Southern Antarctic Circumpolar Current Front to the northeast of

South Georgia: Horizontal advection of krill and its role in

the ecosystem

E. J. Murphy, J. L. Watkins, M. P. Meredith, P. Ward, P. N. Trathan, and S. E. ThorpeBritish Antarctic Survey, Cambridge, UK

Received 3 July 2002; revised 23 June 2003; accepted 17 September 2003; published 31 January 2004.

[1] During December 2000 and January 2001 we conducted a high-resolutionhydrographic and bioacoustic transect (RRS James Clark Ross cruise 57) that extendedacross the South Georgia shelf from close to Cumberland Bay, across the shelf break andslope and into the deep waters of the Georgia Basin beyond. We observed a high biomassof zooplankton between 53.8� and 53.4�S associated with the inshore, northwestwardflow of the Southern Antarctic Circumpolar Current Front (SACCF) that occurred inaround 2500 m of water close to the base of the slope. There was very little zooplanktonbiomass present in the more offshore, eastward flowing waters where a secondmanifestation of the SACCF was also present on the section. The region of enhancedzooplankton biomass was over 50 km in horizontal extent with the highest densities(>10 g m�3) in the area of strongest flow (>35 cm s�1). The majority of the zooplanktonpresent on the section was Antarctic krill and most of it occurred in the upper 100 m. Therate of physically mediated transport of Antarctic krill across the off-shelf sections(�10 km) of the transect showed marked variation, with highest rates (>106 g s�1)associated with the northwestward flow of the SACCF. Farther offshore, where the krillbiomass and flow rates were much reduced, the flux of krill was very low. The integratedhorizontal flux of krill across the offshore sections was large (192 � 103 t d�1) and to thenorthwest. A second occupation of the transect showed that the krill flux is highlyvariable, and we discuss the various physical and biological factors that will generate suchvariability. We show that horizontal flux of krill in ocean currents is a major factor indetermining the abundance of krill around South Georgia. INDEX TERMS: 4207

Oceanography: General: Arctic and Antarctic oceanography; 4815 Oceanography: Biological and Chemical:

Ecosystems, structure and dynamics; 4855 Oceanography: Biological and Chemical: Plankton; 4528

Oceanography: Physical: Fronts and jets; 4899 Oceanography: Biological and Chemical: General or

miscellaneous; KEYWORDS: Southern Ocean, krill, circulation, ecosystem, South Georgia, food web

Citation: Murphy, E. J., J. L. Watkins, M. P. Meredith, P. Ward, P. N. Trathan, and S. E. Thorpe (2004), Southern Antarctic

Circumpolar Current Front to the northeast of South Georgia: Horizontal advection of krill and its role in the ecosystem, J. Geophys.

Res., 109, C01029, doi:10.1029/2002JC001522.

1. Introduction

[2] To be able to predict the dynamics of marineecosystems we require an understanding of the physicaland biological interactions that maintain the systems. Thatocean currents transport biological material was recog-nized in early studies of species and large-scale ecosys-tems [Darwin, 1884; Hardy, 1967], and transport ofplankton in ocean current systems is known to beimportant in the development and maintenance of partic-ular plankton populations [e.g., Bryant et al., 1998;Gurney et al., 2001]. Yet, there has been little quantifi-cation of the importance of horizontal transport in themaintenance of marine food webs, although such external

input of biological material is known to be crucial to themaintenance of some regional ecosystems [Agusti et al.,2001; Ansorge et al., 1999; Pakhomov and Froneman,1999; Perissinotto et al., 1992; Polis and Hurd, 1996;Sanchez-Pinero and Polis, 2000]. A likely consequenceof climate change will be shifts in the pattern of oceancirculation [Bigg, 1996; Houghton et al., 2001] disruptingthese pathways of biological transport, potentially result-ing in the collapse of regional food webs.[3] Some of the world’s largest breeding colonies of

seabirds and seals occur on the island of South Georgia(Figure 1) in the northeastern Scotia Sea [Croxall, 1987;Everson, 1977; Laws, 1984]. During the summer, pen-guins, albatrosses and fur seals rely mainly on a singlespecies of zooplankton, Antarctic krill (Euphausiasuperba Dana), as their major source of food [Boyd,2002; Everson, 1977]. Antarctic krill was also the major

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 109, C01029, doi:10.1029/2002JC001522, 2004

Copyright 2004 by the American Geophysical Union.0148-0227/04/2002JC001522$09.00

C01029 1 of 10

prey of many of the whale species that were exploited soheavily during the first half of the last century [Everson,1977] and is the target of a commercial fishery thatoperates around South Georgia during the winter months[Murphy et al., 1997]. The krill around South Georgia donot form an isolated population [Marr, 1962; Murphy,1995; Murphy et al., 1998; Watkins et al., 1999] as thereis thought to be an input of individuals from the southernScotia Sea (Figure 1), either from the west around theAntarctic Peninsula or from farther east across the SouthScotia Ridge. Indeed, studies of krill populations haveshown that the large interannual fluctuations in theabundance of krill at South Georgia are linked to basin-wide changes in abundance and krill population dynamics[Brierley et al., 1999a; Hofmann et al., 1998; Murphyand Reid, 2001; Murphy et al., 1998; Watkins et al.,1999], which in turn have been linked to large-scaleenvironmental processes [Loeb et al., 1997].[4] Recent work has shown the importance of frontal

regions in determining the distribution of Antarctic krillin the Southern Ocean [Loeb et al., 1997; Nicol et al.,2000; Tynan, 1998]. Coupled biological-physical modelstudies suggest that the large-scale ocean circulation, and

in particular the flow in association with the SouthernAntarctic Circumpolar Current Front (SACCF) [Orsi etal., 1995], may play a role in the transport of krill intothe South Georgia region [Hofmann et al., 1998; Murphyet al., 1998]. A multidisciplinary biological-physicaloceanographic transect was undertaken on the northernside of South Georgia to quantify krill transport rates inthe ecosystem and to determine specifically whether theSACCF is important in transporting krill (Figure 1). Inthe associated paper [Meredith et al., 2003] we describedthe detailed physical oceanography of the section. Herewe report on the detailed field-based estimates of verti-cally resolved upper water column velocities and abun-dance of Antarctic krill that we combine to calculate thehorizontal flux of krill in association with the currentsystems. We consider the magnitude of the flux inrelation to local predator demand for krill in the regionalfood web and the large-scale transport of the krill.

2. Methods

[5] During December 2000 and January 2001, a 160 kmtransect on the northern side of South Georgia between

Figure 1. The geographical and oceanographic setting for the study. South Georgia is on the northernedge of the Scotia Sea in the Southern Ocean. The main flow of the Antarctic Circumpolar Current(ACC) is from west to east, and there are a number of circumpolar fronts (solid black lines) that reflectregional water mass differences [modified from Orsi et al., 1995; Trathan et al., 1998b]. The approximateposition of the study transect is shown as a dashed line. (The 1000 and 3000 m bathymetric contours areplotted, with regions shallower than 3000 m shaded. SAF, Subantarctic Front; PF, Polar Front; SACCF,Southern ACC Front; SB, Southern Boundary of ACC; and Georgia B., Georgia Basin).

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

2 of 10

C01029

53.29�S, 34.26�W and 54.10�S, 36.27�W was undertakenby the RRS James Clark Ross [Meredith et al., 2003;Ward et al., 2002] (Figure 1). A continuous acousticsurvey leg was performed during 26 December 2000using towed and underway instrumentation [Meredith etal., 2003], followed by a return along the transect duringwhich 17 hydrographic stations were sampled. The con-tinuous transect was completed during daylight hours sothat the acoustic survey encompassed periods when mostof the krill would be expected to be deeper in the watercolumn. The second return transect was then undertakenover a period of about 2.5 days so included day andnight sampling.[6] Full depth hydrographic data were collected using

a SeaBird 911 plus conductivity-temperature-depth(CTD) instrument. CTD salinities were calibrated usingdiscrete water sample measurements, and CTD down-casts were averaged at 2 dbar intervals for analysis.Station separation was around 10 km, close to theRossby radius of deformation in the region [Houry etal., 1987]. Measurements of water velocity in the upperpart of the water column were made using a 153.6 kHzvessel-mounted RD Instruments Acoustic Doppler Cur-rent Profiler (ADCP). Underway measurements of acous-tic biomass between 10 and 500 m were made with aSimrad EK500 scientific echosounder operating throughhull-mounted transducers with frequencies of 38, 120,and 200 kHz. A pulse repetition rate of 2 s wascombined with a survey speed of 10 knots. Data werelogged using SonarData EchoLogEK with visualizationand post-data collection processing carried out usingSonarData Echoview software. The echosounder wascalibrated at Stromness Bay, South Georgia, on24 December 2000 using the standard target technique[Foote et al., 1987].[7] Acoustic biomass due to Antarctic krill was sepa-

rated from biomass attributable to other scattering targetsby calculating the difference between acoustic meanvolume backscattering strength (Sv) measured at 120and 38 kHz (d120�38) [Madureira et al., 1993]. Inspec-tion of d120�38 resulting from a series of targeted nethauls conducted during the cruise revealed that 2 <d120�38 < 14 was a reasonable range to include themajority of krill targets found in the study area (Woodd-Walker, personal communication, 2001). Previous studiesusing multifrequency acoustic techniques to delineatekrill have used a variety of ranges for d120�38, from2–12 dB [e.g., Brierley et al., 1999b] to 2–16 dB [SC–CAMLR, 2000]. In an attempt to estimate magnitude oferror due to possible misidentification of acoustic targetsin this paper, we also estimate biomass under twoadditional conditions: where all acoustic targets areassumed to be krill (which produces a maximum bio-mass but which is likely to overestimate the biomass ofkrill), and where only targets with d120�38 = 2 to 12 dBare assumed to be krill.[8] Biomass of krill (g m�3) was derived from acoustic

mean volume backscattering strength (Sv) in decibels (dB)using the procedures described by SC-CAMLR [1991,2000]. The calculations use the length-weighted meantarget strength (TS120) and the mean krill mass ( �W )calculated from the length-frequency distribution sampled

from the local krill population. Biomass (g m�3) is givenby

Biomass ¼ �W 10Sv�TS120ð Þ

10 ; ð1Þ

where TS120 (dB) is related to length (L) in millimeters bythe relationship:

TS120 ¼ �127:45þ 34:85 Log Lð Þ; ð2Þ

with length converted to mass (g) using the allometricrelationship:

W ¼ 2:236� 106L3:314: ð3Þ

The krill length (L) is measured from the front of the eye tothe tip of the telson, rounded down to the nearest mm.[9] To derive a representative estimate of krill length

distribution that could be used to scale the biomass estimate[Watkins et al., 1986], we have included all krill caughtusing 8 m2 Rectangular Midwater Trawl net hauls in thearea between the eastern end of South Georgia and thepresent transect line that were sampled over a 2 day periodimmediately following the present survey [Ward et al.,2002]. The mean size of krill from this area was 32.3 mm(total length). As is usual in acoustic surveys for krill [SC-CAMLR, 2000] no explicit account of krill orientation wastaken, however, averaging over large depth and distancebins should minimize stochastic errors and we assume thattilt angle distributions in field and TS experiments aresimilar [SC-CAMLR, 1991].[10] The derivation of the total velocity field was given

in the paper describing the detailed oceanography[Meredith et al., 2003]. In summary, baroclinic velocitieswere calculated from the CTD data relative to the deepestcommon depth between station pairs. For each stationpair, ADCP velocity vectors were averaged betweenstations and resolved perpendicular to the ship track.Below the ageostrophic layer, the offset between thebaroclinic (CTD) and ADCP velocity profiles is takenas being the barotropic (depth-independent) component ofthe velocity. The magnitude of the ageostrophic compo-nent was derived by considering the difference betweenthe geostrophic and ADCP velocity profiles in the uppersection of the water column. The total velocity field isderived as the sum of the geostrophic and ageostrophiccomponents. Tidal flows were subtracted using the Ore-gon State University (OSU) tidal model [Egbert et al.,1994].[11] Vertically resolved water transport in sverdrup (Sv;

1 Sv = 106 m3 s�1) was calculated between each stationpair (�10 km section) in 2 m depth bins over the upper250 m of the water column. The acoustic biomass datawere horizontally and vertically integrated to estimatebiomass (g m�3) in each 2 m by �10 km bin. The krillflux (g s�1) due to physical advection was then derivedby horizontally and vertically integrating the product ofthe water volume flux and krill density in each 2 m by�10 km bin. Calculations were undertaken using datafrom the section encompassed by the 12 offshore stations.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

3 of 10

C01029

[12] For the food web calculations regional krill stockestimates are based on a 35 g m�2 mean krill densityestimate for the northwest shelf region from January 2001[Brierley et al., 2002]. The biomass was calculated overan arc of 270� centred on the western tip of SouthGeorgia, the location of the most dense predator colonies.Biomass values were calculated for an area with a 75 kmradius with values for a 50 and 100 km radius to give arange. Mass increases due to growth were calculated forvalues of 2 and 5% increases in biomass per day[Atkinson et al., 2001]. A range for the regional increasesin biomass is calculated by applying these figures for theproportional increases in biomass per day to the range ofthe biomass estimates. The estimate (and range) of theconsumption rate of krill by the major krill predators atSouth Georgia during summer was based on seasonalenergy-based calculations for fur seal and macaroni pen-guin populations [Boyd, 2002].

3. Results and Discussion

3.1. Krill Biomass Associated With the SACCF

[13] Detailed analyses of the physical oceanography aregiven in an associated paper [Meredith et al., 2003]. Herewe present a brief description of the key points requiredto interpret the biological data and calculate the fluxes.The SACCF was found close to the base of the slope onthe north side of South Georgia (Figure 1). The locationof the front was marked by strong horizontal gradients intemperature, salinity and density [Meredith et al., 2003],and was close to the position of the 1.8�C isotherm at500 m depth [Orsi et al., 1995]. At the front a strongnorthwestward flow rate reached a maximum of approx.

40 cm s�1 with enhanced flow rates in a narrow bandapproximately 40 km wide. Further offshore there was aweaker southeastward flow (maximum 15 to 20 cm s�1)associated with an isolated ring that had spun off theSACCF retroflexion [Meredith et al., 2003].[14] Associated with the strong northwestward flowing

jet of the SACCF was an enhanced biomass of zooplank-ton (Figure 2). The region of enhanced biomass extendedover about a 40–50 km region, from inshore of thecenter of the front, around 53.8�S, out across the frontto about 53.4�S. Detailed analyses of the upper watercolumn showed that flow rates >25 cm s�1 to thenorthwest were associated with the SACCF, with flowsof �20 cm s�1 occurring over a 30–40 km region acrossthe front (Figure 3a). In the more offshore region wherethere was a generally weak eastward flow there was somevertical structure, with water deeper than 40 m flowingeastward at 15 cm s�1 while there was a very weaknorthwest flow (<5 cm s�1) in the upper 40 m. On thebasis of the net sampling and acoustic analyses themajority of the zooplankton echoes were attributed tokrill [Ward et al., 2002] so the echo-based identificationof krill accounts for the majority of the biomass in theregion. The highest densities of krill occurred in anapproximately 10 km section centered on the front atabout 53.8� to 53.7�S (Figure 3b). The majority of thekrill occurred in the upper 100 m of the water column.The krill were present in the highest densities in the areawith the strongest flow rates to the northwest. In the mostoffshore 40–50 km section there were few krill presentanywhere. What krill were present in the most offshoreregion occurred in the upper 40 m of the water columnassociated with the very weak northwest flow. The krill

Figure 2. The distribution of acoustically detected zooplankton (g m�3) in the upper 200 m of thetransect shown in Figure 1.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

4 of 10

C01029

were almost absent in the depth range between about 40and 250 m in the most offshore regions where the waterwas flowing to the southeast.[15] The majority of the krill biomass on the transect was

therefore associated with the strong northwestward flowingjet of the SACCF. We consider that these krill associatedwith the SACCF will subsequently have been physicallyadvected to the northwest of the transect section. Krill arenot just passive tracers and horizontal and vertical move-ments will modify their position in the water column [Nicol,

2000]. In the present situation however it is difficult toconceive of any way in which the krill would not betransported with the current flow observed. Horizontalswimming speeds of 10 to 20 cm s�1 have been suggestedfor krill [Miller and Hampton, 1989], but to have a markedeffect on the large-scale transport, krill would have tomaintain such speeds in a particular direction for a longperiod; this is unlikely. Krill are known to be able todiurnally vertically migrate [Godlewska, 1993; Tarling etal., 1998], and although this is far from a universal

Figure 3. Data combined for flux calculations: (a) vertically resolved data on the current flow (negativevalues represent northwestward flow) and (b) the abundance and distribution of krill in the water column.The bioacoustic data have been integrated to the same resolution as the current flow data: 2 m vertical andapproximately 10 km horizontal. The figures show the data for the upper 150 m and for the 12 mostoffshore stations.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

5 of 10

C01029

observation in Antarctic krill, such behavior can in theoryaffect their horizontal distribution in regions of verticalcurrent shear. Hardy [1967] discussed how vertical migra-tion of krill around South Georgia could modify theirdistribution. However, in the present study there was nomarked vertical shear structure that would be required toallow the horizontal distribution to be modified greatly.With no strong vertical structure in the current field for krillto exploit by migrating vertically, horizontal physicaladvection will have dominated the movement of krill.

3.2. Krill Flux in the South Georgia Ecosystem

[16] Vertically resolved estimates of water transport andbiomass were combined (Figure 3) to calculate the physicaladvective transport of krill in the region. High krill biomassassociated with the strong northwestward flow resulted in ahigh krill flux in the southern arm of the SACCF (Figure 4).Low krill densities combined with relatively weak south-eastward flow rates resulted in little flow of krill to the east.These field observations of high krill flux associated withthe SACCF confirm model-based suggestions that this frontis important in the Scotia Sea ecosystem [Hofmann et al.,1998]. The instantaneous gross northwestward flux was �211 � 103 tonnes per day, with a net northwestward flux ofkrill of � 192 � 103 tonnes per day integrated across thewhole section. The classification of acoustic targets as krillis based on the backscatter signal at 2 frequencies anddepending on the classification method used, the estimate ofthe net flux varied in the range of �154 to 200 � 103 tonnesper day.[17] The above study cannot reveal the level of varia-

tion in the current flow or in the biomass of krill beingtransported. We have effectively sampled a single point in

a continuous flux series and variability of krill flux intothe region must be expected so our measurements of krillflux are unlikely to be representative of a longer-termaverage. This was emphasized by calculating a secondflux estimate based on the biomass of krill surveyedbetween stations over the 2.5 day period of the CTDtransect [Ward et al., 2002]; this produced a much lowerestimate of krill flux (35.3 � 103 t d�1). Some of thesections on this repeat survey were sampled during thenight when krill migrate into the surface layers outside ofthe range of detection of the acoustic system, so theestimate of krill biomass would be expected to be lower.However, for those sections undertaken during daylighthours in the vicinity of the SACCF there was a reducedbiomass of krill compared to the biomass on the firstsurvey. Such variability may be generated in a variety ofways: physically through fluctuations in ocean currents orfrontal positions [Thorpe et al., 2002] affecting theamount and the pathways of krill transport, or biologi-cally through variations in the input of krill into thecurrent flow as a result of population or behavioral effectsfurther upstream [Murphy et al., 1998]. Whatever thebasis for the variability, the study has shown that advec-tion of krill by ocean currents is an important process,and needs to be considered in studies of regional eco-systems in the Southern Ocean.[18] We have observed an enhanced abundance of krill

in an offshore region toward the eastern end of the islandwhile much of the predator demand for prey occurs at thewestern end of the island both offshore [Boyd et al.,2002] and also on the shelf [Trathan et al., 1998a].Simple views of the circulation in the region indicatethat the westward route along the north coast of South

Figure 4. The ocean current data and acoustically derived biomass data were combined to estimate thepotential transport flux of krill orthogonal to each section of the transect in the upper 250 m. The watervolume transport between each station pair is shown by the red arrows (Sv), and the blue arrows show theassociated krill flux (106 g s�1).

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

6 of 10

C01029

Georgia is likely to be the major pathway of input of krillinto the western region [Murphy et al., 1998]. However,the surface circulation of the region shows complexinteractions with the local bathymetry [Meredith et al.,2003]. Eddy shedding in the area of the retroflexion ofthe SACCF may be important in moving krill out of theSACCF and toward the west rather than returning to theeast in the offshore flow of the SACCF. Cross-shelftransfer processes associated with the complex bathymetryare also likely to be important in krill movement onto theshelf [Brandon et al., 1999]. Where significant cross-shelfexchange occurs is presently not clear; however, krill onthe shelf in the west may have been entrained onto theshelf on the eastern and southern side of the island beforebeing advected west. Understanding the mechanisms forsuch transfers will be crucial for understanding the linksbetween the offshore ocean systems and the more inshorehigh predation regions.[19] The calculated krill flux through the transect sec-

tion is at least 4 times larger than the estimated krillconsumption by the main seal and penguin predators ofkrill in the region at the same time of year (between 22.5and 42 � 103 t d�1) [Boyd, 2002]. Meeting the predatordemand will be the result of a combination of the influxof secondary production (allochthonous) in the form ofkrill and local production (autochthonous) due to growthof individual krill. A comparison of the flux estimateswith the best available estimates of predator consumption[Boyd, 2002] and krill growth [Atkinson et al., 2001]indicates that the measured influx was much larger thanthe krill consumption and growth flows in the food web(Figure 5). The estimates also suggest there will be avery high rate of turnover of the regional krill stock[Murphy, 1995], although efflux rates are unknown.Obtaining robust estimates of all the various flux ratesin the budget requires targeted studies of the various rate

processes, but the calculations further emphasize thatallocthonous production will be vital in supplying thepredator demand in the region.

3.3. Large-Scale Connections in the ScotiaSea Ecosystem

[20] A key question in analyzing the dynamics of krillpopulations at South Georgia is what are the source regionsfor the krill that occur around the island?Analyses of tracks ofgiant icebergs, drifters and model trajectories all indicate thatkrill within the SACCF that we observed in this study arelikely to have been transported around the eastern end of theisland from the southern central Scotia Sea [Hofmann et al.,1998; Ichii et al., 1998; Murphy et al., 1998; Ward et al.,2002]. Timescales of drifters, iceberg tracks and particletracking using output from theOceanCirculation andClimateAdvanced Model suggest that particles could have beentransported across the Scotia Sea in only 2 or 3 months [Wardet al., 2002]. Sea ice covered the southern Scotia Sea regionduring November 2000, suggesting that krill seen in the off-shelf region at South Georgia during summer may have beenunder the ice in the southern central Scotia Sea during spring.This view is consistent with model analyses of the over-wintering and transport of krill in the region [Fach et al.,2002]. This interaction with the winter sea ice emphasizes theuncertainty in attempting to consider the longer-term, larger-scale transport trajectories of krill. The central southern ScotiaSea is a dynamic and variable region, with sea ice advancingand retreating over water masses originating in the Belling-shausen Sea, the Antarctic Peninsula and the Weddell Sea[Murphy et al., 1998]. This means that until we betterunderstand the under-ice movement and transport processesof krill, both passive and active, the ice-ocean interactionprocesses and genetic differences between krill populationsfrom different regions, it is unrealistic to consider whetherkrill at South Georgia have originated either from west of theAntarctic Peninsula or from the Weddell Sea; rather we canonly consider where krill may have come from within aseason. The results reported in this study and the companionpaper by Ward et al. [2002] give the first description of anassociation between the influx of krill into the South Georgiaregion during summer and the ecosystem of the southernScotia Sea during the spring, and indicate that overwinteringprocesses in sea ice-covered regions directly affect the SouthGeorgia ecosystem 500 to 1000 km further north during thefollowing summer.[21] The large flux of krill through the South Georgia area

observed in this study emphasizes the dynamic large-scalenature of the population dynamics of krill [Murphy andReid, 2001]. The distribution of krill biomass will be afunction of horizontal movements and local production. Thedegree to which local spawning and egg production occursat South Georgia remains unclear and the assumption is thatthe majority (or all) the post-larval individuals are ulti-mately generated outside the region. However, it is possiblethat spawning and egg production are occurring in areasaround South Georgia but that high flow rates, and poten-tially high growth rates, mean that the effects are difficult toobserve. The large input of krill into the region observed inthis study also raises the possibility that there is a largeoutflow of krill from the South Georgia area. As we haveseen in the calculations of flux relative to predator demand,

Figure 5. Estimates of regional krill stock (biomass) anddaily rate estimates for processes increasing or decreasingthe stock during the summer period on the northwest coastof South Georgia. Values in parentheses show the estimatedrange.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

7 of 10

C01029

much of the krill in offshore regions may not be available topredators and may flow through the region and be trans-ported north and then east in the main direction of flow ofthe Antarctic Circumpolar Current. Such an outflow pro-vides a potential mechanism for the return of krill back intospawning regions of the krill population further south andeast such as the Lazarev Sea and the eastern Weddell Gyre[Hardy, 1967; Spiridanov, 1996]. Any such return flowcould generate feedbacks that would affect the longer-termpopulation dynamics across the Scotia Sea region [Murphy,1995].[22] There has been little emphasis on the role of such

horizontal transfer of biotic or indeed abiotic material forthe maintenance of oceanic island ecosystems [Pakhomovand Froneman, 1999; Polis and Hurd, 1996], yet suchlarge-scale transport of biological material is a fundamentalaspect of many marine food webs. Large-scale flows are animportant aspect in the maintenance of regional ecosystemsthroughout the oceans, such as in the eastern north Pacific,the eastern north Atlantic [Agusti et al., 2001; Springer etal., 1999] and the Georges Bank region [Manning et al.,2001; Shore et al., 2000]. Many of these shelf dominatedregions support economically and socially important fish-eries, yet the productivity of these regions will be in partdependent upon the horizontal flux of biological material,which could include recruiting individuals that maintainpopulations. In this regard, the recorded flux of krill will beparticularly important in attempts now being made todevelop localized ecosystem-based principles for the sus-tainable management of harvesting in the Southern Oceanunder the international Convention for the Conservation ofAntarctic Marine Living Resources (CCAMLR) [Constableet al., 2000]. Understanding the flux processes will requirededicated oceanographic field effort with year-round mon-itoring of the rates of throughput of krill.[23] Such allocthonous inputs are known to generate

fundamentally different ecological responses to variationand change because of nonlinearities in the interactions[Jefferies, 2000; Sanchez-Pinero and Polis, 2000]. Biolog-ical fluxes associated with the current flows will thereforebe crucial in determining the dynamics of the planktonicsystem and of the dependent predators in the South Georgiaregion. As the biological flux can be such a significantcomponent of the supply of energy in the regional foodweb, variations in the flux will be a major factor generatinginterannual variability and longer term variation in theSouth Georgia ecosystem [Croxall et al., 1988; Murphy etal., 1998; Priddle et al., 1988]. Fluctuation in the input ofbiological material into local ecosystems such as SouthGeorgia, either as a result of physical variation in the oceancurrents, changes in frontal positions or biological changesupstream, also make these systems sensitive indicators oflarger-scale environmental change. Quantifying the hori-zontal flux of biological material in ocean food webs andassociated levels of variability will be a key requirement inunderstanding the major controls on the operation of oceanecosystems and their response to change.

4. Summary

[24] We have observed that physical advection associatedwith the SACCF on the northern side of South Georgia will

have a number of important effects on the regional ecosys-tem. A large biomass of krill was associated with thesouthern manifestation of the SACCF close to the northernslope in about 2500 m of water. The enhanced biomass ofkrill was associated with the strong northwestward flow ofthe current in the vicinity of the SACCF. We observed thatthe strength of the flow and the depth penetration of theenhanced flow rates mean that physical advection will havedominated the movement of krill in this region. We werethus able to show that the SACCF has a profound role intransporting krill in the South Georgia ecosystem.[25] We have compared the instantaneous flux of krill in

the South Georgia ecosystem with the estimated demand forkrill by predators and shown that the measured flux will havebeen much greater than the demand by the major land-basedkrill predators during the breeding season. The current bestestimates of the growth of krill in the region are less than thepredator demand for krill, although suggestions of a highergrowth rate of krill in the region compared to areas furthersouth [Murphy and Reid, 2001] may mean that growth couldmaintain most of the demand. However, the available evi-dence is that the krill in the region are not a self-sustainingpopulation, since eggs and the early stages of krill are notfound in the area. This evidence emphasizes that at least forsome periods of the year krill will be brought into the regionto maintain the regional stock. The present study indicatesthat very large amounts of krill can be flowing through theregion at any one time but also that the flow is likely to behighly variable. The extent to which the krill flowing throughthe region is available to local predators and howmuch of thekrill in the ocean flow becomes entrained in the island regionis unclear and the interaction of the SACCF with the regionalbathymetry will be crucial in these processes [Meredith et al.,2003].[26] Given the large-scale nature of the SACCF in the

Scotia Sea the study supports suggestions that the front willhave a major role in the transport of krill across the Scotia Seaand in the large-scale population dynamics [Hofmann et al.,1998; Murphy et al., 1998; Ward et al., 2002]. The spatialtransfer of secondary production from one area to another willhave a major impact on the operation of such ecosystems andwill be a crucial element of their response to change [Murphy,1995]. Such processes will also be a key element of otherocean ecosystems affecting their regional biogeochemicalcycling, the plankton population and community dynamicsas well as determining the maintenance and development ofhigher trophic level species, including commercial exploitedfish species. Quantification of such horizontal advection willbe important in developing sustainable management proce-dures for regional ocean ecosystems.

[27] Acknowledgments. We are grateful to the scientists, officers andcrew of JR57 and all the BAS staff associated with the DYNAMOEprogram for their input in making the study possible. We are grateful tothe reviewers for their comments, which have improved the manuscript.

ReferencesAgusti, S., C. M. Duarte, D. Vaque, M. Hein, J. M. Gasol, and M. Vidal(2001), Food-web structure and elemental (C, N and P) fluxes in theeastern tropical North Atlantic, Deep Sea Res., Part II, 48, 2295–2321.

Ansorge, I. J., P. W. Froneman, E. A. Pakhomov, J. R. E. Lutjeharms,R. Perissinotto, and R. C. van Ballegooyen (1999), Physical-biologicalcoupling in the waters surrounding the Prince Edward Islands (SouthernOcean), Polar Biol., 21, 135–145.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

8 of 10

C01029

Atkinson, A. W. M., J. Priddle, G. C. Cripps, P. Ward, and M. A. Brandon(2001), South Georgia, Antarctica: A productive, cold water, pelagicecosystem, Mar. Ecol. Prog. Ser., 216, 279–308.

Bigg, G. (1996), The Oceans and Climate, 266 pp., Cambridge Univ. Press,New York.

Boyd, I. L. (2002), Estimating food consumption of marine predators:Antarctic fur seals and macaroni penguins, J. Appl. Ecol., 39, 103–119.

Boyd, I. L., I. J. Staniland, and A. R. Martin (2002), Distribution of fora-ging by female Antarctic fur seals, Mar. Ecol. Prog. Ser., 242, 285–294.

Brandon, M. A., E. J. Murphy, and M. J. Whitehouse (1999), The shelfbreak front to the east of the sub-Antarctic island of South Georgia, Cont.Shelf Res., 19, 799–819.

Brierley, A. S., D. A. Demer, J. L. Watkins, and R. P. Hewitt (1999a),Concordance of interannual fluctuations in acoustically estimated densi-ties of Antarctic krill around South Georgia and Elephant Island: Bio-logical evidence of same-year teleconnections across the Scotia Sea, Mar.Biol., 134, 675–681.

Brierley, A. S., C. Goss, J. L. Watkins, M. T. Wilkinson, and I. Everson(1999b), Acoustic estimates of krill density at South Georgia, 1981 to1998, CCAMLR Sci., 6, 47–58.

Brierley, A. S., C. Goss, S. A. Grant, J. L. Watkins, K. Reid, M. Belchier,I. Everson, M. J. Jessop, V. Afanasyev, and J. Robst (2002), Multipleacoustic estimates of krill density at South Georgia during 2000/2001reveal significant intra-annual and spatial variability, CCAMLR Sci., 9,71–82.

Bryant, A. D., D. Hainbucher, and M. Heath (1998), Basin-scale advectionand population persistence of Calanus finmarchicus, Fish. Oceanogr., 7,235–244.

Constable, A. J., W. K. de la Mare, D. J. Agnew, I. Everson, andD. Miller (2000), Managing fisheries to conserve the Antarctic marineecosystem: Practical implementation of the Convention on the Con-servation of Antarctic Marine Living Resources (CCAMLR), J. Mar.Sci., 57, 778–791.

Croxall, J. P. (1987), The status and conservation of antarctic seals andseabirds—A review, Environ. Int., 13, 55–70.

Croxall, J. P., T. S. McCann, P. A. Prince, and P. Rothery (1988), Repro-ductive performance of seabirds and seals at South Georgia and SignyIsland, South Orkney Islands, 1976–1987: Implications for SouthernOcean monitoring studies, in Antarctic Ocean and Resources Variability.,edited by D. Sahrhage, pp. 261–285, Springer-Verlag, New York.

Darwin, C. (1884), On the Origin of Species by Means of Natural Selection,6th ed., pp. 323–330, John Murray, London.

Egbert, G. D., A. F. Bennett, and F. M. G. G. (1994), TOPEX/Poseidontides estimated using a global inverse model, J. Geophys. Res., 99,24,821–24,852.

Everson, I. (1977), The Living Resources of the Southern Ocean, 156 pp.,FAO, Rome.

Fach, B., E. E. Hofmann, and E. J. Murphy (2002), Modelling studies ofAntarctic Krill (Euphausia superba) survival during transport across theScotia Sea, Mar. Ecol. Prog. Ser., 231, 187–203.

Foote, K. G., H. P. Knudsen, G. Vestnes, D. N. MacLennan, and E. J.Simmonds (1987), Calibration of acoustic instruments for fish densityestimation: A practical guide, in ICES Coop, Res., 1–69.

Godlewska, M. (1993), Acoustic observations of krill (Euphausia-superba)at the ice-edge (between Elephant I and South Orkney I, Dec 1988/Jan1989), Polar Biol., 13, 507–514.

Gurney, W. S. C., D. C. Speirs, S. N. Wood, E. D. Clarke, and M. R. Heath(2001), Simulating spatially and physiologically structured populations,J. Anim. Ecol., 70, 881–894.

Hardy, A. (1967), Great Waters. a voyage of Natural History to StudyWhales, Plankton and the Waters of the Southern Ocean in the Old RoyalResearch Ship Discovery With the Results Brought up to Date by theFindings of the Discovery II, R.R.S., 542 pp., Collins, London.

Hofmann, E. E., J. M. Klinck, R. A. Locarnini, B. Fach, and E. Murphy(1998), Krill transport in the Scotia Sea and environs, Antarct. Sci., 10,406–415.

Houghton,, J. T., Y. Diny, D. J. Giggs,, M. Noguer, P. J. van der Linden,X. Dai, K. Maskell,, and C. A. Johnson (2001), IPCC 2001: ClimateChange 2001—The Scientific Basis: Contribution of Working Group Ito the Third Assessment Report of the Intergovernmental Panel onClimate Change, 881 pp., Cambridge Univ. Press, New York.

Houry, S., E. Dombrowsky, P. de Mey, and J.-F. Minster (1987), Brunt-Vaisala frequency and Rossby radii in the South-Atlantic, J. Phys. Ocean-ogr., 17, 1619–1626.

Ichii, T., K. Katayama, N. Obitsu, H. Ishii, and M. Naganobu (1998),Occurrence of Antarctic krill (Euphausia superba) concentrations inthe vicinity of the South Shetland Islands: Relationship to environmentalparameters, Deep Sea Res., 45, 1235–1262.

Jefferies, R. L. (2000), Allochthonous inputs: Integrating populationchanges and food-web dynamics, Trends Ecol. Evolution, 15, 19–22.

Laws, R.M. (1984),Antarctic Ecology, 850 pp., Academic, SanDiego, Calif.Loeb, V., V. Siegel, O. HolmHansen, R. Hewitt, W. Fraser, W. Trivelpiece,and S. Trivelpiece (1997), Effects of sea-ice extent and krill or salpdominance on the Antarctic food web, Nature, 387, 897–900.

Madureira, L. S. P., I. Everson, and E. J. Murphy (1993), Interpretation ofacoustic data at two frequencies to discriminate between Antarctic krill(Euphausia superba Dana) and other scatterers, J. Plankton Res., 15,787–802.

Manning, J. P., R. G. Lough, C. E. Naimie, and J. H. Churchill (2001),Modelling the effect of a slope-water intrusion on advection of fish larvaein May 1995 on Georges Bank, J. Mar. Sci., 58, 985–993.

Marr, J. (1962), The natural history and geography of the Antarctic krill(Euphausia superba Dana), Discovery Rep., 32, 33–464.

Meredith, M., J. L. Watkins, E. J. Murphy, and P. Ward (2003), The South-ern ACC Front to the northeast of South Georgia: Pathways, character-istics and fluxes, J. Geophys. Res., 108(C5), 3162, doi:10.1029/2001JC001227.

Miller, D. G. M., and I. Hampton (1989), Biology and Ecology of theAntarctic Krill (Euphausia superba Dana): A review, Sci. Comm. onAntarct. Res., Cambridge, UK.

Murphy, E. J. (1995), Spatial structure of the Southern-Ocean ecosystem—Predator-prey linkages in Southern-Ocean food webs, J. Anim. Ecol., 64,333–347.

Murphy, E. J., and K. Reid (2001), Modelling Southern Ocean krill popula-tion dynamics: Biological processes generating fluctuations in the SouthGeorgia ecosystem, Mar. Ecol. Prog. Ser., 217, 175–189.

Murphy, E., P. Trathan, I. Everson, G. Parkes, and F. Daunt (1997), Krillfishing in the Scotia Sea in relation to bathymetry, including the detaileddistribution around South Georgia, CCAMLR Sci., 4, 1–17.

Murphy, E. J., J. L. Watkins, K. Reid, P. N. Trathan, I. Everson, J. P.Croxall, J. Priddle, M. A. Brandon, A. S. Brierley, and E. Hofmann(1998), Interannual variability of the South Georgia marine ecosystem:Biological and physical sources of variation in the abundance of krill,Fish. Oceanogr., 7, 381–390.

Nicol, S. (2000), Understanding krill growth and aging: The contribution ofexperimental studies, J. Can. Fish. Aquat. Sci., 57, 168–177.

Nicol, S., T. Pauly, N. L. Bindoff, S. Wright, D. Thiele, G. W. Hosie, P. G.Strutton, and E. Woehler (2000), Ocean circulation off east Antarcticaaffects ecosystem structure and sea-ice extent, Nature, 406, 504–507.

Orsi, A. H., T. Whitworth, and W. D. Nowlin (1995), On the meridionalextent and fronts of the Antarctic Circumpolar Current, Deep Sea Res.,42, 641–673.

Pakhomov, E. A., and P. W. Froneman (1999), The Prince Edward Islandspelagic ecosystem, south Indian Ocean: A review of achievements,1976–1990, J. Mar. Syst., 18, 355–367.

Perissinotto, R., R. K. Laubscher, and C. D. McQuaid (1992), Marineproductivity enhancement around Bouvet and the South SandwichIslands (Southern Ocean), Mar. Ecol. Prog. Ser., 88, 41–53.

Polis, G. A., and S. D. Hurd (1996), Linking marine and terrestrial foodwebs: Allochthonous input from the ocean supports high secondary pro-ductivity on small islands and coastal land communities, Am. Nat., 147,396–423.

Priddle, J., J. P. Croxall,, I. Everson, R. B. Heywood,, E. J. Murphy, P. A.Prince, and S. C. B. (1988), Large-scale fluctuations in distribution andabundance of krill—A discussion of possible causes, in Antarctic Oceanand Resources Variability, edited by D. Sahrhage, pp. 169–182, Springer-Verlag, New York.

Sanchez-Pinero, F., and G. A. Polis (2000), Bottom-up dynamics ofallochthonous input: Direct and indirect effects of seabirds on islands,Ecology, 81, 3117–3132.

Scientific Committee Convention on the Conservation of Antarctic MarineLiving Resources (SC-CAMLR) (1991), Report of the Working Group onKrill, in Report of the Tenth Meeting of the Scientific Committee (SC-CAMLR-X), pp. 109–204, Conv. on the Conserv. of Antarct. Mar. LivingResour., Hobart, Tasmania, Australia.

Scientific Committee Convention on the Conservation of Antarctic MarineLiving Resources (SC-CAMLR) (2000), Report of the Working Group onEcosystem Monitoring and Management, in Report of the NineteenthMeeting of the Scientific Committee (SC-CAMLR-XIX), pp. 113–273,Conv. on the Conserv. of Antarct. Mar. Living Resour., Hobart, Tasma-nia, Australia.

Shore, J. A., C. G. Hannah, and J. W. Loder (2000), Drift pathways on thewestern Scotian Shelf and its environs, J. Can. Fish. Aquat. Sci., 57,2488–2505.

Spiridanov, V. A. (1996), A scenario of the late-Pleistocene-Holocenechanges in the distributional range of Antarctic krill (Euphausiasuperba), Mar. Ecol., 17, 519–541.

Springer, A. M., J. F. Piatt, V. P. Shuntov, G. B. Van Vliet, V. L. Vladimirov,A. E. Kuzin, and A. S. Perlov (1999), Marine birds and mammals of thePacific subarctic gyres, Prog. Oceanogr., 43, 443–487.

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

9 of 10

C01029

Tarling, G. A., J. B. L. Matthews, R. Saborowski, and F. Buchholz (1998),Vertical migratory behaviour of the euphausiid, Meganyctiphanes norve-gica, and its dispersion in the Kattegat Channel, Hydrobiologia, 376,331–341.

Thorpe, S. E., K. J. Heywood, M. A. Brandon, and D. P. Stevens (2002),Variability of the southern Antarctic Circumpolar Current front north ofSouth Georgia, J. Mar. Syst., 37, 87–105.

Trathan, P. N., E. J. Murphy, J. P. Croxall, and I. Everson (1998a), Use ofat-sea distribution data to derive potential foraging ranges of macaronipenguins during the breeding season, Mar. Ecol. Prog. Ser., 169, 263–275.

Trathan, P. N., M. A. Brandon, E. J. Murphy, and S. E. Thorpe (1998a),Transport and structure of the Antarctic Circumpolar Current to the northof South Georgia, Geophys. Res. Lett., 27, 1727–1730.

Tynan, C. T. (1998), Ecological importance of the southern boundary of theAntarctic Circumpolar Current, Nature, 392, 708–710.

Ward, P., et al. (2002), The Southern Antarctic Circumpolar Current Front:Physical and biological coupling at South Georgia, Deep Sea Res., 49,2183–2202.

Watkins, J. L., D. J. Morris, C. Ricketts, and J. Priddle (1986), Differencesbetween swarms of Antarctic krill and some implications for samplingkrill populations, Mar. Biol., 93, 137–146.

Watkins, J. L., A. W. A. Murray, and H. I. Daly (1999), Variation in thedistribution of Antarctic krill Euphausia superba around South Georgia,Mar. Ecol. Prog. Ser., 188, 149–160.

�����������������������M. P. Meredith, E. J. Murphy, S. E. Thorpe, P. N. Trathan, P. Ward, and

J. L. Watkins, British Antarctic Survey, High Cross, Madingley Road,Cambridge, CB3 0ET, UK. (ejmu@bas.ac.uk)

C01029 MURPHY ET AL.: ADVECTION OF ZOOPLANKTON

10 of 10

C01029