coupling between upper ocean layer variability and size

MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 499: 3546, 2014doi: 10.3354/meps10668

Published March 3

INTRODUCTION

The dynamics of phytoplankton in the upper oceanis linked to both the upper mixed layer (UML) and tomesoscale and submesoscale variability. The UML is

the result of complex interactions between atmos-pheric forcing and this variability (e.g. Ferrari & Boc-caletti 2004). For example, mesoscale eddies maycause deepening or shoaling of the UML, therebyinfluencing the turbulent mixing and affecting the

Inter-Research 2014 www.int-res.com*Corresponding author: [email protected]

Coupling between upper ocean layer variability and size-fractionated phytoplankton

in a non-nutrient-limited environment

Pablo Sangr1,*, Cristina Garca-Muoz2, Carlos M. Garca3, ngeles Marrero-Daz4, Cristina Sobrino5, Beatriz Mourio-Carballido5,

Borja Aguiar-Gonzlez4, Cristian Henrquez-Pastene6, ngel Rodrguez-Santana4,Luis M. Lubin2, Mnica Hernndez-Arencibia4, Santiago Hernndez-Len1,

Elsa Vzquez5, Sheila N. Estrada-Allis4

1Instituto Universitario de Oceanografa y Cambio Global, Universidad de Las Palmas de Gran Canaria, 35017 Las Palmas de Gran Canaria, Spain

2Departamento de Ecologa y Gestin Costera, Instituto de Ciencias Marinas de Andaluca (ICMAN-CSIC), 11510 Puerto Real, Cdiz, Spain

3Departamento de Biologa, Facultad de Ciencias del Mar y Ambientales, Universidad de Cdiz, 11510 Puerto Real, Cdiz, Spain4Departamento de Fsica, Universidad de Las Palmas de Gran Canaria, 35017 Las Palmas de Gran Canaria, Spain

5Departamento de Ecoloxa e Bioloxa Animal, Universidade de Vigo, 36200 Vigo, Pontevedra, Spain6Departamento de Geofsica, Universidad de Concepcin, 160-C Concepcin, Chile

ABSTRACT: We describe the coupling between upper ocean layer variability and size-fraction-ated phytoplankton distribution in the non-nutrient-limited Bransfield Strait region (BS) of Antarc-tica. For this purpose we use hydrographic and size-fractionated chlorophyll a data from a transectthat crossed 2 fronts and an eddy, together with data from 3 stations located in a deeply mixedregion, the Antarctic Sound (AS). In the BS transect, small phytoplankton (20 m ESD) accounted for 80% of totalchl a. The proportion of large phytoplankton increases as the depth of the upper mixed layer(ZUML), and the corresponding rate of vertical mixing, increases. We hypothesize that this changein phytoplankton composition with varying ZUML is related to the competition for light, and resultsfrom modification of the light regime caused by vertical mixing.

KEY WORDS: Physical-biological coupling Mesoscale Submesoscale Vertical mixing Phytoplankton composition Antarctica

Resale or republication not permitted without written consent of the publisher

FREEREE ACCESSCCESS

Mar Ecol Prog Ser 499: 3546, 2014

rate of vertical mixing of phytoplankton (e.g. Thomp-son et al. 2007). One major source of submesoscalevariability is small filamentous fronts that originate ateddy boundaries or at mesoscale fronts. Intense ver-tical velocity associated with vertical circulation cellsmay cause the fronts to slump, resulting in restratifi-cation and enhancement of turbulent mixing (Thomaset al. 2008, DAsaro et al. 2011). Vertical circulationcells related to meso scale and submesoscale variabil-ity, mixing processes, and their interaction may mod-ify the vertical ex changes of tracers and hence influ-ence phytoplankton dynamics (e.g. Nagai et al. 2006,Klein & Lapeyre 2009).

Vertical mixing may influence light availability,nutrient entrainment and uptake, zooplankton en -counter probability, and settling velocity (e.g. Peter -sen et al. 1998, Huisman et al. 2002, Man & Lazier2006, Kirboe 2007, Behrenfeld 2010). Turbulence isnot always active in the UML and therefore it isimportant to distinguish between the terms UML andmixing layer. The mixing layer is also a uniformdensity layer but where turbulence is active through-out (Brainerd & Gregg 1995). The vertical extent ofthe upper mixing layer is particularly important forphytoplankton modulation by turbulence, because itdetermines the rate of vertical mixing of the cells. Ina non-nutrient-limited environment it may be ex pec -ted that this rate will essentially control the positionand movement of the cells in the light gradient, aswell as the zooplankton-phytoplankton encountersthat can reduce/increase the grazing pressure.

Since the early Sverdrup (1953) critical depth hypo - thesis, many studies have used either the mixinglayer depth or the upper mixed layer depth (ZUML) asproxies for turbulence forcing. The effects of thisforcing in less complex non-nutrient-limited environ-ments are twofold: modulation of the population netgrowth rate, and modulation of the composition ofthe plankton community (e.g. Petersen et al. 1998,Huisman et al. 2004, Behrenfeld 2010). Hewes et al.(2008) studied the relationship between hydro-graphic properties, nutrients and phytoplankton bio-mass using data collected during a summer cruise ina region near the South Shetland Island that includesour study region (Fig. 1). In areas of high iron con-centrations they observed an inverse relationshipbetween ZUML and total chl a, but with a very lowcoefficient of determination (R2 = 0.176). When clima-tological data for the region are considered (Heweset al. 2009), this relationship is much more robust(R2 = 0.73). Climatological data analysis also showedthat blooms in the region co-vary with shallow UMLdepths (e.g. Mitchell & Holm-Hassen 1991, Hewes et

al. 2009). Contrary to these observations, Holm-Hansenet al. (2004) did not observe any signi ficant correla-tion between ZUML and phytoplankton abundance ina much broader survey. They concluded that low ironconcentrations are the major factor controlling phyto-plankton biomass. Recent observations by Mendes etal. (2012) in the vicinity of James Ross Island (Wed-dell Sea) show bloom levels of phytoplankton (~5 mgchl a m3) in deep mixed layers (ZUML ~ 80 m) andpoorly stratified water.

Several studies in our surveyed region have alsoexamined the possible relationship between the mix-ing layer depth/water stratification and phytoplank-ton composition (e.g. Kopczynska & Ligowsky 1985,Kang & Lee 1995, Kang et al. 2001, Mendes et al.2012, 2013). Kopczynska & Ligowsky (1985) first sug-gested a connection between variability of phyto-plankton composition and water mass properties forthis region. Kang & Lee (1995) observed the domi-nance of nano-sized flagellates in the more stratifiedwaters of the Bransfield Strait, whereas nano- and

36

59.5 59 58.5 58 57.5 57 56.5 56

62

62.5

63

63.5

64

Longitude

Latit

ude

Fig. 1. Location of sample stations for the study of upperocean layer variability and size-fractionated phytoplanktondistribution in the Bransfield Strait (Stns 1 to 12) and Antar-tic Sound (Stns 19, 21 and 22), Antarctica. Blue dots: stationswhere standard measurements were taken; red dots: sta-tions additionally sampled with a microstructure turbulenceprofiler. Gray lines are isobaths with a contour interval of

100 m

Sangr et al.: Ocean surface layer and phytoplankton dynamics

micro-sized dia toms dominate in the more homoge-neous waters of Drake Passage. Recent results from 3summer data collections by Mendes et al. (2013)showed that dia toms dominate in deeper mixed layerconditions whereas nano-sized cryptophytes domi-nate in more stratified conditions. Their February2010 observations across Bransfield Strait are of par-ticular interest because they almost coincided in timeand location with the present study. As depicted intheir Fig. 6, there is a clear correlation between thevariability of the ZUML and phytoplankton composi-tion along the cited transect. For stratified environ-ments (ZUML ~ 15 to 25 m) nano-sized phytoplanktondominated. In a homogeneous water column with avery deep mixed layer (ZUML > 160 m) diatoms domi-nated instead.

This study aimed to explore observationally howthe distribution and composition of phytoplanktonare connected with the variability of the ocean sur face layer. The observations by Mendes et al.(2013), although made in the same area surveyedby us, were, however, mainly decoupled from themesoscale variability, as their spatial resolutionwell exceeded the mean local first baroclinicRossby radius of deformation (Rd) (~10 km; Cheltonet al. 1998). Therefore, these observations do notresolve the effects of mesoscale variability andtheir interaction with the UML on the planktondynamics. In January 2010, during the mid-australsummer, an interdisciplinary survey was made overa region located near the South Shetland Island atmesoscale resolution (stations ~9 km apart, Fig. 1).The general coupling between the mesoscale andphytoplankton distributions and their taxonomiccomposition has already been outlined in the studyby Garca-Muoz et al. (2013) from a more detailedbiological perspective. In the present study wefocused attention on the processes underlying thiscoupling. Particular em pha sis was given to theBransfield Strait region (Fig. 1), where distinctwater masses meet and there is a rich mesoscalevariability (Sangr et al. 2011). Moreover nutrientconcentrations in this southern region were abovelimiting thresholds for phytoplankton growth, bothduring this cruise and in past observations (Heweset al. 2009, Teira et al. 2012). This allowed ouranalysis of the effects of the physical environmentto focus on light rather than nutrient availability asthe main factor affecting phytoplankton growth inconditions of vertical mixing. Both circumstancesmake the selected region suitable for the study ofthe coupling between ocean surface layer variabil-ity and phytoplankton dynamics.

MATERIALS AND METHODS

Physical data collection

Data were acquired during the interdisciplinarycruise COUPLING conducted in January 2010 onboard RV Hesprides, during the austral mid- summer near the South Shetland Islands. This surveyformed part of the project COUPLING (Physical-bio-logical coupling at the mesoscale range around theSouth Shetland Islands) of the Spanish Antarctic Pro-gram. For this study we selected 1 transect of 12 sta-tions across the Central Bransfield Strait and 3 stations (Stns 19, 21 and 22) in the Antarctic Sound(Fig. 1). Distinct water mass compositions and richmesoscale variability were the motivation for thisselection. In order to resolve the mesoscale, stationswere sampled every 5 nautical miles (~9 km). Thetransect sampling can be viewed as quasi-synoptic as


ZUML was inferred from CTD profiles using theKara et al. (2000) algorithm. First, the potential den-sity anomaly at 10 m is chosen as an initial referencedensity ref. Second, a search is made of the poten-tial density anomaly profile for regions of uniformpotential density anomaly. These regions were de -fined as any pair of adjacent values where the poten-tial density anomaly variation , corresponding to achange of potential temperature of 0.8C, was


the ADCP indicate a jet-like structure that is associ-ated with the Bransfield Current; its axis is centered atStn 2, where the current velocity can reach up to 0.4 ms1. This current is the major component that drivesthe BCS, transporting TBW along the slope of theSouth Shetland Islands. In Fig. 2 we can also identifythe signature of an anticyclonic eddy through thedeepening of the isopycnals be tween Stns 3 and 6 justsouth of the Bransfield Front. The signal of the eddy isvisible until ~240 m depth within the TBW range. Dueto geostrophic adjustment there is a deepening of the

isopycnals at the center of the eddy that induces an in-crease of ZUML between Stns 3 and 6. Eddy samplingtook 28 h; during this period it is unlikely that theeddy displacement was greater than 10 km and thusits sampling can be considered to be quasi-synoptic.This is supported by the fact that westward shelf-ad-vection speeds of the eddies at high latitudes are verylow (Chelton et al. 2007); advection by the mean flowcan be discounted as outside the Bransfield Front itdoes not exceed 0.1 m s1 (Sangr et al. 2011). Finally,close to the Antarctic Peninsula, we can recognize thesignal of a shallow frontal region through the steeplytilted isopycnals between Stns 7 and 11 that Sangr etal. (2011) named the Peninsula Front. TBW and TWWconverge at the surface of the Peninsula Front. Theabove structures have been detected in all mesoscaleresolution surveys of the region and, hence, they canbe considered as permanent features of the circulationof this region during the austral summer (Sangr et al.2011).

From the combined distribution of fluorometer cal-ibrated total chl a and isopycnals show in Fig. 2, wecan extract signs of relatively strong subductionregions at both fronts. These regions are indicated byrather high values of total chl a well below ZUML, andthe euphotic layer forming deep column-like struc-tures. At the Bransfield Front, values of total chl a >0.2 mg m3 were observed down to 180 m depth, indi-cating phytoplankton subduction on the light (anticy-clonic) side of the front at Stn 1 (Fig. 2). At the Penin-sula Front heavy (cyclonic) side (between Stns 9 and11) there is also clearly evidence of subductingphytoplankton. The distribution of total chl a has acolumn-like structure between Stns 9 and 11, that issub ducted down as far as 300 m at Stn 10 (Fig. 2).Models and observations (e.g. Nagai et al. 2006, Pal-ls-Sanz et al. 2010) have related these frontal sub-duction regions with the descending part of verticalcirculation cells linked to ageostrophic secondary cir-culation of the front.

Although our data set is too incomplete (i.e. 2-dimensional and with insufficient resolution at bothfronts) to diagnose the vertical circulation, the occur-rence of deep chl a columns supports, at least quali-tatively, a downward motion at one side of bothfronts. Observations in other frontal regions (e.g. Pol-lard & Regier 1992, Palls-Sanz et al. 2010, Thomas &Joyce 2010), show that the subducting region on oneside of a front is accompanied by an upwelling regionon the other side. On the upwelling side of the frontrestratification may take place due to slumping of theisopycnals induced by the ascending water (Nagai etal. 2006, Thomas et al. 2008). At the heavy (cyclonic)

39

Chl a(mg m3)

Stn number

0 20 40 60 80 100

Distance (km)

Dep

th (m

)

0

20

40

60

80

100

120

140

160

180

200

220

240

260

280

300

320

340

1 2 3 4 5 6 7 8 9 10 11 12

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

BF AE PF

Fig. 2. Mesoscale variability in the Bransfield Strait (Stns 1 to12): Thin black lines are isopycnals (potential density anom-aly , kg m3); the color filled contour plot shows total chl a(mg m3) calculated from fluorometer calibrated data (colorfilled contour plot); the bold black line shows the euphoticlayer depth (ZEU) and bold red line shows the upper mixedlayer depth (ZUML). Circled stations on the top axis indicateStns 2, 6 and 9 where turbulence profiles were recorded.The locations of the Bransfield Front (BF), the anticycloniceddy (AE) and the Peninsula Front (PF) are indicated on thetop axis. The bottom topography is indicated by the shaded

grey area at the right lower corner


side of the Bransfield Front (between Stns 2 and 3)there are signs of restratification evidenced by thesqueezing of isopycnals at the depth range 30 to 60 mand by the slumping of the 2 nearest surface isopyc-nals accompanied by a shoaling of the UML. On thelight (anticyclonic) side of the Peninsula Front, be -tween Stns 8 and 9, the UML also shoals, indicatingrestratification. This shoaling is due to the decreasedtilt of the 27.5 and 27.55 kg m3 isopycnals near thesurface (20 to 40 m) and the opposing tilt of the 27.6and 27.65 kg m3 isopycnals at deeper layers (60 to80 m). Although the restratification observed on theheavy (cyclonic) side of the Bransfield Front and onthe light (anticyclonic) side of the Peninsula Frontmay be due to internal motions, the occurrence ofan upwelling region accompanying the subductingregion should not be disregarded.

Upper mixed layer (UML) variability

As shown in Fig. 2, the ZUML is modulated by thepresence of mesoscale structures. It is deeper on the

light (anticyclonic) side of the Bransfield Front atStn 1, in the center of the eddy (Stns 4 and 5), on theheavy (cyclonic) side of the Peninsula Front (Stns 10and 11), and at TWW (Stn 12), coinciding with geo -strophically depressed isopycnals and/or with thesubducting regions. At Stn 12 the ZUML reaches thebottom because the whole water column is occupiedby TWW which is homogeneous. The ZUML is shal-lower at the axis and at the heavy side of the Brans-field Front (between Stns 2 and 3) and at the axisand the light side of the Peninsula Front (Stns 7, 8and 9). It is important to investigate whether turbu-lent mixing is active over the whole depth range ofthe observed UMLs and thus if the UMLs are actu-ally mixing layers. Regarding this question, the profile at Stn 6 located inside the eddy shows a tur-bulent upper mixing layer throughout the depthrange of the UML, where values of are maximum(Fig. 3b). If we choose the depth of this mixing layeras the depth where the value of drops by oneorder of magnitude (Brainerd & Gregg 1995), thisdepth coincides ap proximately with the ZUML asindicated by the corresponding density profile

40

Fig. 3. Averaged vertical profiles of kinetic energy dissipation rate (bars, W kg1), potential density anomaly (black lines,kg m3), and fluorescence (green lines, relative units) as obtained from microstructure turbulence profilers at (a) the BransfieldFront (Stn 2), (b) the anticyclonic eddy (Stn 6), (c) the Peninsula Front (Stn 9) and (d) the Antarctic Sound (based on average

values measured at Stns 19, 21 and 22)


(Fig. 3b). The profile for Stn 2 also shows a maxi-mum of in the UML portion of the water column(Fig. 3a). Vertical profiles of show a deeper mixinglayer at the Peninsula Front axis (Stn 9) when com-pared with those at the Bransfield Front axis (Stn 2)and at the eddy (Stn 6) (Fig. 3a,b,c). This deep-reaching mixing may be related to frontal in -stabilities that may en hance mixing as observed byDAsaro et al. (2011) in a small submesoscale front.Stn 9 was located at the mouth of a submarinecanyon. This complex bathymetry may favor thegeneration of internal motion that could be alsoresponsible for the ob served deep mixing.

Unfortunately we do not have direct measurementsof turbulence from all the stations. An indirect ap -proach to infer whether the mixed layer correspondsto an active turbulent layer is to quantify the Thorpescale (Thorpe 1977). The Thorpe scale measures theinversions in a vertical density profile (e.g. Gargett &Garner 2008). Thorpe (1977) related these inversionsto the 3-dimensional turbulent eddy field overturn-ing. Therefore, the appearance of the inversion in avertical density profile is an indicator of turbulentmixing (Gargett & Garner 2008). Fig. 4 shows the dis-tribution of the Thorpe scale along the BransfieldStrait transect. In general, the ZUML effectivelymatches the base of an upper active turbulent layerwhere inversions are frequent. The most active tur-bulent layer is observed within the UML of the anti-cyclonic eddy between Stns 3 and 6. Another activeturbulent region corresponds to Stns 10, 11 and 12,where the inversion reaches the ZUML indicating thatthe whole UML is actually an active turbulent layer.Although intermittent, the density profile of Stn 1also shows inversions down to the ZUML, indicatingturbulent activity at the depth range of the UML. Atthe axis of the Peninsula Front (Stn 9) there areintense inversions below the UML; this indicates thatthe UML and the mixing layer are decoupled, as theturbulence profiles further point out.

Antarctic Sound turbulent environment

Fig. 3d shows mean profiles of density, fluorescenceand as obtained from the microstructure turbulenceprofiler in the Antarctic Sound. Notice the higher lev-els of turbulence throughout the water column incomparison with the Bransfield Strait profiles. In factthe mean value of the 9 profiles recorded in theAntarctic Sound for the depth range 20 to 100 m, =5.73 107 W kg1, was ~5-fold higher than thoserecorded along the Bransfield Strait transect, =

1.22 107 W kg1. Antarctic Sound density profileswere homogeneous over the full depth range(Fig. 3d), and composed of Weddell Sea Shelf Water(WSSW). As TWW (potential temperature, = 1.30to 0.50C, salinity, S = 34.37 to 34.5) is a modifica-tion of WSSW ( = 1.73 to 0.93C, S = 34.37 to 34.5),the 2 water masses are not very distinct; they arecold, salty and homogeneous, with WSSW beingslightly colder. Therefore the water mass propertiesare very similar south from the axis of the PeninsulaFront and at the Antarctic Sound. Because the den-sity profiles are homogeneous, instead of defining aZUML we computed the mixing layer depths as thedepth where is reduced by an order of magnitude:hence 350 m for Stn 19, 220 m for Stn 21, and 320 mfor Stn 22. Notice that those mixing layer depths aremore than 3 fold higher than the observed meanZUML (60 m) at the Bransfield Strait.

41

0 20 40 60 80 100Distance (km)

0

20

40

60

80

100

120

140

Dep

th (m

)

1 2Stn number

3 4 5 6 7 8 9 10 11 12

0.50.91.31.72.12.52.93.33.74.14.54.95.35.76.16.56.97.37.78.18.58.99.39.7

Thorpescale(m)

Fig. 4. Thorpe scale distribution (filled color contour plot) inthe Bransfield Strait. Thorpe scale units are in meters andprovide a measure of vertical overturning and thus an indica-tor for turbulent mixing. Turbulent mixing will be moreactive for large Thorpe values and less active for low Thorpevalues. Note that the largest values are observed in the UML,coinciding with the largest values of (see Fig. 3). See Fig. 2legend for explanation of superimposed black and red lines


Coupling total chl a with mesoscale variability and mixing

Maximum values of total chl a, as obtained fromthe fluorometer calibrated data (Fig. 2b), were ob - served near the surface on the light (anticyclonic)side and at the axis of the Peninsula Front (Stns 7, 8,and 9), and at the edge of the Anticyclonic Eddy(Stn 6). At the Peninsula Front stations these maximacoincide with the shallowest UML (54, 34, and 18 mres pectively) and are thus located on the restratifiedside of the front. From this maximum, the phyto-plankton is subducted and mixed within the heavy(cyclonic) side of the front (between Stns 9 and 11).This subducting region acts as a sink for the cells thatare transported to the deep ocean interior down to atleast 300 m. Therefore the cross-frontal distributionof total chl a is asymmetric, with a shallow local maxi -mum on the restratified side of the front and a deepcolumn-like structure on the subducting side.

At TWW (Stn 12), total chl a distribution is alsodiluted over the whole water column, probably dueto a high rate of vertical mixing as suggested by anUML reaching the bottom (115 m), and by the highdensity of inversions (Fig. 4). As shown in Fig. 2, themaximum at the Peninsula Front diffuses towards theeddy, attenuating its intensity and diluting in theUML. This suggests that the eddy entrains the phyto-plankton from the Peninsula Front source, althoughin situ net growth should not be disregarded. Theentrained phytoplankton will be subject to mixingand accumulate at the eddy UML. This can explainthe relatively high values of total chl a at subsurfacelevels (30 to 60 m) inside the eddy. The largest den-sity inversions were observed at the eddy, suggestingstrong mixing (Fig. 4). This is confirmed by directmeasurements of turbulence at Stn 6 (Fig. 3b).

At the axis and at the heavy (cyclonic) side of theBransfield Front (Stns 2 and 3) a relative maximum oftotal chl a at the base of the UML was observed but ithad lower concentrations than those at the PeninsulaFront. As for the case of the Peninsula Front, thiscoincides with the restratified side of the front evi-denced by a local minimum of the ZUML. On the light(anticyclonic) side located at Stn 1, the concentrationof total chl a is again diluted in the water column,coinciding with a deep UML. As already mentioned,significant values of total chl a were observed until180 m, well below the ZUML and ZEU, indicating sub-duction. Similarly to the case of the heavy (cyclonic)side of the Peninsula Front, this subduction and highrate of vertical mixing may be responsible for theobserved total chl a dilution throughout the water

column. Again, as was the case for the PeninsulaFront, the cross-frontal distribution of total chl a wasasymmetric, with a deep column-like structure indi-cating subduction on one side, and a shallow localmaximum on the restratified side.

A plausible explanation for this asymmetric distri-bution may be related to the occurrence of verticalcirculation cells as observed in other frontal regions(e.g. Palls-Sanz et al. 2010) as previously suggested.The descending part of these cells may cause thesubducting, column-like structure of chl a in the deepUML part of the fronts. At the restratified part of thefronts, the shallow local maximum may be related tothe ascending part of the vertical cell that will trans-port phytoplankton cell to the surface where higherlight levels will increase their net growth rate. Otherplausible mechanisms that may explain the observedlocal maxima on the sides of the fronts are the occur-rence of a confluent flow and/or the advection ofphytoplankton from an upstream source.

Preference of large plankton for well-mixedenvironments

When correlating phytoplankton size fraction vari-ability with the physical environment, the first no -ticeable point is that in the relatively high turbulentenvironment of the Antarctic Sound large phyto-plankton dominate, whereas in the low turbulentregion of the Bransfield Strait small phytoplanktondominate (Fig. 5). In the Antarctic Sound, where theupper turbulent active layer is very deep (>300 m,Fig. 3d), the chl a profiles are homogeneous, with amean value of 1.8 mg m3 for larger phytoplanktonand one order of magnitude less, 0.3 mg m3, forsmall phytoplankton. In the Bransfield Strait (loweraverage turbulence) the picture is reversed. There,the maximum values for small phytoplankton largelyexceed 1 mg m3 whereas the values for large phy -toplankton do not exceed 0.5 mg m3 (Fig. 5a,b). Garca-Muoz et al. (2013), using Flow-CAM andCHEM TAX software, observed that the nano- phytoplankton size range is the most abundant alongthe Bransfield Strait transect. It is mainly composedof small diatoms followed by haptophytes, prasino-phytes and cryptophytes. At the Antarctic Sound,they observed that phytoplankton is largely com-posed of microplanktonic diatoms, mainly Thalassio -sira sp. (Garca-Muoz et al. 2013).

The general pattern described above suggests thatlarge phytoplankton have a preference for thoseenvironments where the rate of vertical mixing is

42


high, with very deep active turbulent upper layerssuch as those we observed at the Antarctic Soundstations. In this regard, if we inspect in more detailthe Bransfield Strait section corresponding to thelarge phytoplankton (Fig. 5b) we can appreciate that,except for the near surface at Stn 2, clearly higher(>0.25 mg m3) concentrations of large phytoplank-ton chl a are found only at those stations coincidingwith the deepest UMLs. This is the case of the lightside of the Bransfield Front (Stn 1, ZUML = 91 m), theheavy side of the Peninsula Front (Stn 11, ZUML =117 m) and the TWW (Stn 12, ZUML = 115 m). TheThorpe scale distributions at these stations indicatesthat the UML coincides with an active turbulentupper layer (Fig. 4), suggesting a high rate of verticalmixing. Therefore, a close inspection of the Brans-field Strait transect supports the above hypothesisconcerning the preference of large phytoplankton forthose environments where the rate of vertical mixingassociated with deep UMLs is high. In contrast to thelarge phytoplankton distribution, size-fractionatedchl a for small phytoplankton shows appreciable val-ues (>0.25 mg m3) all along the Bransfield Straittransect. As the phytoplankton are mainly composedof small phytoplankton (80%) in this transect, theirdistribution mirrors the distribution for total chl a

(Fig. 2b). Thus the phytoplankton distribution res -ponds to the same physical forcing as suggestedabove for total chl a, which is mainly coupled withmesoscale structure variability.

Phytoplankton composition and ZUML variability

The above observations of size-fractionated chl adistribution suggest a covariance between large phy -to plankton and the ZUML, and hence a relationshipbetween this component and the rate of vertical mix-ing. To explore this, in Fig. 6 we plotted the percent-age of large phytoplankton versus ZUML for all ourstations. Fig. 6 also shows the mean values of for theBransfield Strait and Antarctic Sound regions.

There is a clear linear relationship (R2 = 0.86) be -tween the ZUML/mixing layer depth and the percent-age of large phytoplankton (Fig. 6), indicating thatthe proportion of large phytoplankton increases withthe deepening of the ZUML. For the shallowest ZUML,as is the case on the light side of the Peninsula Front(Stns 8 and 9) and the edge of the Bransfield Front(Stn 3), the percentage of large phytoplankton is lessthan 15%. In those stations of the Bransfield Straitwhere the ZUML is deep and well below the ZEU, as is

43

Chl a < 20 m

mg m3 mg m3

Antarctic Sound

a bc

Stn number

Chl a > 20 m

Chl

a

20

m

Chl a (mg m3)Stn number

0 20 40 60 80 1000 20 40 60 80 100

Distance (km)Distance (km)

0

20

40

60

80

100

120

140

0

201.651.551.451.351.251.151.050.950.850.750.650.550.45

0.35

0.25

0.15

40

60

80

100

120

140

1 0 0.5 1 1.5 2 2.52 3 4 5 6 7 8 9 10 11 121 2 3 4 5 6 7 8 9 10 11 12

0.15

0

25

50

75

100

0.2

0.25

0.3

0.35

0.4

Dep

th (m

)

Fig. 5. Distributions of (a) small phytoplankton (chl a < 20 m equivalent spherical diameter [ESD]) and (b) large phytoplank-ton (chl a > 20 m ESD) in the Bransfield Strait (filled contour plots). See Fig. 2 legend for explanation of superimposed blackand red lines. (c) Depth profiles of small (blue) and large (red) phytoplankton at the Antarctic Sound stations. The bold black

line shows the euphotic layer depth (ZEU)


the case for Stns 1, 11 and 12, the percentage of largephytoplankton doubles, reaching values greater than30% of the total chl a. As already mentioned, theThorpe scale distribution indicates that turbulence isactive throughout the ZUML. At the Antarctic Soundstations, where the upper active turbulent layer is 3to 4-fold deeper than the mean ZUML at the Bransfieldstations, the proportions of large and small phyto-plankton are reversed, with large phytoplanktonaccounting for about 80% of the total. The largephytoplankton biomass in terms of depth-integratedchl a concentration down to 100 m (~130 mg m2) isone order of magnitude greater than the average bio-mass at the Bransfield Strait stations (~14 mg m2).Recall that in the first 100 m and thus the level ofturbulence of the Antarctic Sound stations was about5-fold higher than at Bransfield Strait stations (Stns 2,6, and 9). This covariance of the ZUML and the per-centage of large phytoplankton reinforces the hypo -thesis that this size fraction has a preference for thosephysical environments where the upper turbulentlayer is deep and, hence, where the rate of verticalmixing is high. Therefore it is also suggested that thevariability of the rate of vertical mixing modulatesthe phytoplankton composition.

One may argue that the increase of large phyto-plankton biomass at the Antarctic Sound (Stns 19, 21and 22) may be due to its advection from an upstreamsource. However this scenario is unlikely as ourADCP data indicates that during the sampling of theAntarctic Sound stations there was a mean inflow of~0.2 m s1 entering from the Bransfield Strait, wherelarge phytoplankton biomass was very low. Althoughour ADCP records are only 30 h long, ADCP climato-logical observations of Savidge & Amft (2009) alsoshow this inflow from the Bransfield Strait for thesummer period.

As models predict (e.g. Huisman et al. 2002), wepropose 2 mechanisms for explaining the observedhigh biomass of large phytoplankton in deep well-mixed layers. The first mechanism is that largephytoplankton have a moderate sinking rate, so thata minimum level of turbulence is necessary to main-tain cells for long enough in the water column tohave a positive net growth (Huisman et al. 2002). Thesecond mechanism is based on the assumption thatgrowth rates in the upper part of the water columnexceed vertical mixing (Huisman et al. 2002). For thiswe assume that the rate of vertical mixing does notexceed a critical level. The order of magnitude of ourturbulence measurements for high turbulent areas( ~ 107 W kg1) may be considered as an intermedi-ate level of turbulence when compared to other ob -ser vations of UMLs where values of usally rangebe tween ~109 and

20

m)

Low turbulence = 1.22 x 107 (W kg1)

High turbulence = 5.73 x 107 (W kg1)

y = 0.3x + 4.8R2 = 0.86

0 50 100 150 200 250 300 350

100

90

80

70

60

50

40

30

20

10

21

1

12

11

10

5387

4

6

2

9

2219

Fig. 6. Large phytoplankton (chl a > 20 m) as a percentageof total extracted chl a integrated down to 100 m in relationto the depth of the upper mixed layer (ZUML). Numbered dotsshow results for each station. Mean values of the kineticenergy dissipation rate for the relatively low turbulenceenvironment of the Bransfield Strait (Stns 1 to 12) and for therelatively high turbulence environment of the Antarctic

Sound (Stns 19, 21 and 22) are also indicated


also been reported that Antarctic diatoms activelyutilize the photoprotective xanthophyll cycle forheat dissipation, minimizing photoinhibition (Krop-uenske et al. 2009, Van de Poll et al. 2011). As re -ported in the literature, Thalassio sira sp., which inthis case is one of the main taxa composing largephytoplankton, exhibits one of the largest ranges forthe light intensity minimum and maximum growthrates (Richardson et al. 1983). This suggests a lowercritical light level compared to dominant small phyto -plankton species. It also suggests that this diatommay maintain optimal photosynthetic rates in thedeep UML where variability in the light environ-ment is significantly high for phytoplankton (Lavaudet al. 2007). There are observations in other oceanicareas that indicate that the deep UML may create aniche for large diatoms thanks to the aforemen-tioned superior capacity to exploit fluctuating lightand low irradiance regimes (e.g. Goldman & Mc -Gillicuddy 2003, Thompson et al. 2007).

In conclusion, our observations suggest thatphytoplankton in the Bransfield Strait region do notreact as a whole to physical forcing; instead, differ-ent selective responses occurs depending on thephysical process involved. Small phytoplankton dis-tributions strongly correlate with the presence ofmesoscale structures. An asymmetric distributionwas observed across the 2 frontal regions. A deepcolumn-like structure was observed on the deepen-ing ZUML sides of the fronts indicating subduction. Alocal shallow maximum was observed on the othersides, coinciding with restratification and the ac -companying shoaling of the UML. Large phyto-plankton strongly correlate with the rate of verticalmixing with a preference for environments wherethis rate is high and thus dominant in the highly tur-bulent environment of the Antarctic Sound. On theother hand, in the Bransfield Strait, where meanvertical mixing is low, their proportion is very lowexcept in those stations where vertical mixing isclearly above the mean.

Acknowledgements. We express our gratitude to the techni-cal staff and crew of RV Hesprides for supporting our workat sea. This work was supported by the Spanish governmentthrough project COUPLING (CTM2008-06343-CO2-01).The authors are grateful to Celia Marras and FrancescPeters from the Institut de Cincies del Mar-CSIC for pro-viding useful suggestions and comments. C.G-M. was sup-ported by a predoctoral fellowship from the Spanish Councilfor Scientific Research (CSIC), JAE-Predoc 2009. B.M.-Cwas supported by the Ramn y Cajal program from theSpanish Ministry of Education and Science.

LITERATURE CITED

Behrenfeld MJ (2010) Abandoning Sverdrups CriticalDepth Hypothesis on phytoplankton blooms. Ecology 91: 977989

Brainerd K, Gregg M (1995) Surface mixed and mixing layerdepths. Deep-Sea Res I 42: 15211543

Chelton DB, Deszoeke RA, Schlax MG, Naggar K, Stwertz N(1998) Geographical variability of the first baroclinic Ross -by radius of deformation. J Phys Oceanogr 28: 433460

Chelton DB, Schlax MG, Samelson RM, de Szoeke RA(2007) Global observations of large oceanic eddies. Geo-phys Res Lett 34: L15606, doi:10.1029/2007GL030812

Cuttelod A, Herv C (2010) ALMOFRONT 2 cruise in Albo-ran sea: chlorophyll fluorescence calibration. J OceanogrResearch Data 3: 611

DAsaro E, Lee C, Rainville L, Harcourt R, Thomas LN (2011)Enhanced turbulence and energy dissipation at oceanfronts. Science 332: 318322

Elliott JA, Irish AE, Reynolds CS (2002) Predicting the spa-tial dominance of phytoplankton in a light limited andincompletely mixed eutrophic water column using thePROTECH model. Freshw Biol 47: 433440

Ferrari R, Boccaletti G (2004) Eddy-mixed layer interactionsin the ocean. Oceanography 17: 1221

Firing E (1991) Acoustic Doppler current profiler measure-ments and methods. In: WOCE (World Ocean CirculationExperiment) operations manual, Vol 3, Section 1, Part 3: WHP (WOCE Hydrographic Program) operations andmethods. WHP Office Report WHPO 91-1, WOCE ReportNo. 68/91, Woods Hole Oceanographic Institution, WoodsHole, MA

Franks PJS (2001) Turbulence avoidance: an alternativeexplanation of turbulence-enhanced ingestion rates inthe field. Limnol Oceanogr 46: 959963

Garca-Muoz C, Lubin LM, Garca CM, Marrero-Daz A,Vernet M, Sangr P (2013) A mesoscale study of phyto-plankton assemblages around the South ShetlandIslands. Polar Biol 36: 11071123

Gargett A, Garner T (2008) Determining Thorpe scales fromship-lowered CTD density profiles. J Atmos Ocean Tech-nol 25: 16571670

Goldman JC, McGillicuddy DJ Jr (2003) Effect of large marine diatoms growing at low light on episodic newproduction. Limnol Oceanogr 48: 11761182

Hewes CD, Reiss CS, Kahru M, Mitchell BG, Holm-HansenO (2008) Control of phytoplankton biomass by dilutionand mixing depth in the western Weddell-Scotia Conflu-ence. Mar Ecol Prog Ser 366: 1529

Hewes CD, Reiss CS, Holm-Hansen O (2009) A quantitativeanalysis of sources for summertime phytoplankton vari-ability over 18 years in the South Shetland Islands(Antarctica). Deep-Sea Res I 56: 12301241

Holm-Hansen O, Naganobu M, Kawaguchi S, Kameda Tand others (2004) Factors influencing the distribution,biomass, and productivity of phytoplankton in theScotia Sea and adjoining waters. Deep-Sea Res II 51: 13331350

Huisman J, Arrays M, Ebert U, Sommeijer B (2002) How dosinking phytoplankton species manage to persist? AmNat 159: 245254

Huisman J, Jonker RR, Zonneveld C, Weissing FJ (1999)Competition for light between phytoplankton species: experimental tests of mechanistic theory. Ecology 80: 211222

45

http://dx.doi.org/10.1890/0012-9658(1999)080[0211%3ACFLBPS]2.0.CO%3B2http://dx.doi.org/10.1086/338511http://dx.doi.org/10.1016/j.dsr.2009.01.010http://dx.doi.org/10.3354/meps07515http://dx.doi.org/10.4319/lo.2003.48.3.1176http://dx.doi.org/10.1175/2008JTECHO541.1http://dx.doi.org/10.1007/s00300-013-1333-5http://dx.doi.org/10.5670/oceanog.2004.26http://dx.doi.org/10.1046/j.1365-2427.2002.00813.xhttp://dx.doi.org/10.1126/science.1201515http://dx.doi.org/10.1029/2007GL030812http://dx.doi.org/10.1175/1520-0485(1998)028%3C0433%3AGVOTFB%3E2.0.CO%3B2http://dx.doi.org/10.1016/0967-0637(95)00068-Hhttp://dx.doi.org/10.1890/09-1207.1


Huisman J, Sharples J, Stroom JM, Visser PM, KardinaalWFA, Verspagen JMH, Sommeijer B (2004) Changes inturbulent mixing shift competition for competition forlight. Ecology 85: 29602970

Kang SH, Lee SH (1995) Antarctic phytoplankton assem-blage in the western Bransfield-Strait region, February1993: composition, biomass, and mesoscale distributions.Mar Ecol Prog Ser 129: 253267

Kang SH, Lee SH, Chung KH, Kim D, Park MG (2001)Antarctic phytoplankton assemblages in the marginal icezone of the northwestern Weddell Sea. J Plankton Res23: 333352

Kara AB, Rochford PA, Hurlburt HE (2000) An optimal defi-nition for ocean mixed layer depth. J Geophys Res 105: 1680316821

Kirboe T (2007) Small-scale turbulence, marine snow for-mation, and planktivorous feeding. Sci Mar 61 (Suppl 1): 148158

Klein P, Lapeyre G (2009) The oceanic vertical pump in -duced by mesoscale and submesocale turbulence. AnnuRev Mar Sci 1: 351375

Kopczynska EE, Ligowsky R (1985) Phytoplankton com -position and biomass distribution in the southern DrakePassage, the Bransfield Strait and the adjacent watersof the Weddell Sea in December 1983January 1984(BIOMASS-SIBEX). Pol Polar Res 6: 6567

Kropuenske LR, Mills MM, van Dijken GL, Baile S, Robin-son DH, Welschmeyer NA, Arrigo KR (2009) Photophysi-ology in two major Southern Ocean phytoplankton taxa: photoprotection in Phaeocystis antarctica and Fragilari-opsis cylindrus. Limnol Oceanogr 54: 11761196

Lavaud J, Strzepek RF, Kroth PG (2007) Photoprotectioncapacity differs among diatoms: possible consequenceson the spatial distribution of diatoms related to fluctua-tions in the underwater light climate. Limnol Oceanogr52: 11881194

Lozovatsky I, Figueroa M, Roget E, Fernando HJS, Shapo-valov S (2005) Observations and scaling of the uppermixed layer in the North Atlantic. J Geophys Res 110: C05013, doi: 10.1029/2004JC002708

Mann KH, Lazier JRN (1996) Dynamics of marine ecosys-tems: Biologicalphysical interactions in the oceans, 2ndedn. Blackwell Science, Oxford

Mendes CRB, de Souza MS, Garcia VMT, Leal MC, BrotasV, Garcia CAE (2012) Dynamics of phytoplankton com-munities during late summer around the tip of the Ant -arctic Peninsula. Deep-Sea Res I 65: 114

Mendes CRB, Tavano VM, Leal MC, de Souza MS, Brotas V(2013) Shifts in the dominance between diatoms andcryptophytes during three late summers in the BransfieldStrait (Antarctic Peninsula). Polar Biol 36:537547

Mitchell BG, Holm-Hansen O (1991) Observations and mod-eling of the Antarctic phytoplankton crop in relation tomixing depth. Deep-Sea Res A 38: 9811007

Nagai T, Tandon A, Rudnick DL (2006) Two-dimensionalageostrophic secondary circulation at ocean fronts due tovertical mixing and large-scale deformation. J GeophysRes 111: C09038, doi:10.1029/2005JC002964

Oakey N, Elliott J (1982) Dissipation within the surfacemixed layer. J Phys Oceanogr 12: 171185

Padman L, Fricker HA, Coleman R, Howard S, Erofeeva L(2002) A new tide model for the Antarctic ice shelves and

seas. Ann Glaciol 34: 247254Palls-Sanz E, Johnston TMS, Rudnick DL (2010) Frontal

dynamics in a California Current System shallow front: 2.Mesoscale vertical velocity. J Geophys Res 115: C12068,doi:10.1029/2010JC006474

Passarge J, Hol S, Escher M, Huisman J (2006) Competitionfor nutrients and light: stable coexistence, alternativestable states, or competitive exclusion? Ecol Monogr 76: 5772

Petersen JE, Sanford LP, Michael Kemp P (1998) Coastalplankton responses to turbulent mixing in experimentalecosystems. Mar Ecol Prog Ser 171: 2341

Pollard RT, Regier LA, (1992). Vorticity and vertical circula-tion at an Oceanic Front. J Phys Oceanogr 22: 609625.

Richardson K, Beardall J, Raven JA (1983) Adaptation ofunicellular algae to irradiances: analysis and strategies.New Phytol 93: 157191

Sangr P, Gordo C, Herndez-Arencibia M, Marrero-Daz Aand others (2011) The Bransfield current system. Deep-Sea Res I 58: 390402

Savidge DK, Amft JA (2009) Circulation on the West Antarc-tic Peninsula derived from 6 years of shipboard ADCPtransects. Deep-Sea Res I 56: 16331655

Sverdrup HS (1953) On conditions for vernal blooming ofphytoplankton. J Cons Int Explor Mer18: 287295

Teira E, Mourio-Carballido B, Martnez-Garca S, SobrinoC, Ameneiro J, Hernndez-Len S, Vzquez E (2012)Controls of primary production and bacterial carbonmeta bolism around South Shetland Islands. Deep-SeaRes I 69: 7081

Thomas LN, Joyce TM (2010) Subduction on the northernand southern Flanks if the Gulf Stream. J Phys Oceanogr40: 429438

Thomas LN, Tandon A, Mahadevan A (2008) Submesoscaleprocesses and dynamics. In: Hecht MV, Hasumi H (eds)Ocean modeling in an eddying regime. GeophysicalMonograph Series Vol 177, American GeophysicalUnion, Washington, DC, p 1738

Thompson PA, Pesant S, Waite AM (2007) Contrasting thevertical differences in the phytoplankton biology of adipole pair of eddies in the south-eastern Indian Ocean.Deep-Sea Res II 54: 10031028

Thorpe SA (1977) Turbulence and mixing in a Scottish loch.Philos Trans R Soc Lond A 286: 125181

UNESCO (1994) Protocols for the Joint Global Ocean FluxStudy (JGOFS) core measurements. IOC Manuals andGuides 29, UNESCO, Paris

Vaillancourt RD, Sambrotto RN, Green S, Matsuda A (2003)Phytoplankton biomass and photosynthetic competencyin the summertime Mertz Glacier region of East Antarc-tica. Deep-Sea Res II 50: 14151440

Van de Poll WH, Lagunas M, de Vries T, Visser RJW, BumaAGJ (2011) Non-photochemical quenching of chloro-phyll fluorescence and xanthophyll cycle responses afterexcess PAR and UVR in Chaetoceros brevis, Phaeocystisantarctica and coastal Antarctic phytoplankton. Mar EcolProg Ser 426: 119131

Weissing FJ, Huisman J (1994) Growth and competition in alight gradient. J Theor Biol 168: 323336

Wolk F, Yamazaki H, Laurent S, Lueck RG (2002) A newfree-fall profiler for measuring biophysical microstruc-ture. J Atmos Ocean Technol 19: 780793

46

Editorial responsibility: Katherine Richardson, Copenhagen, Denmark

Submitted: May 23, 2013; Accepted: November 25, 2013Proofs received from author(s): February 12, 2014

http://dx.doi.org/10.1175/1520-0426(2002)019%3C0780%3AANFFPF%3E2.0.CO%3B2http://dx.doi.org/10.1006/jtbi.1994.1113http://dx.doi.org/10.3354/meps09000http://dx.doi.org/10.1016/S0967-0645(03)00077-8http://dx.doi.org/10.1016/j.dsr2.2006.12.009http://dx.doi.org/10.1175/2009JPO4187.1http://dx.doi.org/10.1016/j.dsr.2012.07.002http://dx.doi.org/10.1093/icesjms/18.3.287http://dx.doi.org/10.1016/j.dsr.2009.05.011http://dx.doi.org/10.1016/j.dsr.2011.01.011http://dx.doi.org/10.1111/j.1469-8137.1983.tb03422.xhttp://dx.doi.org/10.3354/meps171023http://dx.doi.org/10.1890/04-1824http://dx.doi.org/10.1029/2010JC006474http://dx.doi.org/10.3189/172756402781817752http://dx.doi.org/10.1175/1520-0485(1982)012%3C0171%3ADWTSML%3E2.0.CO%3B2http://dx.doi.org/10.1016/0198-0149(91)90093-Uhttp://dx.doi.org/10.1016/j.dsr.2012.03.002http://dx.doi.org/10.1029/2004JC002708http://dx.doi.org/10.4319/lo.2007.52.3.1188http://dx.doi.org/10.4319/lo.2009.54.4.1176http://dx.doi.org/10.1146/annurev.marine.010908.163704http://dx.doi.org/10.1093/plankt/23.4.333http://dx.doi.org/10.3354/meps129253http://dx.doi.org/10.1890/03-0763

cite43: cite28: cite5: cite14: cite42: cite3: cite27: cite13: cite41: cite39: cite12: cite40: cite25: cite38: cite11: cite24: cite37: cite10: cite8: cite23: cite36: cite6: cite22: cite35: cite4: cite21: cite34: cite19: cite20: cite33: cite18: cite32: cite17: cite31: cite9: cite29: cite30: cite15:

coupling between upper ocean layer variability and size

Documents