOMBS, S.H., 1994: Identification of eggs 01' hake, ~lerluccius merluccius. J. mal'. biol.

Ass. U.K., 74: 449-450.

13

COO1\lBS, S.lI., amI l\lITCHELL, C.E., 1982: The dc\"clopmcnt rate 01' eggs and larvac ofthc hake, Merluccius merluccius (L.) and their distribution to thc west 01' the British Isles.L. Cons. int. Explor. 1\Ier. ,10: 119-126.

DIAZ del RIO C, J. ALONSO, 1\1..1. CARCIA, J.1\1. CABANAS and J. 1\[OLINEHO, 19!J2°:Estudio <.1inamico en la plataforma contincntal dei norte de Galicia (Abril-junio 19!)1). Inr.Tee. I.E.O. NI'. 13-1, 2:3 pp.

FARINA, A.C. and ABAUNZA, P., 1991: Contribueion al cstudio dc los juveniles de merluzaentre Cabo Villano y Cabo Prior (NW Galieia) mediante prospeeciones pesqueras. Bo!. Inst.Esp. Oceanogr. 7 (2): 155-1G3.

GODIN G., 1991: The analysis of tides and currcnts. Tidal hydrodynamics. EJitl'd by B.Parker 1991.

GUICIIET, R., 1988: Etudc dl' Ia. croissClnce du lllerlu Cllropeen (l\Icrlut'{:ius lllcrluccius) ilU •

conrs dc ses prl'miares annces. Analys par NOR1\lSEP des distributions en taille obser\"cestrimcstrielImcnt en mer de 1980 - 1987. ICES, C.1\1. 1988/0:53.

INSTITUOT NACIONAL dc 1\IETEOROLOGIA, 1995: Boletin mcteorologico diario, 1\Ia­drid, February - 1\Iay 1995.

LEVITUS, S., HlS2: Climatologieal Atlas 01' thc \Vorld Oeean. NOAA Technieal Paper, 17:lpp.

l\IARRALE D., P. ALVAREZ and L. 1\IO'1'OS, in press. Early lire dc\·clopillent. and idcnti­ficat.ion of the European hake, l\lerluccius mcrluccius L. Ozeanografilm (In press).

1\101'OS, L., P. ALVAREZ and A. UIUARTE, 1996: Spawning of hake, l\lcrluccius mcrIue­cius, in the Bay of ßiscay in wint.er 1996. ICES Cl\I 1996/S:8.

OLIVAR, 1\1.P., P. RUBIES ami J. SALAT, (1988): Early life history and spawning 01'Merluccius capcnsis CastcInall in the northern Bengllcla Currcnt. S. Afr. _J. mal'. Sei. 6: •2-15-254.

PAULUS, R., 1983: Vortcx - like cloud strllcture o\"er the North Sea. Beiträge zur Physikder Atmosphäre, 56, ,W5 - ·lOG.

PINGREE R.D. and B. Le CANN, l!.HJO: Structure, Strcngtll and Seasouality 01' t.he Slopeeurrents in the Bay of ßiseay Region. J. mal' bio!. Ass. U.K. 70, 857 - 885.

POREBSKY, J., 1975: Application 01' the Surface Adhesion Test to identify the eggs of thehake l\IerIuecius spp. Colln. scient. Pap., Int. Commn. SE Atl. Fish., 2: 102 - 106.

PORTEIRO C., J.l\L CABANAS, L. VALDES, P. CARRERA, C. FRANCO and A. LAVIN,1996: Hydrography features and dynamics of Blue \Vhiting, l\Iackerel in the ßay of ßiscay.1994-1996. A multidisciplinary study on SEFOS. ICES C.l\1. S: 13

RUSSELL, F.S., 1976: The eggs amI planktollic stages of british lIlarille fishes, LOlldoll,Aeademic Press, 524 pp.

14

SANCHEZ" F., 1993: Patrones de distribucion y abundancia de la Illerluza eu aguas dela plataforma norte de la Peninsula Iberica. In Gonzalez-Garces, A. y Pereiro. F.J. eds.1994. Jornadas sobre el estaclo actual c1elos conocimientos c1e las poblaciones de merluzaque habitan 1a plataJorma continental Atlantica y Meeliterranea de la Union Europea conespecial atencion a la peninsula Iberica. Publicacion privada, pp. 255-279.

SHERMAN, K. and K.A. HONEY, 1971: Size selectivity of the GulE III anel bongo zoo­plankton sam pIers. ICNAF Res. Bult., 8: ·15 - 48.

S1\IITH, P.E. amI S.L. RICHARDSON, 1977: Standard techniques for pelagic fish cggs andlarval surveys. FAO Fish. Tec. Papers, 175: 100 pp.

STAUENO, P. J., HERMANN, N. A., BOND, N. A. anel S. J. BOGRAD. 1D9.S: ~Iode­

ling the impact 01' climate variability Oll the advection 01' larval walleye pollock (l'/w/'ag/'a

chalcogl'amma) in the Gulf of Alaska. In: Climate Change anel Northern Fish Populations,• R. J. Beamish (cd.), Can. Spec. Pub!. Fish. Aqu. Sci., 121, pp. 719 - 727.

von STORCH, H., 1!)84: A comparative stuely of observed anel GCM - simulated turbulentsurface fluxes at the positions 01' Atlantic weather ships. DYllamics 01' Atmospheres andOceans, 8, 343 - 359.

\VERNER, F. E., PAGE, F. H.. LYNCH. 1). H., LODER. J. \V .• LOlfGH. H. C; .. PI~:Inn·.

R. l., GREENBEHG, D. A. and M. 1\1. SINCLAlH, I~H):3: luflllcuces ur 1l1CcUI ddvccLiutl amIsimple behaviour on the distribution 01' co<! amI haddock early lire stages on Geürges Bank.Fisheries Ocea,nography, Vol. 2, Nr. 2, pp. ·13 - 6"1.

•

15

~------------ -- ----

Fig. 1 Southern SEFOS model area and topography showing the 100m, 400m, 300m and2000m isobaths. .

M 2-Tide, Wiudst.ressand Dcnsit.y Fields

Circulnt.iou ModclSout.hern SEFOS Area

•

Fish LarvneDist.ributions

Sca Sm'face Elevat.ion3-D Cmrcnt Fields •

'l'ran~port Mod(!l(Advect.iou, Diffusion nlld

Vertical Migrat.ion of Lnrvac)

Drift RontesTn\cer Distributions

Fig...2 Schematic diagram or model system. Oval boxes denote input and output data.

..45° 500m

200m100m

44°

·90 _8° -60

• Fig. 3 Location of moorings. Mooring 1 was deployed at 1150 m depth, mooring 2 wasdeployed at 568 m depth at the shelf break and mooring 3 was deployed at 90 m depth onthe continental shelf. The 100 m, 200 m and 500 m isobaths are indicated. ,

4

4

N

Bn.BAO SAN sEBAmAN

~·o~. "CVCt[jU \

~-""'" ..· 7 \

_[,0 ~' . i\'j~. ~~-~~. ~j.

" :. ,c 6:c \

-""--"" .•,~ •. 11

CTD

­o-rit°o

-CTD

O_ONCG-4lI• 80Nc0.40 + erD

00.......1 MI I., ~~..

• Opelll../d Mt 2" e"'"

';""CTD SEroS .._..-tI

100.... ••

0::'

-'-"""!1J~ •.•.•y'':~:; \. ". .' 00 ,

0; '. otea;

(,orD : .. ,fP O'CIO ~'o • ~

- °41oo

W6 5

"

Fig. 4 Sampling grid of the hake egg and larval surveys carried out on board RV INVESTI­GADOR in 1995. Two consecutive cruises covered the time periods 15 . 24 February alld22 - 30 March respectively. The same grid was to be covered in both cruises. See legend fordetails on the sampling scheme. Areas surrounded by closed lines were not covered duringthe February cruise.

O'1/S

("HIS

FEB HAR ""PR HAY

•FEB HAR ""PR HAY

Fig. 5 a) Observed current components (u bottom, v top) at mooring 3 (74 m) betweenFebruary - May 1995.

Oll'

..~ .

•CM/S

FEB HAR ""PR HAY

FEB HAR ""PR HAY

Fig. 5.. b) Simulated current components (u bottom, v top) at mooring 3 in modellayer 5(80 m) between February - May 1995. .

•

CM/S

l- --L -..l. --I- --'

('"HIS

FEB KAR APR MAY

,.~ ,

. ,. .L-- --L .1-- ....L.- --'

FEB KAR APR KAY

Fig. 6 a) Observed current components (u bottom, v top) at mooring 2 (80 m) betweenFebruary - May 1995.

CHIS

,,~ .'L-~__----L _L___'__ __l_ ~

CHIS

FEB KAR APR KAY

.,~ .

FEB KAR APR KAY

Fig. 6 b) Simulated current components (u bottoin, v top) at mooring 2 in model layer 5(80 m)' between February - May 1995.

"~ .

CHIS

CHI!';

FEB HAR APR HAY

" .L- -L.. -L.. _

FEB HAR APR HAY

Fig. 7 a) Observed current components (u bottom, v top) at mooring 2 (180 m) betweenFebruary - May 1995.

CHIS

alls

FEB HAR APR HAY

..~ .

FEB HAR APR HAY

Fig. 7 b) Simulated current components (u bottoin, V top) at mooring 2 in modellayer 7(175 l~) hetween February - May 1995.

J

--------1

~(;v ",,

N

48

NANTES

47

SAN SEBAS11AN

•

'".-..'

••••RN "INVESTIGAVOR"

16-12 Ji'ebruary 19~.'S

STAGE I HAKE EGGS <111 1_1)

GIJON

.-".....+....:

BILBAO

f- 45

44

w , 5 4 3

Fig. 8a: Hake stage I egg abundances found during the survey period 16-22 February 1995 (mid date:19 February). Egg abundances are given in numbers per 10m2

•

NANTES

lvI) 11'

. -< h)(" "'_ ..'"

RN "INVESTIGADOR"11-17 M.rcla 1995

'mm HAKE LARVAE (IIIIOm1)

.'..'

,,­.;.'

'".~

-""

SAN SEBASTIAN

N

48

47

45

44

43

w , 5 4 3

Fig. 8b: Abundances of6 mm hake larvae found durlng the survey period 22-27 March 1995 (mid date:25 March). Larval abundances are given in numbers per 10m2

,

•

././,/./ ••••••• "\\\1\\1.,, ••••••• 1·'"~. _ .. ., ••• ,. ...... Ilt 110\ \ \ \,." ••• ." •• ~",., t \

• 1II .. " 11 " •• \ Ilt " , I ".... , - ".,.......................... 11 ••• .,., .. "",;'"- -", \"

...... ,."".,. ....... ., ...... " .;,/tJI- , .,.,." /,;'tI' ",

".",. ...... " ••• •••••• 11 ... • " •••• 11 '·· ..f/I .. ,. ,. ,. ,. '4 • ,. " " ,. ,,, _-'I" ,. Ilt •• , t • t ..

• , • , • t , f .. ., , , .. "tI'" ~ __ 1" __ ,. " " ' ....

.. " tI .. .. • • .. I' 'li 111111 1111 I ,. ".. .. .. I I

.. "" ., "' .. ,. ..... ,. llllillfil 111:'1:11,11111'111111:,,,,1 111l'l 11 1.... , • 't ,1t ,., 111 ,)1;111 11::111 1,III:lllllfllllllllnl,

111"" 11 ,. llllll'll!II"I'IIII'III'lllllllfflltllllfll I........ " •• t I1I 111'1".lIIl'I"IIIl'II'lllllf 1111I I•• " •••• + + + + ••• ., ,1111111'1.1111.111 I1 111,' t 1/111111 1I11 1

'I 11' '1"1 1/"'1"" "" 'I""" ,::::::::::::::::.' :\ 11 11" 'I" '111 I' "1

111'1,,1 'I..: ::::::,' I I'!III 11 :

·}::::::"::::::::::::::::/;IIIIIIIIIIIIIIIIIIII, illlitlli'llllllilll!111•• " •••••••• ,. " 111111111111!!II:I:I !lli!i!lll. 1.11 Illi !li••••••• , •••• • ••• , ::111111111111111111111111111111111111111 1I

......... " .. '&, •••• ., • • .. f 1111"1111111"111"11"11"111"1" 'I" I"•••••• , ••••••••• ' :11,111101111111,11 1/1"' oll 1 111 I•••••••• , ••••••• ' I I ' Ii'lll'l I 11 1/111111'1 1111 11• , '" • , " \ \ \ , - • • ., 1 - ' I! dl: I I !:! HIP ddi I d q

....... , "",\. .. " ------, ':::11 t 1::1111:!1111'11 11 1, I• , \. \. \ , I _ - -_........... 'li I 111 t I11 11 I I t I

.- . , " "" " \. \ \ \. " ,,-, , ili! 111 HPiPl:l d :i I., ,\., "",,,,, ,'" I' ::11 I Illlll!:dlll 11 11 :

--_ ,,,,'\.,, .. , ,"""' ...--,-~, 1:11: I :llllll:l1111 11 11 J...... ", , """, ---- , 1111111' l"lltlllll I 11" ,......... .. , , , "", _ ......... _, ':111111111111111111 I I1 l,••• -. ,\\, ••• ,. 'I' 1IIIII/1llll1lllll!II!II!lll llili•••- , •• - , , \ \ , ••• "li ,1111111111111111111111111111 1I11...._\/./."' .. ,\\ •••• 1I1 111 111111111 1111 I

9 b)

9 a)

--_.-.-.-. ••••••• \ \ \ \ \ \ \ \ \ • , , , •• \ • 1 '" \ ,,.,.- "",..." '" ~ 4 • " .. " " 1Ii \ \ \ \ , , \ I " " • • .. \ ,

• ........ 111 • " " " " 111 " " , \ \ , •• _ _ • • ' .... \ \ , .......

........ ", •••• 111" •• • .. • .. ' \"", ...... .. .. l1li , • 11 11 .. '" , , I , ~ __ ", '" " ,.......... - " ••• "/I'/~· .. ···""·.".~,,,· '"

... ,·,· .. ·······,.,,"1111 ··_ •• _'1 ••• ...., ••••• "." ••• , •• "\\ ...... _, ,,, ••• , ••••• I1

...... " ............. " ....... , _ , •• ,..-_ '" .... __ __ , .. " ..... 11 t ' .. " I J;...................... lIt .,,- 1 I. i ' ... .. .. :I:

---.-.-- ••• • • • , , • ••• " ,1'1 11 '/11 1/'1'/"1 1'11" 1 'I 111 '111'1:, I ••••• " t I I I I I I 11: I I I: I I' ,\ \ •••. ,., ~ ...•.. , 'lill ,11!!!I;111111111 11111 !!!I111111 1III1· . , " , , 'I" I , '11"'11 "" 111 I '1 11"1' 1/1'1'"................... 11" .. "." "1111111 1.111 • .11111 "lI I I11 I1 'l'd:1

'I" 'li ''''11'''1'11'1'' '/"'1"11" 'li'.......... ,. "" II11 11111111'11111111 IIIt I;

.:::::::..: '.',' '!II,II! II!lIllil lll 11 11111 '1/l1!11 1111111111, 1"'11

'I' 1'1 I' I "" '1"1 I 1111'" •, 11,,' _ r t /' ,I 'I 11 1 1 '1.1111 11 11',1. I 1111 1I lltl"...... 111 _ ,11 t .I 1(1 111 11111 11 11 " 11I11f11l 11111)

..... " •••• - ,," • • " " .. , i!lllilli Id:lHli!: Hf!' i!h1llli !!!l!1.... , 111 "" 111" " •••• /' IIH I 11I 111111111 11 1111 111.11111 111111

1111/1 "I "1"1111111111 "1111"1111111, " 111' I '''·11111111 I1 lldl 111 II 111; Iltllftt: 111111••••••••••• • \ \. , 1'11"111: 1'1'11'11 11 '1'1 1I1111I11 IIII11., - ,\ \ '\... 1IH:::: :1:1:11:: JJ !l!1 11 1111111 dlll i

......... , , .. \. " .. ,. l:lnlll 111::::11 1111::1 :llltll'll lfll:f•• - _. " " -" It ·'_:1111111 IlIlllllt Illftlt llll:llllllll:

, •• " 111 " " :::1: If ::1:,1:: 1::1:11 JlJI:IIII !'I" I" "" ,..... I 1111111111111 111111111 11 I

• .. • • •••••••• • • •• - .... 1· li!!I!I!III!1 illl!!!ll!!l!1 I,:: 111': I: 'I 111': 11 '1111' I, ol"ol, ",I." 1,,,.1 ,I"11"I'ltlllIlIl 1" 11 1 I'111 :i I "111/11 .1111, I

IllI!!lllllllll' P!III !I,1I' I'!1111 1 I' I!11'1 ":I, I' , ,ol : "I, 11,

1'11/111'111/"11 '1'111 'I:I 'I" I,,, lol", I"

1111111.1111111111111/111:111111111: 11 f 11I f ::11 fI : !: :

Fig. 9 Simulated circulation field on 10. 3. 1995 in modellayer 5 (SOm, (a)) and In modellayer 6 (125m, (b)). Smallest arrows denote speeds< 2 cm/s. As arrow size lllcreases,current speeds increase by 2 cm/s increments..'

•

•

10 a)

10 b)

10 c)

44°N

44°N

Fig. 10 a) Observed hake egg (stage Ia and Ib) distribution in the Bay of Biscay on 17February 1995 interpolated onto the model grid. Initial distribution of tracer simulation.

Fig. 10 b) Observed hake larvae (6 mm) distribution in the Bay of Biscay on 27 March 1995interpolated onto the model grid.

Fig. 10 c) Simulated tracer distribution after 39 days dispersion in the Bay of Biscay on 27March 1995.

f-==·~-·==-=~"==-=~·=··~~.--LÄ·';,~=-=oo1l=:,=::,,;=(=:-=-i-=:~=.-=-~~~-'~_-!-_~=====~'

; '~.BREST1r-, -~:~~

"/ ':~"'-' LORIENT

•

•

- ZOOm

44°';I 200m •.

I '., -100m-

..... : i .

-'~".n~,," .

.,.J '-, ' ,-.-.-•.. ' i

(>, '1 LA ROCHELL? I

\, ;_1'" :

'. \.\\\\ I

Fig. 11: Dynamic heights (0.2cm intervals) from 10/100 dbars, indicating the general geostrophiccirculation (arrows) during the period 22-27 March 1995. Open circles illustrate the abundance of6mmhake larvae (the biggest circle represents 7 individuals ofthe 6mm lenght class per 10 m2

).

Table 1: Summary ofthe infonnation available on the egg development and larval growth rates ofhake.VaJues in the table are time (in days) to reach the different developmental stages: Hatching, end ofyolk­sac larva and different sizes ofpost-Iarva. In the latter, G refers to the Gompertz curve fit and L to theIinear fit to observed larval growth figures (Bailey, 1981). This author pointed out that larval growthwas better explained by a linear increment of0.16 nun d'l up to approximately the 32nd day of live. Initalies, time to the end ofthe yolk-sac larval stage assuming that the growth pattern of M merlucciusat this stage is similar to that of M productus. See text.

Merluccius productus - Haller (1981)ternperature 10 11 12 13

hatching 5.28 4.81 4.40 4.05

end of yolk-sac larva 14.68 12.73 11.08 9.68

length G L

post-Iarva 4 rnrn 22 18

5 rnrn 26 24

6 mm 30 317 rnrn 33 378 mm 36 43

M. rnerluccius - Coornbs & Mitchell (1982)1~ 10 11 12 13 15

3.49 6.82 5.96 5.27 4.71 3.85

7.50 16.47 14.09 12.12 10.49 7.96

G L

23 1927 25

31 3234 38

37 44 •

•