water masses and distribution ofphysico-chemical properties in...

Deep-Sea Research II 49 (2002) 585–602

Water masses and distribution of physico-chemical propertiesin the Western Bransfield Strait and Gerlache Strait during

Austral summer 1995/96

M.A. Garc!ıaa,b,*, C.G. Castroc, A.F. R!ıosc, M.D. Dovalc, G. Ros !ond, D. Gomise,f,O. L !opezg

aLaboratori d’Enginyeria Mar!ıtima, ETS d’Enginyers de Camins, Canals i Ports, Universitat Polit "ecnica de Catalunya, Barcelona, SpainbDirecci !o General de Ports i Transports, Generalitat de Catalunya, Av. Josep Tarradellas 2-6, 08029 Barcelona, Spain

c Instituto de Investigacions Mari *nas, Consejo Superior de Investigaciones Cient!ıficas, c/Eduardo Cabello 6, 36208 Vigo, SpaindDepartamento de F!ısica, Facultad de Ciencias del Mar,Universidad de Vigo, Apdo. 874, 36200 Vigo, Spain

eDepartament de F!ısica, Universitat de les Illes Balears, Spainf Institut Mediterrani d’Estudis Avan-cats, IMEDEA (CSIC-UIB), Carretera de Valldemossa, km 7.5, 07071 Palma de Mallorca, Spain

g In. nova Oceanografia Litoral, c/Independ "encia 361, ent. 3er, 08024 Barcelona, Spain

Received 10 May 1999; received in revised form 16 January 2001; accepted 15 June 2001

Abstract

In the framework of the FRUELA project, two oceanographic surveys were conducted by R/V Hesp !erides in the

eastern Bellingshausen Sea, western basin of the Bransfield Strait and Gerlache Strait area during December 1995 andJanuary 1996. The main hydrographic structures of the study domain were the Southern Boundary of the ACC and theBransfield Front. The characteristics and zonation of local water masses are discussed in terms of temperature, salinity,dissolved oxygen, nutrient and inorganic carbon concentrations. Concentration intervals for water mass labelling, on

the basis of chemical parameters in addition to the common y=S-based classification, are defined. Silicate seems to be avery good discriminator for local water masses. r 2001 Elsevier Science Ltd. All rights reserved.

1. Introduction

The Bransfield Strait is a semi-enclosed Antarc-tic sea located between the South Shetlandsarchipelago and the Antarctic Peninsula coast.The extent of the Strait is about 50,000 km2 and it

can be divided into three major basins which areseparated from each other by sills shallower than1000 m. The western basin of the Strait isconnected to the neighbouring BellingshausenSea through passages existing between the wes-ternmost South Shetland islands and also throughthe Gerlache Strait, and to the Drake Passage viathe Boyd Strait principally. The Gerlache Strait isflanked by the western Antarctic Peninsula coastand the Palmer archipelago. The sill connectingthe Gerlache Strait to the Bellingshausen Sea isabout 350 m deep.

*Corresponding author. Current address: Direcci !o General

de Ports i Transports, Generalitat de Catalunya, Av. Josep

Tarradellas 2-6, 08029 Barcelona, Spain. Tel.: +34-93-495-

8070; fax: +34-93-495-8196.

E-mail address: [email protected] (M.A. Garc!ıa).

0967-0645/01/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved.

PII: S 0 9 6 7 - 0 6 4 5 ( 0 1 ) 0 0 1 1 3 - 8

In the framework of the FRUELA carbon fluxstudy funded by the Spanish National Programmeon Antarctic Research, two oceanographic surveyswere conducted by R/V Hesp !erides in the easternBellingshausen Sea, western basin of the BransfieldStrait and Gerlache Strait area in December 1995and in January 1996. From the point of view ofphysical oceanography, the aims of the surveyswere to gain insight into the characterization of thelocal water masses including the details of theirintraseasonal variability during Austral summerand to establish the local mesoscale circulation inthe study region with a resolution which should beadequate to meet the project needs. This paperdeals with the distribution of water masses asobserved during the FRUELA 1995/96 cruises andthe companion article by Gomis et al. (2002)discusses the 3D circulation and mass transportpatterns.

The Bransfield Strait is one of the most denselyobserved Antarctic seas due to its geographicalaccessibility, its favourable sea ice conditions andthe existence of multiple logistic resources in thearea. Modern oceanographic research in the zonecan be dated back to the Discovery expeditions inthe late 20s (Clowes, 1934). The internationalISOS and BIOMASS programmes, which werecarried out during the late 70s and the early 80srespectively, produced a basic understanding ofthe local water mass structure and circulationpatterns on the basis of coarse-resolution surveys(see e.g. Gordon and Nowlin, 1978; Grelowskiet al., 1986; Tokarczyk, 1987). Hofmann et al.(1996) provided descriptions of the large scalewater mass distribution and circulation patterns inthe Bransfield Strait on the basis of historicalhydrographic data.

A number of eddy-resolving field studies aus-piced by the US and Spanish national Antarcticprogrammes targeted the mesoscale circulation atthe different Bransfield Strait basins in the late 80sand early 90s (Niiler et al., 1991; Garc!ıa et al.,1994; L !opez et al., 1994; L !opez et al., 1999). Inparticular, Niiler et al. (1991) carried out a detailedanalysis of the geostrophic circulation in thewestern Bransfield Strait region with a focus onthe circulation variability at seasonal scale.Although Niiler et al.’s (1991) work provided a

very useful reference for the oceanographic aspectsof the FRUELA project, it was clear from thebeginning that additional hydrographic observa-tions of the area of interest should be a keycomponent of the Spanish study. On one hand, theflow variability documented by Niiler et al. (1991)implied that ad hoc physical measurements wereneeded to compute fluxes of biogeochemicalvariables contemporary to data acquisition. Onthe other hand, Niiler et al. (1991) could notextend their measurements deeper than 200 m dueto limitations of resources and shiptime imposedby the funding programme. This resulted in anunderestimation of the magnitude of the majorlocal circulation featureFthe Bransfield Curren-tFpossibly by one half (Gomis et al., 2002). Giventhe presumed importance of the advective trans-port term in the local carbon budget (see Huntleyet al., 1991), reassessing the geostrophic transportwith a deeper reference layer was regarded as amust.

2. Material and methods

R/V Hesp !erides conducted two cruises in theFRUELA study area in December 1995 and inJanuary 1996, respectively. The December 1995cruise comprised two hydrographic legs. In leg 1(hereafter referred to as MACRO ‘95), a series of 5hydrographic transects were visited across themajor hydrographic fronts of the project studyarea. This leg was carried out between 3rd and11th December and was composed of 33 stations(Fig. 1a). The characteristic spacing between ad-jacent stations was about 17 nautical miles on eachtransect and the spacing between adjacent trans-ects was twice as long. Leg 2 (named MESO ‘95 inwhat follows) consisted of a higher-resolutionsurvey of the western Bransfield basin and theGerlache Strait carried out between 12th and 19thDecember. In this leg, 107 hydrographic stationswere visited and the typical spacing betweenadjacent stations was about 8 nautical miles. Asfor the January 1996 cruise, it included one singlehydrographic leg (MACRO ‘96) which replicatedthe MACRO ‘95 survey except for the western-most transect in the Bellingshausen Sea (Fig. 1b).

M.A. Garc!ıa et al. / Deep-Sea Research II 49 (2002) 585–602586

Twenty seven stations were occupied in theMACRO ‘96 leg from 21st to 27th January1996. During both MACRO ‘95 and MACRO‘96, a section along the Gerlache Straitwas sampled. As the main aim of this paper is toexplore the intraseasonal variability of thelocal hydrography and to discuss concen-tration thresholds of the local water mass label-ling, we shall only focus on the MACRO ‘95and MACRO ‘96 data sets. Both sets and theMESO ‘95 data are used in the companionpaper by Gomis et al. (2002) to compute thelocal 3D mesoscale circulation and related trans-ports.

On each hydrographic station, a surface-to-bottom CTD cast was performed with a GOMkIIIC WOCE probe provided with extra dis-solved oxygen, fluorescence and light transmissionsensors. Water samples were obtained routinely at24 levels with a GO Rosette equipped with 10 lNiskin bottles. Some of the Niskin bottles carriedRTM SiS 4002 digital reversible thermometers.Salinity was obtained from water samples bymeans of a Guildline Autosal 8600 B. Dissolvedoxygen concentrations were measured followingWinkler’s method, with an estimated error of71 mmol/kg. Nutrient concentrations were deter-mined by segmented flow analysis with Technicon

AAII autoanalyzers, following Hansen andGrasshoff (1983) with some improvements(Mouri *no and Fraga, 1985; !Alvarez-Salgadoet al., 1992). Alkalinity and pH(NBS) weredetermined by potentiometric methods accordingto P!erez and Fraga (1987a, b). The accuracy of pHand alkalinity, estimated using CO2 certifiedreference material, was 70.004 and 71.4mmol/kg,respectively. Total inorganic carbon (TIC) con-centrations were computed from pH and alkalinitydata by using the thermodynamic equations of thecarbonate system (Dickson, 1981) and the con-stants determined by Mehrbach et al. (1973) withan accuracy of 74 mmol/kg (Millero, 1995; Leeet al., 1996). TIC data were normalized at salinity35 psu by means of the formalism NTIC=TIC*35/S. Water sample analyses were carried out onboard.

The contouring package used to map thedistributions is the Surface Mapping System v.6.03 from Golden Software Inc. The y and Sdistributions have been derived from CTD profileswhereas water sample data (available on everyother CTD station) have been used for the spatialinterpolation of chemical properties. The landareas have been masked after the interpolation, sodetails of the interpolated fields near coastalboundaries may not be properly assessed.

Fig. 1. (a) Map of hydrographic stations where water samples were obtained during the MACRO ‘95 leg, December 1995 cruise. (b)

Same for the MACRO ‘96 leg, January 1996 cruise. The CTD station coverage of each survey is shown in Figs. 2a and 4a. The dashed

straight lines indicate the locations of the vertical sections represented in Figs. 6 and 7. The 0-m, 200-m, 500-m and 1000-m depth

contours are depicted.

M.A. Garc!ıa et al. / Deep-Sea Research II 49 (2002) 585–602 587

3. Local water masses and fronts

The water mass zonation of the FRUELA studyarea in terms of physical water properties is quitewell established in the literature. A summary of thelocal water mass classification used follows forreference. The conventions are similar to thoseused by Garc!ıa et al. (1994), Garc!ıa (1996) andL !opez et al. (1999), which were distilled fromliterature and were applied to analyses undertakenin the framework of previous Spanish Antarcticreserach projects. For the sake of clarity, thedescription has been splitted into three parts: (i)eastern Bellingshausen Sea and southern DrakePassage, (ii) Bransfield Strait and (iii) GerlacheStrait.

The typical water masses sequence in the easternBellingshausen Sea and southern Drake Passagesector consists of four main water masses:Antarctic Surface Water (AASW), Upper Circum-polar Deep Water (UCDW), Lower CircumpolarDeep Water (LCDW) and Antarctic BottomWater (AABW). AASW is a cold water massoriginated all around Antarctica in early winterand typically extending down to 200 m. Thehydrographic signature of the AASW core is asubsurface temperature minimum (typicallyyo01C) embedded in a strong halocline. BelowAASW, a two-layer system of Circumpolar DeepWater (CDW) occupies most of the water column.The difference in physical properties between bothCDW cores, upper CDW (UCDW) and lowerCDW (LCDW) corresponds to their differentextra-Antarctic origin and is subtle. The deepy ¼ 01C isoline is considered to be the boundarybetween LCDW and AABW. In the sector,AABW mostly consists of Weddell Sea DeepWater which is advected westward after leavingthe Weddell basin through deep gaps across theSouth Scotia Ridge and also further east.

The Southern ACC Front (SACCF) and theSouthern Boundary of the ACC (SbyACC) as theywere defined by Orsi et al. (1995) are the mainhydrographic features in the eastern Bellingshau-sen Sea and southern Drake Passage. The SACCFis the only ACC front that does not separatedistinct water masses. In the FRUELA studydomain, the SbyACC shows as a structure which

separates the ACC waters from an area ofrelatively vertical hydrographic homogeneity ex-tending on the South Shetlands continental shelf.In our case, the locus where the 01C isothermintersects the surface layer can be roughly con-sidered as the position of the SbyACC. Eastward-flowing geostrophic jets are related to both theSACCF and the SbyACC in the southern DrakePassage, although the SbyACC jet is not acircumpolar feature like the SbyACC itself (thehydrographic expression of the SbyACC is not adensity front everywhere).

From the hydrographic point of view, theBransfield Strait is best defined as a transitionzone between the Bellingshausen Sea and theWeddell Sea. The Strait basins are mostly occupiedby water masses whose properties are controlledby the characteristics of the inflows from theadjacent seasFa warm and relatively fresh inflowfrom the Bellingshausen Sea (typically 0.5–3.01Cand 33.1–33.9 psu in summer) and a cool andrelatively salty inflow of Weddell Sea surface anddeep waters (typically with negative temperaturesand salinity ranging from 34.1 to 34.6 psu insummer) -. The main local water masses can thenbe referred to as Transitional Zonal Waters withBellingshausen Sea influence (TBW) and Transi-tional Zonal Waters with Weddell Sea influence(TWW) depending on the dominant source waters.Both water masses are separated by a shallowhydrographic front which has a clear surfacethermal signatureFseveral summer cruises per-formed in the area suggest that the 1.01C isothermis a good discriminator between TBW and TWWin mid summer. TBW appears confined to anarrow stretch of the mixed layer extending alongthe northern half of the Strait whereas TWWoccupies almost the entire volume of the Bransfieldbasins. As mentioned before, modified LCDWmay flow into the Bransfield Strait through theBoyd Strait and other passages. When it occurs, itreplaces TWW and produces a characteristic deeptemperature maximum (y > 01C). Still anotherwater mass can be traced below the TWWlayerFthe Bransfield Deep Water (BDW). It isbelieved that this water mass is an isopycnalmixture of Weddell Sea shelf waters and WarmDeep Water which is advected into the eastern part


of the Bransfield Strait and may be furthermodified in situ by winter convection. The�1.01C and 34.5 psu y and salinity values definethe shallowest extent limit of BDW under TWW.

Two hydrographic fronts have been described inthe Bransfield Strait: the aforementioned TBW/TWW surface front and the so-called BransfieldFront (BF)Fthe density front extending along thesouthern South Shetlands continental slope, whichseparates TWW from waters sitting on thearchipelago’s shelf. Of the two, only the BF isrelevant from the dynamical point of view. Asboth fronts occupied the same geographic positionduring the FRUELA cruises, the combined frontalzone will be referred to as BF.

The hydrography of the Gerlache Strait hasbeen little discussed in the open literature, to theauthors’ knowledge. As a first approach, theGerlache Strait can be understood as a westwardextension of the western basin of the BransfieldStrait. Given the depth range of the GerlacheStrait and the shallow sill existing on its westernend, no BDW should be present in the area. Thetypical Gerlache Strait water column should thenconsist of an upper layer of TBW and an under-lying, bottom-reaching layer of TWW. Limitedintermediate intrusions of CDW could be expectedto penetrate either from the west or from the east.Local TBW should be fresher and cooler than inthe Bransfield Strait due to the influence offreshwater inputs from local glaciers.

4. Results

4.1. Horizontal zonation

Figs. 2 and 3 show the horizontal distributionsof y; S; oxygen, nitrate, silicate and NTIC at seasurface and at 300 m depth in the western basin ofthe Bransfield Strait and eastern BellingshausenSea as deduced from the MACRO ‘95 data set.Figs. 4 and 5 show similar distributions for theMACRO ‘96 data.

The MACRO ‘95 leg was carried out in earlyDecember shortly after sea ice retreat in theBransfield Strait. This is clearly noticeable inFigs. 2a and b, which display low sea surface

temperature and salinity values with respect to thecanonical local summer range of y and S: Contraryto this, the sea surface distributions of temperatureand salinity in late January 1996 (Figs. 4a and b)match better with the paradigm for local summerconditionsFin our case, the BF front can betraced through the 34.0 psu isohaline. A compar-able situation is found, e.g., when comparing datafrom the previous BRANSFIELD 9112 cruisecarried out in December 1991 (Rojas et al., 1996)with the distributions of the BIOANTAR 93 cruiseperformed in January 1993 (Garc!ıa et al., 1994). Asimilar but less pronounced intraseasonal changeof temperatureFof about 0.21C between Decem-ber 1995 and January 1996Fcan be noticed in the300 m distributions of y in the southern half of thewestern Bransfield Strait. The signature of the BFis visible in both deep distributions of y: Thecomparison of the temperature distributions sug-gests a limited variability of the BF between thetwo study periods. The deep horizontal salinitydistributions show very little structure (Figs. 3band 5b).

The sea surface distribution of oxygen forMACRO ‘95 shows a zone of high oxygenconcentrations (O2>390 mmol/kg) coinciding withlow nitrate (NO3o20 mmol/kg) and low NTIC(NTICo2150 mmol/kg) levels in the northeasternBellingshausen Sea (Figs. 2c–e), in particularwhere a meandering jet associated with theSbyACC was observed (Gomis et al., 2002).Strong phytoplankton accumulation in this regioncaused this signal in the chemical field (Varelaet al., 2002). No similar feature was observedduring MACRO ‘96 one month later. In thissecond cruise, the highest oxygen (O2>380 mmol/kg), lowest nitrate (NO3o16 mmol/kg) and lowestNTIC (NTICo2170 mmol/kg) values were regis-tered in the Gerlache Strait related to phyto-plankton nutrient consumption/oxygen produc-tionFhigh chlorophyll concentrations weremeasured at these stations (see Varela et al.,2002; !Alvarez et al. 2002; Castro et al., 2002). At300 m depth, the distributions of the chemicalproperties clearly follow the thermohaline distri-butions and a high percentage of the chemicalvariability can be explained by variations in thethermohaline properties. No significant variability


Fig. 2. MACRO ‘95 leg. Surface horizontal distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate(mmol/kg), (e) NTIC (=TIC � 35/S; mmol/kg) and (f) silicate (mmol/kg) in the western basin of the Bransfield Strait and BellingshausenSea.


Fig. 3. MACRO ‘95 leg. Horizontal distributions at 300m of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate(mmol/kg), (e) NTIC (=TIC � 35/S; mmol/kg) and (f) silicate (mmol/kg) in the western basin of the Bransfield Strait and BellingshausenSea.


Fig. 4. MACRO ‘96 leg. Surface horizontal distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate(mmol/kg), (e) NTIC (mmol/kg) and (f) silicate (mmol/kg) in the western basin of the Bransfield Strait and Bellingshausen Sea.


Fig. 5. MACRO ‘96 leg. Horizontal distributions at 300m of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate(mmol/kg), (e) NTIC (mmol/kg) and (f) silicate (mmol/kg) in the western basin of the Bransfield Strait and Bellingshausen Sea.


was observed between the two surveys. Thedistributions of all the chemical properties(Figs. 3c–e, 5c–e), show strong gradients in theSbyACC and BF regions. The low oxygen(O2o180 mmol/kg), high nitrate (NO3>35 mmol/kg) levels north of the SbyACC reflect the south-ward extension of the LCDW. On the other hand,waters with Weddell Sea influence can be easilytraced by high oxygen (O2>250 mmol/kg) and lownitrate (NO3o32 mmol/kg) and NTIC (NTICo2270 mmol/kg) in the southeastern region of theBransfield Strait.

The distributions of silicate behave in a moreconservative way than the other chemical proper-ties (i.e. nitrate distributions) in the study area.Figs. 2f and 4f show a consistent picture of thesilicate distribution at sea surface which appar-ently is not affected by the December 1995phytoplankton bloom nor by any other source ofintraseasonal variability. The surface waters of theBransfield Strait have silicate concentrations ex-ceeding 65 mmol/kg whereas the correspondingvalues are o60 mmol/kg in the eastern Belling-shausen Sea. Moreover, the position of theSbyACC can be easily traced via the horizontalgradient of the sea surface silicate concentrations.This is not so clear, however, for the BF. Theposition of the SbyACC is also noticeable in thesilicate distribution at 300 m depth, although atthis level waters pertaining to the ACC region donot have significantly different silicate concentra-tions from Bransfield Strait waters.

4.2. Vertical distribution of water masses

Figs. 6 and 7 display the vertical distributions(up to 1200 m) of the thermohaline and chemicalproperties along the eastern Bellingshausen Sea–Bransfield Strait transect through the Boyd Straitcorresponding to the MACRO ‘95 and MACRO‘96 surveys. Despite the y and the S distributionsexhibit the same overall pattern in both legs, acloser view reveals two noteworthy differences(Figs. 6a and b, 7a and b). One is that sea surfacetemperature and salinity of the MACRO ‘95 legare lowerFby roughly 11C and 0.1 psu, respecti-velyFthan in the MACRO ‘96 leg (Figs. 2a and b,4a and b), which is a clear signal of hydrographic

intraseasonal variability in the study area (theMACRO ‘95 y and S values were influenced bymelting of the sea ice cover occurred shortly beforethe cruise whereas the MACRO ‘96 distributionsdisplay full summer conditions). On the otherhand, the intrusion of modified LCDW into theBransfield Strait was larger in January 1996 thanin December 1995, as the temperature distribuionsreveal (see Figs. 6a and 7a). Moreover, a carefulscrutiny of the details of the y distributions showsthat this feature is not an interpolation arti-factFthe area covered by the 1.01C isotherm doesnot include any of the data points of station 8 inthe MACRO ‘95 distribution shown in Fig. 6a,whereas it includes all points below 300 m in thecorresponding station (stn 215) of the MACRO‘96 distribution depicted in Fig. 7a. We think thatdifferential intrusion of CDW is related to under-sampled deep mesoscale variability of the LCDWinflowFGarc!ıa et al. (1994) reported time seriesobservations of temperature at 400 m acquiredsouth of Deception Island during Austral summer1992/93 which displayed transient temperatureincreases which could only be explained byvariability of the intrusion of CDW at mesoscale.On the other hand, there was no signature ofBDW on the transect in any of the two occupa-tions during the FRUELA study.

As in the horizontal distributions shown before,the vertical distributions of oxygen, nitrate andNTIC (Figs. 6c–e, 7c–e) closely mirror the tem-perature distributions. However, the temperatureminimum of AASW, at about 80 m depth, is nottraced in the chemical fields. At this depth, stronggradients are observed in the chemical properties.Below AASW, the poleward extension of CDW isclearly discerned by minimum oxygen values andmaximum concentrations of nitrate and NTIC.However, the chemical properties extrema are notlocated at the same depth. For the two cruises, wefound the nitrate and NTIC maxima at about250 m depth and the oxygen minimum a littledeeper (B350 m depth). In addition, the differen-tial intrusion of CDW into the Bransfield Strait isalso evident in these vertical distributions. DuringMACRO ‘96, lower oxygen (O2o200 mmol/kg),higher nitrate (NO3>33 mmol/kg) and higherNTIC (NTIC>2280 mmol/kg) values were reached


at stn 215 in comparison with concentrationsmeasured at stn 8 during the December 1995cruise. On the other hand, the strong verticalchemical gradients between stns 6–8 for MACRO‘95 and stns 217–215 for MACRO ‘96 mark theposition of the BF closely following the tempera-ture distributions (Figs. 6a and 7a). South of thefront, the domain of Weddell Sea water influenceis defined by high oxygen (O2>280 mmol/kg),low nitrate (NO3o32 mmol/kg) and low NTIC(NTICo2260 mmol/kg) levels.

As regards to silicate, it is seen again to be abetter tracer of local water masses than nitrate(Figs. 6f and 7f). This is indeed clear in the upper100 m, where the MACRO ‘95 and MACRO ‘96

vertical distributions of silicate are very similar. Atdepth, the maximum silicate concentration valueswere attained just south of the Boyd Strait duringboth cruises. This fact and the shape of the silicateconcentration isolines suggest that these maximawere related to the inflow of LCDW into theBransfield Strait. With respect to this, the increaseof the deep silicate concentrations in about10 mmol/kg between December 1995 and January1996 would be an expression of larger LCDWinflow.

Figs. 8 and 9 depict interpolated vertical dis-tributions of potential temperature, salinity, oxy-gen, nitrate, NTIC and silicate along the GerlacheStrait obtained from the MACRO ‘95 and

5 6 7 8 9 10 11 12

Station

1200

800

400

0D

epth

(m

)

MACRO'95

θ

5 6 7 8 9 10 11 12

Station

1200

800

400

0

Dep

th (m

)

MACRO'95

O2

MACRO'95NO3

1200

800

400

0

Dep

th (m

)

MACRO'95

S

MACRO'95

NTIC

MACRO'95

SiO2

Fig. 6. MACRO ‘95 leg. Vertical distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate (mmol/kg), (e)NTIC (mmol/kg) and (f) silicate (mmol/kg) along the Bellingshausen Sea–Bransfield Strait transect through the Boyd Strait. The DrakePassage is on the right and the Bransfield Strait is on the left.


MACRO ‘96 data sets. The y and S sections(Figs. 8a and b, 9a and b) show a consistentpattern whose overall features are repeated in bothlegs. The vertical distributions of temperature(Figs. 8a and 9a) are similar to that obtained byRoese and Speroni (1995) using XBTs in February1991. The water sequence west of the BismarckStraitFi.e., the western end of the GerlacheStraitFconsists of a surface layer of AASWincluding a temperature minimum and a deeplayer of CDW which hosts both the temperatureand the salinity maxima. In the eastern Belling-shausen Sea, UCDW reaches the shelf break and isobserved near 300 m (Klinck, pers. comm.). It maybe that LCDW has risen to 300 m northwest of the

Bransfield Strait as the ACC (and UCDW) shiftsoffshore. If this was the case during the FRUELAcruises, CDW flowing from the Bellingshausen Seainto the Gerlache Strait would be LCDW. But itmight also be that the CDW entering the GerlacheStrait is uplifted UCDW. In any case, the warmwater found in the Gerlache Strait at depth iscertainly some sort of CDW modified throughheat exchange with the surface and mixing withcoastal waters.

Water flows from the Bellingshausen Sea intothe Gerlache Strait through multiple paths like viaBismarck Strait and through Dallman Bay and theScholaert Channel (Klinck, pers. comm.). TheBellingshausen Sea inflow has to pass over sills

1200

800

400

0

Dep

th (m

)

MACRO'96

O2

217 216 215 214 213 212 211

Station

1200

800

400

0

Dep

th (m

)

217 216 215 214 213 212 211

Station

MACRO'96

S

MACRO'96

NO3

MACRO'96SiO2

1200

800

400

0

Dep

th (m

)

MACRO'96

NTIC

MACRO'96

θ

Fig. 7. MACRO ‘96 leg. Vertical distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate (mmol/kg), (e)NTIC (mmol/kg) and (f) silicate (mmol/kg) along the Bellingshausen Sea–Bransfield Strait transect through the Boyd Strait. The DrakePassage is on the right and the Bransfield Strait is on the left.


whose depths are less than 400 m and this placesan important constraint to CDW. Most of theCDW layer is ‘‘lost’’ and is replaced by TWW tothe extent that the deep temperature maximumwhich characterizes CDW is totally eroded at thecentre of the Gerlache Strait. At surface, a sharpertransition from AASW to TBW is observed. Asexpected, TBW in the Gerlache Strait is somewhatcooler and fresher than in the Bransfield Strait.Back to deep layers, an isolated temperaturemaximum (y > 0:21C) is observed near the easternend of the Gerlache Strait at depths of the order of200 m. Given the fact that the salinity correspond-ing to this maximum is lower than the salinity ofthe CDW core in the western part of the Strait, its

oxygen concentration is higher and its silicatelevels are lower, and also taking the local bathy-metric setting of the Gerlache Strait and surround-ings into account, we believe that this feature maycorrespond to a filament of modified CDWadvected from the western Bransfield basin.Another possible explanation is that thiswater body consists of Bellingshausen Sea CDWflown into the Gerlache Strait through theScholaert Channel or other passage furthereast. The differences in properties with respectto CDW found in the western part of theGerlache Strait would then result from mixingwith waters adjacent to the western flanks of theStrait.

34 37 39 40 4142 43 47

Station

1200

800

400

0

Dep

th (m

)

34 37 39 40 4142 43 47

Station

MACRO'95S

MACRO'95

NO31200

800

400

0

Dep

th (m

)

MACRO'95

O2

1200

800

400

0

Dep

th (m

)

MACRO'95

NTIC

MACRO'95SiO2

MACRO'95θ

Fig. 8. MACRO ‘95 leg. Vertical distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate (mmol/kg), (e)NTIC (mmol/kg) and (f) silicate (mmol/kg) along the Gerlache Strait. The Bellingshausen Sea is on the left and the Bransfield Strait ison the right.


The only remarkable differences between theMACRO ‘95 and MACRO ‘96 y and S distribu-tions along the Gerlache Strait are constrained tothe upper layers. As was the case in the westernbasin of the Bransfield Strait, the MACRO ‘96 seasurface temperatures were higher in the GerlacheStrait than the corresponding MACRO ‘95values (see Figs. 8a and 9a) due to seasonalwarming, but unlike in the Bransfield, the surfacesalinity was lower in January 1996 than inDecember 1995. This probably reflects the increaseof the freshwater supply related to meltingprocesses at local glaciers and passing icebergsas Austral summer progresses. In a limited-areabasin as the Gerlache Strait, such freshwater

inputs must have a significant effect on surfacewater properties.

The distributions of oxygen, nitrate and NTICalong the Gerlache for the two cruises are stronglydetermined by phytoplankton consumption in theupper meters of the water column. High oxygenconcentrations, low nitrate and low NTIC valueswere obtained on both sides of the Gerlache Straitdue to local high phytoplankton accumulation(Varela et al., 2002; !Alvarez et al., 2002; Castroet al., 2002). In fact, the lowest nitrate (NO3o12 mmol/kg) and NTIC (NTICo2195 mmol/kg)concentrations for the two surveys were registerednear the southwestern end of the Gerlache Straitduring MACRO’ 96. In the central sector of the

MACRO'96NO3

1200

800

400

0

Dep

th (m

)

MACRO'96NTIC

1200

800

400

0

Dep

th (m

)

MACRO'96O2

187 189 191 192193 195 186

Station

1200

800

400

0

Dep

th (m

)

187 189 191 192193 195 186

Station

MACRO'96S

MACRO'96SiO2

MACRO'96 θ

Fig. 9. MACRO ‘96 leg. Vertical distributions of (a) potential temperature (1C), (b) S (psu), (c) O2 (mmol/kg), (d) nitrate (mmol/kg), (e)NTIC (mmol/kg) and (f) silicate (mmol/kg) along the Gerlache Strait. The Bellingshausen Sea is on the left and the Bransfield Strait ison the right.


Gerlache Strait, a deep mixed layer probablyprevented phytoplankton accumulation (Castroet al., 2002) and consequently lower oxygen andhigher nitrate and NTIC levels were observed.

The vertical distributions of oxygen, nitrate andNTIC in the Gerlache Strait precisely mimic thetemperature distributions at layers deeper than100 m (Figs. 8c–e, 9c–e). The deep oxygen mini-mum (O2o220 mmol/kg), nitrate maximum(NO3>32 mmol/kg) and NTIC maximum(NTIC>2280 mmol/kg) are related to the abovecited inflows of LCDW. Again as in the BransfieldStrait and in the eastern Bellingshausen Sea,silicate seems to be a good tracer of LCDW (seeFigs. 8f and 9f). The silicate concentration rangewithin the LCDW body in the Gerlache Straitmatches the values displayed by LCDW in the restof the FRUELA study area.

5. Discussion

In Table 1, we show the average values of thethermohaline and chemical properties for thedifferent water masses sampled in the study area.In the eastern Bellingshausen Sea, the presence ofAASW is clearly noticeable through the tempera-ture minimum at about 80 m depth. The averagethermohaline properties of this minimum werewarmer and saltier during the MACRO’ 96 due tointraseasonal variability (Table 1). The averageconcentrations of oxygen, nitrate and NTIC werenot very different between the two cruises and nocorrelation with the thermohaline properties wasfound. However, about 52% and 68% of theobserved variability of silicate concentrationin the MACRO ‘95 and MACRO ‘96 datasets, respectively, can be explained by temper-ature and salinity. Whitehouse et al. (1995),using data from a transect sampled in theBelllingshausen Sea west of our study area duringthe UK STERNA project, showed that silicateconcentration in the core of AASW was deter-mined by the hydrographic variability in contrastwith nitrate, which presented an homogeneousdistribution.

North of the SbyACC, we can distinguish thecores of UCDW and LCDW in the chemical field.

The UCDW maximum temperature is associatedto an oxygen minimum and both nitrate andNTIC maxima (Table 1). UCDW’s averagetemperature in the area of the Bellingshausen Seasampled by R/V Hesp !erides was higher duringMACRO ‘95 than in MACRO ‘96 whereasLCDW presented the highest average tempera-tures in MACRO ‘96. The temperature changes inthe UCDW and LCDW cores were correlated tothe changing levels of nitrate. On the other hand,strong gradients in sea surface silicate distributionswere observed across the SbyACC in the twosurveys. Veth et al. (1997) also described a strongsilicate surface gradient associated with thisstructure at 61W.

In the southwestern part of the Gerlache Strait,we found similar thermohaline and chemicalproperties of AASW for the two cruises (Table1). AASW in the Gerlache Strait was less salinethan at the eastern Bellingshausen Sea. Alsoduring the two cruises, the layer of LCDWpresented lower temperature and salinity valueswith respect to the Bellingshausen Sea. Dilution ofthe LCDW signal was also observed in oxygen,probably due to mixing with highly oxygenatedTWW. However, nitrate and NTIC levels were notsignificantly different from the concentrationsfound in the Bellingshausen area and silicate levelswere even higher, which might be a product ofsediments supplied by local glaciers. In the north-east part of the Gerlache Strait, the water columnwas occupied by TBW over TWW, though a littleintrusion of LCDW was registered. RegardingTBW, we found similar thermohaline propertiesfor the two FRUELA cruises. Oxygen, nitrate andNTIC levels were influenced by phytoplanktonconsumption, being stronger during the MACRO‘96 survey. However, silicate was not affected byphytoplankton activity as the region was domi-nated by Cryptomonas populations (Varela et al.,2002). In fact, silicate levels were very similar forthe two cruises and higher than silicate concentra-tions for TBW in the western Bransfield region.TWW was found between 300–700 m depth in theGerlache Strait, while in the Bransfield Strait itencompasses the whole water column at someareasFin particular, on the Antarctic Peninsulacontinental shelf. Consequently, we cannot


directly compare the average property values fromthe two regions.

In the Bransfield Strait, the average temperaturevalues for the upper water masses, i.e. TBW andTWW, reflect the seasonal heating (Table 1). Forthe two water masses, the low oxygen values forMACRO ‘96, associated with no significantchanges in nitrate and NTIC concentrationsbetween the two cruises, suggest the influence ofremineralization processes during MACRO ‘96.During the two surveys, TBW was characterizedby lower silicate values than TWW. The averagetemperature for LCDW was higher duringMACRO ‘96, which we intrepret as the expression

of a larger intrusion of this water mass in January1996. Also the higher nutrient levels and loweroxygen values support this idea.

To our knowledge, the paper by Tokarczyk(1987) is the only paper published in the openliterature in which an objective classification ofwater masses has been proposed for the BransfieldStrait on the basis of a statistical multidimensionalanalysis method involving both physical andchemical water properties. This author suggesteda classification of Bransfield Strait waters in sixgroups to which the four local water massesdiscussed in Section 3 roughly correspondFTBWapproximately corresponds to Tokarczyk’s water

Table 1

Average and standard deviation of the thermohaline and chemical properties for the different waters masses found in the eastern

Bellingshausen Sea (EBELLINGS), the Gerlache Strait (GERLACHE) and the western Bransfield Strait (BRANSFIELD) during the

MACRO ‘95 and MACRO ‘96 surveys. We also have introduced the maximum and minimum values in the western Bransfield region

during the MACRO’96 for comparison with Tokarczyk’s water masses classification (see text and Table 2)a

Area WT y S O2 NO3 SiO2 NTIC

MACRO ‘95

EBELLINGS AASW �0.94l70.24 33.97l70.05 309l716 29l71 66l711 2253l79UCDW 2.15l70.06 34.55l70.06 180l74 35.1l70.4 77l73 2279l74LCDW 1.6l70.2 34.69l70.02 188l77 32.9l70.3 97l76 2272l73

BRANSFIELD TBW �0.38l70.25 33.95l70.05 343l711 27l71 72l72 2239l710TWW �0.80l70.15 34.18l70.08 330l716 27l71 81l72 2249l78LCDW 0.57l70.41 34.51l70.11 228l717 32.4l70.3 92l73 2273l71

GERLACHE AASW �0.68l70.47 33.92l70.07 278l72 29l71 88.2l70.2 2262l72CDW 0.96l70.30 34.57l70.07 195l713 32.9l70.1 102l73 2276l74TBW 0.44l70.53 33.88l70.13 321l724 26l72 87l71 2247l714TWW �0.56l70.19 34.54l70.02 251l74 31.84l70.06 87l71 2266l72

MACRO ‘96

EBELLINGS AASW �0.69l70.40 34.02l70.08 292l721 29l71 68l714 2258l712UCDW 2.07l70.04 34.635l70.003 175l71 34.2l70.1 87l75 2285l71LCDW 1.82l70.17 34.66l70.02 181l74 33.2l70.5 90l76 2280l73

BRANSFIELD TBW 0.84l70.51 33.93l70.11 322l711 25l72 75l76 2240l714TBW max 1.65 34.16 335 29.0 84 2268

TBW min 0.02 33.72 291 21.3 68 2214

TWW �0.3l70.5 34.28l70.13 318l722 29l72 83l74 2252l712TWW max 0.66 34.53 351 32.7 91 2273

TWW min �0.92 34.08 271 25.2 76 2222LCDW 0.63l70.45 34.48l70.16 226l723 33.1l70.4 94l710 2279l72LCDW max 1.15 34.66 261 33.7 107.1 2283

LCDW min 0.02 34.27 203 32.4 81 2275

GERLACHE AASW �0.71l70.21 33.79l70.01 285l712 29.4l70.6 91l74 2272l73CDW 1.15l70.20 34.61l70.06 195l78 33.4l70.2 107l72 2281l74TBW 0.47l70.19 33.71l70.25 328l747 20l76 86l76 2217l739TWW �0.35l70.07 34.54l70.01 252l73 31.8l70.3 94l72 2276l72

ay in 1C, salinity (S) in psu and chemical property concentrations in mmol/kg. WT: water type.


types 1 and 1a; TWW involves Tokarczyk’s types4a and 4b; modified LCDW is Tokarczyk’s watertype 3; BDW was not individualized by Tokarczykbut included in type 4b.

Table 1 shows the range of values of waterproperties measured in the western BransfieldStrait and Gerlache Strait during the MACRO‘96 cruise and Table 2 displays equivalent valuesobtained by Tokarczyk for his water classes usingthe FIBEX 1981 data set (FIBEX 1981 was one ofthe cruises auspiced by the international BIO-MASS programme and roughly covered the samestudy area as the FRUELA cruises in February–March 1981. Here we use the MACRO ‘96 data setfor comparison because it was obtained underseasonal conditions which are closest to those ofFIBEX 1981). The comparison of both tablessuggests that Tokarczyk’s (1987) water types areexcessively broad in hydrographic terms. They arebest discriminated in y� S space but the rangesfor oxygen, nitrate and silicate are too wide andoverlap, so assigment of water samples to aparticular water type on the basis of a singlechemical property does not seem feasible. On thecontrary, Bransfield Strait local water massesdefined according to the conventions exposed inSection 3 do have the same distinct character iny� S space but have much narrower ranges ofchemical properties and little intersection amongthem.

Table 1 suggests that silicate is a very gooddiscriminator of water masses in the BransfieldStrait. Moreover, the distributions shown inFigs. 8f and 9f evidence that silicate concentrationhas very little variability (at least, on a sub-

seasonal scale), so it is a robust water massindicator in the Gerlache Strait area. No intervalcan be assessed for BDW as it was not found inour study area. Another choice would be labellingwater masses by using oxygen and nitrate con-centrations jointly, but it has the obvious draw-back that values in the euphotic zone are modifiedby non-physical causes.

Acknowledgements

We acknowledge the help and the support of thecrew of R/V Hesperides and of all colleagues whomade possible data acquisition during the FRUE-LA December 1995 and January 1996 cruises.We are specially grateful to Julia Figa, ManuelGonz!alez, Joan Puigdef"abregas, Maribel Lloret,Pilar Rojas, Mario Manr!ıquez, Marcel. l!ı Farr!an,Jorge Guill!en and Pere Masqu!e (hope to meet onboard again!). Our special thanks to T. Rell!an andM.V. Gonz!alez for the oxygen and CO2 analysesduring the two cruises and to M.J. Paz !o for thenutrient analyses during the MACRO ‘96 cruise.This study has been funded by the SpanishComisi !on Interministerial de Ciencia y Tecnolo-gia, contract no. ANT94-1010.

References

!Alvarez, M., R!ıos, A.F., Ros !on, G., 2002. Spatio-temporal

variability of air-sea fluxes of carbon dioxide and oxygen in

the Bransfield and Gerlache Straits during Austral summer

1995–96. Deep-Sea Research II 49, 643–662.

Table 2

Range of values of water properties measured during the FIBEX cruise, February–March 1981 (after Tokarczyk, 1987)a,b

Type y S O2 NO3 SiO2 NTIC

1a 0.5–3.3 33.1–33.9 322–357 9–25 2–26 n.d.

1 0.1–0.8 32.9–33.9 327–379 2–16 20–60 n.d.

3 0.9–2.1 34.3–34.8 174–244 22–27 24–102 n.d.

4a �1.3–1.5 33.6–34.4 261–353 10–24 11–73 n.d.4b �1.5–1.0 33.8–34.6 218–335 13–26 39–100 n.d.

aNote: Correspondence between water types from Tables 1 and 2: TBWB1, 1a; TWWB4a, 4b; LCDW B3; BDW included in 4b.See text for more details.

by in 1C, salinity (S) in psu, O2 in mmol/kg, NO3 in mmol/kg, SiO2 in mmol/kg and NTIC in mmol/kg.


!Alvarez-Salgado, X.A., Fraga, F., P!erez, F.F., 1992. Determi-

nation of nutrient salts by automatic methods both in

seawater and brackish waters: the phosphate blank. Marine

Chemistry 39, 311–319.

Castro, C.G., R!ıos, A.F., Doval, M.D., P!erez, F.F., 2002.

Nutrient utilisation and chlorophyll distribution in the

Atlantic sector of the Southern Ocean during Austral

summer 1995–96. Deep-Sea Research II 49, 623–641.

Clowes, A.J., 1934. Hydrology of the Bransfield Strait.

Discovery Reports 9, 1–64.

Dickson, A.G., 1981. An exact definition of total alkalinity and

procedure for the estimation of alkalinity and total

inorganic carbon from titration data. Deep-Sea Research I

28 (6), 609–623.

Garc!ıa, M.A., 1996. BIO Hesperides covered WOCE SR1b in

February 1995. International WOCE Newsletter 23, 33–35.

Garc!ıa, M.A., L !opez, O., Sospedra, J., Espino, M., Gr"acia, V.,

Morrison, G., Rojas, P., Figa, J., Puigdef"abregas, J., S.-

Arcilla, A., 1994. Mesoscale variability in the Bransfield

Strait region (Antarctica) during Austral summer. Annales

Geophysicae 12 (9), 856–867.

Gomis, D., Garc!ıa, M.A., L !opez, O., Pascual, A., 2002. Quasi-

geostrophic 3D circulation and mass transport in the

western sector of the South Sheltland Islands. Deep-Sea

Research II 49, 603–621.

Gordon, A.L., Nowlin, W.D., 1978. The basin waters of the

Bransfield Strait. Journal of Physical Oceanography 8,

258–264.

Grelowski, A., Majewicz, A., Pastuszak, M., 1986. Mesoscale

hydrodynamic processes in the region of the Bransfield

Strait and the southern part of Drake Passage during

BIOMASS-SIBEX 1983/84. Polish Polar Research 7,

353–369.

Hansen, H.P., Grasshoff, K., 1983. Automated chemical

analysis. In: K. Grasshoff, M. Ehrhardt, K. Kremlig

(Eds.), Methods of Seawater Analysis. Verlag Chemie,

Weinheim, 419pp.

Hofmann, E.E., Klinck, J.M., Lascara, C.M., Smith, D.A.,

1996. Water mass distribution and circulation west of the

Antarctic Peninsula and including Bransfield Strait. In:

R.M. Ross et al. (Ed.), Foundations for Ecological

Research west of the Antarctic Peninsula, Vol. 70, Antarctic

Research Series, pp. 61–80.

Huntley, M., Karl, D.M., Niiler, P., Holm-Hansen, O., 1991.

Research on Antarctic Coastal Ecosystem rates (RACER):

and interdisciplinary field experiment. Deep-Sea Research II

38, 911–941.

Lee, K., Millero, F.J., Campbell, D.M., 1996. The reliability of

the thermodynamic constants for the dissociation of

carbonic acid in seawater. Marine Chemistry 55, 233–245.

L !opez, O., Garc!ıa, M.A., S.-Arcilla, A., 1994. Tidal and

residual currents in the Bransfield Strait, Antarctica.

Annales Geophysicae 12 (9), 887–902.

L !opez, O., Garcia, M.A., Gomis, D., Rojas, P., Sospedra,

J., S.-Arcilla, A., 1999. Hydrographic and hydro-

dynamic characteristics of the eastern basin of the

Bransfield Strait (Antarctica). Deep-Sea Research I 46,

1755–1778.

Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicz,

R.M., 1973. Measurements of the apparent dissociation

constant of carbonic acid in seawater at atmospheric

pressure. Limnology and Oceanograhy 18, 897–907.

Millero, F.J., 1995. Thermodynamics of the carbon dioxide

system in the oceans. Geochimica et Cosmochimica Acta 59,

661–677.

Mouri *no, C., Fraga, F., 1985. Determinacion de nitratos en

agua de mar. Investigacion Pesquera 49, 81–96.

Niiler, P., Amos, A., Hu, J.-H., 1991. Water masses and 200m

relative geostrophic circulation in the western Bransfield

Strait region. Deep-Sea Research II 38, 943–959.

Orsi, A.H., Whitworth III, T., Nowlin Jr., W.D., 1995. On the

meridional extent and fronts of the Antarctic Circumpolar

Current. Deep-Sea Research I 42, 641–673.

P!erez, F.F., Fraga, F., 1987a. A precise and rapid analytical

procedure for alkaliniy determination. Marine Chemistry

21, 169–182.

P!erez, F.F., Fraga, F., 1987b. The pH measurements in

seawater on NBS scale. Marine Chemistry 21, 315–327.

Roese, M., Speroni, J., 1995. Din!amica de las aguas en Bahia

Para!ıso y su vinculaci !on con el Estrecho de Gerlache,

Ant!artida. Instituto Antartico Argentino, contribucion No.

416, 1–29.

Rojas, P., L !opez, O., Garc!ıa, M.A., Ar!ıstegui, J., Torres, S.,

Hern!andez-Le !on, S., Bastarrechea, G., Amengual, B.,

Esc!anez, J., Morales-Nin, B., 1996. Hidrograf!ıa del Estrecho

de Bransfield durante el verano Austral 91/92. Actas del V

Simposio Espa *nol de Estudios Ant!articos, 403–411.

Tokarczyk, R., 1987. Classification of water masses in the

Bransfield Strait and southern part of the Drake passage

using a method of statistical multidimensional analysis.

Polish Polar Research 41, 629–641.

Varela, M., Fernandez, E., Serret, P., 2002. Size-fractioned

phytoplankton biomass and primary production in the

Gerlache and south Bransfield Straits (Antarctic Peninsula)

in Austral summer 1995–1996. Deep-Sea Research II 49,

749–768.

Veth, C., Peeken, I., Scharek, R., 1997. Physical anatomy of

fronts and surface waters in the ACC near the 6W meridian

during Austral spring 1992. Deep-Sea Research II 44,

23–50.

Whitehouse, M.J., Priddle, J., Woodward, E.M.S., 1995.

Spatial variability of inorganic nutrients in the marginal

ice zone of the Bellingshausen Sea during the Austral spring.

Deep-Sea Research II 42, 1047–1058.


water masses and distribution ofphysico-chemical properties in...

Documents