MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser

Vol. 491: 15–31, 2013doi: 10.3354/meps10461

Published October 2

INTRODUCTION

Global biogeochemical as well as trophic webmodels require information on phytoplankton bio-mass and on the role of the different groups that com-pose it in nature. However, obtaining informationon phytoplankton diversity using classic taxonomic

methods requires the dedicated work of an expertfor an entire day or more for each sample. There-fore, although microscopic analysis remains themost trustworthy way to determine phytoplanktontaxonomic composition, several alternative tech-niques have been developed that classify phyto-plankton into groups, commonly known as phyto-

© Inter-Research 2013 · www.int-res.com*Email: vsegura@inidep.edu.ar

Phytoplankton types and primary production in the Argentine Sea

Valeria Segura1,*, Vivian A. Lutz1,2, Ana Dogliotti3,4, Ricardo I. Silva1, Ruben M. Negri1, Rut Akselman1, Hugo Benavides1

1Instituto Nacional de Investigación y Desarrollo Pesquero (INIDEP), Paseo Victoria Ocampo No. 1, B7602HSA, Mar del Plata, Argentina

2Instituto de Investigaciones Marinas y Costeras, Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Dean Funes 3350, B7602AYL, Mar del Plata, Argentina

3Instituto de Astronomía y Física del Espacio (IAFE), CONICET/UBA, CC 67, Suc. 28, 1428 Buenos Aires, Argentina4Instituto Franco-Argentino para el Estudio del Clima y sus Impactos (UMI IFAECI/CNRS-CONICET-UBA),

Ciudad Universitaria Pabellón II Piso 2, C1428EHA, Ciudad Autónoma de Buenos Aires, Argentina

ABSTRACT: Several satellite models classify phytoplankton functional types (PFT) based on cellsize. In this study we used field data from the Argentine Sea on both the photosynthetic and thebio-optical properties of phytoplankton to distinguish photosynthetic and bio-optical phytoplank-ton types (PBPT). Cluster analyses were run using data from 70 stations sampled during 3 periodsto distinguish different PBPT, and principal component analysis was used to describe them. Weexamined the main taxonomic composition and percentage of chl a in the

Mar Ecol Prog Ser 491: 15–31, 2013

plankton functional types (PFT). Phytoplankton in aPFT share common ecological roles in nature but donot necessarily have the same phylogenetic origin(Nair et al. 2008). The PFT can be distinguishedaccording to different criteria. Le Quéré et al. (2005)defined functional types of plankton (phyto- and zooplankton) using a combination of different prop-erties, including their biogeochemical role (e.g. N2 fixers, calcifiers, DMS producers, silicifiers), theirnutritional requirement, and the behavior of onegroup with respect to others.

There are models that differentiate phytoplanktontypes at large scale (regional and global), from satel-lite data. Some of these models use algorithms thatallow differentiation between individual taxonomicgroups, such as coccolithophores (Brown & Podestá1997), the cyanobacteria Trichodesmium spp. (Sub-ramaniam et al. 1999), and diatoms (Sathy endranathet al. 2004), based on their unique optical character-istics. Recently, models have been developed thatattempt to describe the distribution of all phytoplank-ton forms, grouping them into different PFT at bothregional and global scales. In these models, the crite-rion to distinguish phytoplankton types is usuallyrelated to cell size (micro-, nano- and picophyto-plankton). The principle is based on differentiating,indirectly, cell size through the composition of phyto-plankton pigments (by HPLC) and associating thesesize fractions to some particular bio-optical charac-teristic that can be derived from satellite estimations,such as radiance, phytoplankton absorption coeffi-cient and chl a concentration (Alvain et al. 2005, Uitzet al. 2006, Hirata et al. 2008). Nevertheless, withinthe same size fraction (e.g. nanoplankton), therecould be phytoplankton with a diversity of biogeo-chemical functions, such as DMS producers and cal-cifiers (Nair et al. 2008). Some remote sensing studies(e.g. Raitsos et al. 2008) have attempted to distin-guish different phytoplankton functional groupswithin the same size class (for example, diatoms anddinoflagellates). Some of these models have showna good agreement between satellite and field estimates for some regions such as the southernBenguela ecosystem (Aiken et al. 2007). However,many regions of the world ocean, including theArgentine Sea, have no field data for validation. Theimportance of obtaining more in situ observations toadvance these models has been highlighted by manyauthors (e.g. Brewin et al. 2011).

Photosynthetic and bio-optical properties changeaccording to the composition, abundance and sizestructure of the phytoplanktonic communities, andwith the environmental conditions (e.g. light, tem-

perature, availability of nutrients). For example, val-ues for the photosynthetic parameters αB (initial slopeof production versus irradiance at low values, nor-malized by chl a) and PBm (maximum production atlight saturation, normalized by chl a) usually showspatial and seasonal variations (Platt et al. 1992). Aninverse relationship was found between cell size indiatoms and the value of PBm by Geider et al. (1986)and Bouman et al. (2005), though this may not alwaysbe the case. Several studies have found that smallcells (picophytoplankton) have higher values forphotosynthetic parameters than large cells (e.g. Plattet al. 1983, Geider et al. 1986). Variations in cell sizeand pigment composition in different species, as wellas in their photophysiological state, influence theshapes of the absorption spectra of phytoplankton.The specific absorption coefficient at wavelength (λ)of phytoplankton (absorption per unit chl a), aBph(λ), isknown to decrease with an increase in cell size andintracellular pigment concentration (Sathyendranathet al. 1987). The ratios of absorption at differentwave lengths, especially aph(440)/aph(676), aph(490)/aph(555) and aph(443)/aph(555), have been used as anindication of dominant phytoplankton groups (Stuartet al. 2000). The ratio Chla/Fl (chl a concentrationversus in vivo fluorescence) has also been related tothe phytoplankton composition and its physiologicalstate (Kiefer 1973).

The influence of environmental conditions on pho-tosynthetic parameters has been extensively docu-mented, particularly that of the light field (Falkowski1980), and that of temperature in the case of PBm(Bouman et al. 2005), as well as a dependency of αB

on nitrate concentrations (Platt et al. 1992).The southwestern Atlantic Ocean, and the Argen-

tine Sea in particular, are among the richest chl aareas globally according to ocean color satelliteimages (e.g. Gregg et al. 2005, Rivas et al. 2006).Recent field work has shown wide ranges in primaryproduction during spring (Schloss et al. 2007, Garciaet al. 2008, Lutz et al. 2010). Phytoplankton in thearea show a high degree of heterogeneity in theirdistribution, not only in biomass, but also in commu-nity structure (e.g. Negri et al. 1992). Maxima of chl ahave been associated with different taxonomicgroups: diatoms and dinoflagellates in the north(above 41° S; Carreto et al. 1981); diatom blooms inthe Argentine shelf break (Lutz & Carreto 1991, Gar-cia et al. 2008), including one of Thalassiosira cf.oceanica (Sabatini et al. 2012); the dinoflagellate Pro-rocentrum minimum in Bahia Grande ~51° S (Gómezet al. 2011, Sabatini et al. 2012); and Synechococ -cus sp. (picoplankton fraction) during the summer

16

Segura et al.: Phytoplankton types in the Argentine Sea

months at the coastal time-series station EPEA (38°28’ S, 57° 41’ W; Silva et al. 2009).

Studies describing phytoplankton according totheir function, PFT, are just beginning in the Argen-tine Sea. Moreno et al. (2012) described PFT in 27stations in this area during the end of summer/begin-ning of fall using pigment composition to infer cellsize. They found that diatoms dominated the micro-phytoplankton south of the Argentine shelf break,dinoflagellates occurred close to Peninsula Valdés,nanophytoplankton was abundant in the middleshelf and close to Mar del Plata, and picophytoplank-ton were found in frontal areas in the south. Using asatellite model, Omachi et al. (2010) showed a widespatial heterogeneity in PFT prescribed according tocell size (micro-, nano- and picophytoplankton) indifferent zones of the Argentine Sea from the coast(up to 50 m in depth), middle shelf (up to 200 m) andthe shelf break (beyond 200 m).

Most of the studies assigned PFT based on the cri-terion of cell size, rather than on a given function asdefined by Le Quéré et al. (2005). In this study, weconsidered the following aspects in the distinction ofphytoplankton types: (1) that there is an interest instudying the distribution of phytoplankton typesusing bio-optical remote sensing information, (2) thatsatellite models of primary production require theassignation of photosynthetic parameters (which canonly be estimated by field work), and (3) that a rela-tionship between bio-optical and photosyntheticcharacteristics could be expected. Thus, our mainobjective was to distinguish photosynthetic and bio-optical phytoplankton types (PBPT) using bio-opticaland photosynthetic field measurements in the Argen -tine Sea. The main taxonomic composition and thepercentage of concentration of chl a

Mar Ecol Prog Ser 491: 15–31, 2013

the photosynthetic parameters α and Pm. These pa-rameters were normalized by Chla to obtain αB (mg C(mg chl a)−1 h−1 (W m−2)−1) and PBm (mg C (mg chl a)−1

h−1). Hourly production at the surface, p0, and dailywater-column-integrated primary production, PZ,T,were calculated at each station; for the latter, it wasassumed that the photosynthetic parameters wouldremain constant with depth and throughout the day.There are some caveats to these assumptions, e.g. it isknown that photosynthetic parameters may changeat different depths according to the mixing regime(influencing species distribution and the physio -logical state of the cells because of changes in lightand nutrients; e.g. Platt & Gallegos 1980).

Identification and quantification of surface phyto-plankton. Seawater samples were preserved withdifferent fixing solutions: neutralized formaldehyde(0.4% final concentration) for nano phytoplankton(5 to 20 µm) and ultraphytoplankton (20 µm) (Throndsen 1978). Identification and quan-tification were carried out using the sedimentationtechnique and an inverted microscope (Hasle 1978,Tomas 1997). The identification of taxonomic groupswas not complete; the abundance of cyanobacteriawas not analyzed (since epifluorescence microscopydata were not available), and for GEF-3, only partialinformation on microphytoplankton was available.

Other environmental measurements. Irradiance atthe surface, E0, was measured continuously with a LI-COR cosine detector. Noon irradiance at the surfacefor each station, En, was estimated as the average ofall E0 recorded within 2 h around local noon andevery 2° latitude (Lutz et al. 2010). The downwellingattenuation coefficient of light, Kd(PAR), was com-puted using the model of Sathyendranath & Platt(1988). The inputs for the model were the absorptionand scattering spectra of pure seawater (Pope & Fry1997), the values of Chla and absorption coefficientsobtained for discrete depths, and the correspondingaverage E0. Using these values of Kd(PAR) per meter,the irradiance profiles at different times were calcu-lated. The depth of the euphotic zone, Zeu, was takenas the depth at which irradiance, EZ, was equal to 1%E0. Wind velocity, Wi, was measured with an on-board meteorological station. The depth of the mixedlayer, MLD, was taken as the depth where the differ-ence between the density and the surface densitywas >0.05 kg m−3 (Brainerd & Gregg 1995).

Data analysis

Distinction of PBPT. A total of 70 stations with acomplete bio-optical and photosynthetic dataset(Fig. 1) from the 3 cruises (22 from GEF-1, 25 from

18

1

7 9 11

131517192123

35 36 37

384042

44 46 48 50

515355

70° 68° 66° 64° 62° 60° 58° 56°W

200 m

4 678

1013

33 35 37

38434749

5052

57

58 5962

65

6971

70° 68° 66° 64° 62° 60° 58° 56°56°

54°

52°

50°

48°

46°

44°

42°

40°

38°S GEF-1 (Spring 2005)

200 m

17 9

101315

2025273133

3437

3941 42

43454849

52 55

626466

70° 68° 66° 64° 62° 60° 58° 56°

GEF-2 (Summer 2006)

200 m

GEF-3 (Winter 2006)

Grande Bay

Peninsula Valdés

Grande Bay Grande Bay

Peninsula Valdés Peninsula Valdés

San Jorge Gulf

San Jorge Gulf

San Jorge Gulf 100 m

50 m50 m

100 m

100 m

50 m

Chla + Abs + PP

Chla

Fig. 1. Location of the stations in the 3 cruises (GEF-1, GEF-2 and GEF-3). Symbols indicate collection of samples for mainanalyses at the stations. Chla represents points where only surface chl a was measured, and Chla + Abs (Absorption) + PP

(field primary production) represent the primary production stations

Segura et al.: Phytoplankton types in the Argentine Sea

GEF-2 and 23 from GEF-3) were used to distinguishPBPT. The photosynthetic parameters used were αB,PBm and maximum quantum yield (φmax); and withinthe bio-optical variables, we chose the following:spectrally averaged specific absorption coefficient( ), aBph(440), aBph(676), aph(440)/ aph(676), aph(490)/aph(555), aph(443)/aph(555), total Chla, and the ratioChla/Fl . A matrix with the selected standardizedproperties from all stations in the 3 cruises was built,and a cluster analysis following Ward’s method as acriterion to specify the internally most homogeneousgroup and the square Euclidean distance as a meas-ure of dissimilitude were run using the STATISTICA7 software package. When the distance of associationamong stations was

Mar Ecol Prog Ser 491: 15–31, 2013

Relationship between PBPT and environmentalconditions. The available environmental propertiesincluded basic physical (temperature, salinity, windspeed, total depth, depth of the mixed layer, and inci-dent irradiance) and underwater bio-optical (depthof the euphotic layer, detritus absorption, attenuationcoefficient) estimations (Table 3). These were used tocharacterize the PBPT according to environmentalconditions using PCA.

Satellite estimation of phytoplankton size classes.The model to determine phytoplankton size classesfrom satellite ocean color developed by Hirata et al.(2008) was applied to moderate resolution imagingspectroradiometer (MODIS)-Aqua satellite compos-

ite images of aph(443) using the quasi-analytical algo-rithm (Lee et al. 2002) for each cruise. MODIS Level1A and ancillary data from each cruise period weredownloaded from the NASA Web site (http://oceancolor. gsfc. nasa. gov/), processed using SeaDASv6.2 (2009 reprocessing) and mapped to a cylindricalequidistant projection with a spatial resolution of1 min (~2 km). Maps of phytoplankton size classeswere obtained using aph(443) thresholds to detect thedominance of micro- (aph(443) > 0.069 m−1), nano-(0.023 < aph(443) ≤ 0.069 m−1) and picophytoplankton(aph(443) ≤ 0.023 m−1).

RESULTS

Chl a and primary production distribution

Chl a concentrations showed a high degree of vari-ation in the 3 periods analyzed (Fig. 2A), with maxi-mum values recorded in spring (24.01 mg m−3 atStn 73, GEF-1, in Grande Bay, with no primary pro-duction data) and minimum values recorded at theend of winter (0.42 mg m−3 at Stn 51, GEF-3). Markedblooms were observed in spring, while at the end ofsummer, these were of minor intensity (Fig. 2A). Atthe end of winter, Chla values were low in the southand high in the north, indicating the beginning ofspring (GEF-3, Fig. 2A). The concentration of chl a inthe

Segura et al.: Phytoplankton types in the Argentine Sea 21

56°

54°

52°

50°

48°

46°

44°

42°

40°

38°

200 m

GEF-1 (Spring 2005)

0.000.090.140.190.400.650.902.102.603.506.008.2012.0017.0022.00

San Jorge Gulf

Peninsula Valdés

100 m

50 m

GEF-2 (Summer 2006)

Peninsula Valdés

San JorgeGulf

100 m

50 m

200m

A

200 m

GEF-3 (Winter 2006)

Peninsula Valdés

San JorgeGulf

100 m

50 m

mg Chl a m-3

GEF-3 (Winter 2006)

200 m

Peninsula Valdés

San JorgeGulf

100 m

50 m

56°

54°

52°

50°

48°

46°

44°

42°

40°

38°

0

25

50

75

100

200 m

100 m

50 m

GEF-2 (Summer 2006)

Peninsula Valdés

San JorgeGulf

200 m

GEF-1 (Spring 2005)

Peninsula Valdés

San JorgeGulf

B

100 m

50 m

4678

1013

3335 37

38434749

5052

57

585962

65

6971

70° 68° 66° 64° 62° 60° 58° 56° 54°56°

54°

52°

50°

48°

46°

44°

42°

40°

38°

200 m

C

100 m

50 m

52 51 60-150150-300300-600

900-1500

0-60

600-900

1500-2500

2500-3000

3000-5500

GEF-1 (Spring 2005)

mg C m-2 d-1

17

1315

2025273133

3437

3942

43454849

52 55

626466

69° 67° 65° 63° 61° 59° 57° 55°

200 m

100 m

50 m

GEF-2 (Summer 2006)1

7 9 11

1315171921

35 36 37

384042

46 48 50

5355

69° 67° 65° 63° 61° 59° 57° 55°

200 m

100 m

50 m

GEF-3 (Winter 2006)

Peninsula Valdés Peninsula Valdés Peninsula Valdés

GulfSan Jorge

GulfSan Jorge

GulfSan Jorge

4441

10

23

51

Fig. 2. Distribution of (A) total surface chl a, (B) percentage of surface chl a in the

Mar Ecol Prog Ser 491: 15–31, 2013

Variations in surface bio-optical and photosynthetic properties

High variations in all properties were observedduring the 3 cruises (Table 2), e.g. PBm varied between0.68 and 10.05 mg C (mg chl a)−1 h−1, and aBph(440)varied between 0.015 and 0.067 m2 (mg chl a)−1. Sig-nificant correlations were found between aph(440)/aph(676) and Chla (rS = −0.37, p < 0.05; Fig. 4A),between aBph(440) and Chla

Segura et al.: Phytoplankton types in the Argentine Sea 23

0 10 20 30 40 80 90 100(Dlink/Dmax) x 100

36+15+11+53+42**33**7**

59*58*42+35+46+38+13+38*7+

37+21+19+1+

37**48**37*23+49**33*55**10*

8*35*

7*50*69*62*57*52*49*65*47*

6*34**66**62**64**43**10**71*52**41**39**31**27**15**

9**1**

20**55+45**51+25**13**44+40+

9+13*43*17+48+50+4*

50- 70

PBPT-1

PBPT-2

PBPT-3

PBPT-4

PBPT-5

PBPT-6

PBPT-7

PBPT-8

PBPT-10

PBPT-9

PBPT-11

* : GEF-1

**: GEF-2

+ : GEF-3

Fig. 5. Cluster analysis to distinguish photosynthetic and bio-optical phytoplankton types (PBPT) based on photosynthetic andbio-optical properties including all stations from the 3 cruises. The dashed line represents the threshold line of the cluster

marked at 10% of the total distance to distinguish the PBPT

PBPT1 PBPT2 PBPT3 PBPT4 PBPT5 PBPT6 PBPT7 PBPT8 PBPT9 PBPT10 PBPT11

n(stations): 4 1 18 7 8 1 2 20 4 2 3

Station GEF-1: GEF-1: GEF-1: GEF-1: GEF-1: GEF-1: GEF-1: GEF-1: GEF-1: GEF-2: GEF-3: numbers: 4 43 13 71 6,47,49,52, 50 7,35 8,10,33, 58,59 42 11,15,36 GEF-3: GEF-2: GEF-2: 57,62,65,69 37,38 GEF-2: GEF-3:

17,48,50 1,9,13,15, 10,34,43, GEF-2: 7,33 53 20,25,27,31, 62,64,66 37,48,49,55 39,41,45,52 GEF-3:

GEF-3: 1,7,13,19,21,23,

9,40,44,51,55 35,37,38,42,46 PBm 1.20 1.24 1.87 2.69 6.21 6.71 1.87 1.38 3.86 1.51 0.96αB 0.09 0.06 0.07 0.10 0.38 0.84 0.10 0.07 0.09 0.07 0.05φmax 0.004 0.005 0.012 0.005 0.019 0.042 0.005 0.004 0.004 0.004 0.002p0 1.16 1.14 2.26 7.24 9.87 8.51 32.31 2.69 4.68 4.34 1.02Chla 0.83 0.92 1.18 1.72 0.99 1.26 15.00 1.77 1.25 3.38 1.06Chla/Fl 2.00 0.64 1.07 0.71 0.58 0.63 2.37 1.67 1.13 1.08 7.00

0.012 0.013 0.016 0.022 0.011 0.010 0.008 0.010 0.007 0.013 0.010

aBph(440) 0.039 0.046 0.046 0.061 0.035 0.028 0.019 0.027 0.018 0.032 0.029

aBph(676) 0.015 0.009 0.016 0.022 0.012 0.016 0.011 0.012 0.008 0.023 0.012aph(440)/aph(676) 2.52 5.25 2.91 2.85 2.75 1.78 1.73 2.21 2.34 1.40 2.37aph(490)/aph(555) 9.97 8.88 4.16 4.37 4.18 5.62 2.72 4.26 2.72 2.79 5.73aph(443)/aph(555) 16.43 17.28 6.33 6.22 6.79 8.53 4.16 6.64 4.19 4.15 9.51

aphB

Table 4. Mean values of the properties analyzed in each photosynthetic and bio-optical phytoplankton type (PBPT). Bold num-bers represent properties with coefficients of variation within each PBPT

Mar Ecol Prog Ser 491: 15–31, 2013

shapes of the absorption spectra at the different sta-tions within a PBPT were usually similar, but somePBPT had strong differences in the blue region, suchas in PBPT5 and PBPT10 (Fig. 8).

The available information on taxonomic composi-tion within the PBPT showed a high heterogeneity

(Table 6). For example, PBPT1 and PBPT6, with%Chla

Segura et al.: Phytoplankton types in the Argentine Sea

mainly associated with bio-optical variables [Zeu,Kd(PAR) and ad(440)], while PC2 was mainly associ-ated with physical variables (Sal, Z, SST and En). PC3and PC4 explain only approximately 12% of the vari-ation (Table 7). The projection of the different sta-tions in relation to the PCs showed that some PBPTsuch as PBPT3 and PBPT7 were characterized byhigh values of Kd(PAR) and ad(440) and low values ofZeu (PC1) (Fig. 9B,D). The rest of the PBPT did notshow a clear association with a given PC.

Distribution of satellite phytoplankton size classesand field PBPT. The model of the distribution ofphytoplankton types according to their size (Hirata etal. 2008) follows the satellite chl a distribution ob -

served in the mean images for the cruises (data notshown). Overlapping the PBPT distinguished here onthis satellite classification, we can see that for a sin-gle phytoplankton size class, there are more than 1PBPT and not always of the assumed size class; e.g.in the ‘micro class’ during spring, there are 3 PBPT(PBPT2, PBPT7 and PBPT8) which have a taxonomiccomposition that includes the nano size class as well.By contrast, PBPT3 was found prevalently in the‘nano size’ but also in an area marked as ‘micro’ andin another marked as ‘pico’ (Fig. 10). Note, however,that our description of prevalent cell sizes for fielddata for the

Mar Ecol Prog Ser 491: 15–31, 2013

DISCUSSION

Oceanographic conditions in the study area

The large area of the Argentine shelf studiedhere — encompassing 18° of latitude (from temper-ate to subantarctic waters) and >1000 m differencein total depth (from the coast to the shelf break)during 3 seasons (spring, summer and winter) —showed a highly heterogeneous environment. Froma biological point of view, the wide ranges in chl aconcentration in these waters were caused by theoccurrence of a large variety of different phyto-plankton taxa. Among the physical and biologicalfactors examined in this study, the most variablewere the bio-optical properties (i.e. euphotic depth,attenuation coefficient and detritus absorption),while the main hydrographic measurements (e.g.SST and Sal) were more consistent (Table 7). Therich diversity in the phytoplankton communitiesfound in the Argentine Sea could be attributed tothe environmental heterogeneity, as shown inTable 3.

Chl a and primary production

The variability in the photosynthetic and bio-optical properties of phytoplankton producedlarge variations in the hourly and integrated primary production (Table 2). The PZ,T valuesreported here, considering the caveats of extra -polation of surface parameters, provide a firstapproximation of integrated production in differ-ent areas of the Argentine Sea. The positive andsignificant correlation found between surfaceChla and p0 for all stations in the 3 cruises (rS =0.61, p < 0.05) coincided with that in spring (Lutzet al. 2010). There were also significant correla-tions between Chla and PZ,T (rS = 0.59, p < 0.05).Nevertheless, if one considers a linear directrelation between surface Chla and PZ,T, the dis-persion of the data is high; in some cases, theproduction is much lower (e.g. in GEF-1 Stn 35,in GEF-2 Stn 55 and in GEF-3 Stns 7 and 53) andin other cases is much higher (e.g. in GEF-1Stns 8, 49, 71, 50 and 52) than that predictedfrom Chla (Fig. 3).

26

PBPT1

0.00

0.02

0.04

0.06

0.08PBPT2

PBPT9

400 500 600 7000.00

0.02

0.04

0.06

0.08

PBPT8

400 500 600 700

400 500 600 700PBPT11

PBPT7

PBPT3

Wavelengths (nm)

400 500 600 700

PBPT10

PBPT4

aB ph

(m2

(mg

Chl

a)–1

)

0.00

0.02

0.04

0.06

0.08PBPT5 PBPT6

Fig. 8. Shapes of the specific ab sorption spectra of phytoplankton (normalized by chl a) of all the samples within each photo-synthetic and bio-optical phytoplankton type (PBPT)

Segura et al.: Phytoplankton types in the Argentine Sea

Phytoplankton types

The combination of the diverse taxa found andtheir physiological flexibility, according to our clusteranalysis, revealed 11 PBPT which were associatedwith different photosynthetic and bio-optical proper-ties (PCA, Fig. 7). This number of PBPT is higher thanwhat is usually modeled by remote sensing accord-ing to cell size (pico-, nano- and microphytoplank-ton). Nevertheless, this grouping of the phytoplank-ton populations into 11 PBPT represents an importantsimplification with respect to the large amount ofphytoplankton species present in the region; morethan 330 species have been described for dinoflagel-lates alone (Balech 1988).

Several studies have found a relationship betweenthe bio-optical and physiological properties of phyto-plankton and their cell size, e.g. Hoepffner &Sathyendranath (1992) showed that the highest val-ues of aBph(λ) corresponded to small cells, andBouman et al. (2005) showed that low values of PBmcorresponded to microphytoplankton. Some clearcases of these relationships were also found here,e.g. PBPT1 formed mainly by small cells showed highspecific absorption coefficients. Regarding ratios ofabsorption at different wavelengths (aph(440)/aph(676) and aph(490)/aph(555)), PBPT3 (which hadPrymnesiophyceae) had values close to thosedescribed for prymnesiophytes, and PBPT7 (com-posed of a bloom of diatoms) had values close tothose reported for diatoms by Stuart et al. (2000).However, in other cases, variations in the photosyn-thetic and bio-optical properties did not always fol-low the expected trend according to the cell size. Forexample, PBPT5, with %Chla

Mar Ecol Prog Ser 491: 15–31, 2013

rial Pacific, the microphytoplankton showed highervalues of PBm than small-sized cells.

The high variation in photosynthetic parameterswithin the PBPT indicates that, at least for this dataset, the relationship between bio-optical and photo-synthetic characteristics was not strong. Therefore,the assignation of representative photosyntheticparameters for different phytoplankton types tomodel primary production in this region is quite difficult.

Some of the PBPT could be related to some partic-ular environmental variables according to the PCA(e.g. PBPT7 and PBPT3 were associated with highvalues of Kd(PAR) and ad(440)), although in most

cases, there was no clear pattern (Fig. 9). This wouldpreclude the use of models of distribution of phyto-plankton types according to physical characteristicsof the environment, as has been proposed by Raitsoset al. (2008).

Comparison between satellite and field phytoplankton types

Satellite and field distributions of phytoplanktontypes are difficult to compare with any accuracysince they are not defined by the same principles. Inthis case, we chose to distinguish our types using

28

1

2

34

5

6

78

9

-1.0 -0.5 0.0 0.5 1.0

PC 3 : 12.17%

-1.0

-0.5

0.0

0.5

1.0

PC

4 :

11.4

7%

1-Z 2-Zeu3-Kd(PAR)4-MLD5-SST6-Sal7-En8-ad (440) 9-Wi

1

23

4

5

6

7

8 9

-1.0 -0.5 0.0 0.5 1.0

PC 1 : 28.95%

-1.0

-0.5

0.0

0.5

1.0

PC

2 :

24.1

1%

PC3

PC

4

-4 -3 -2 -1 0 1 2 3 4-3

-2

-1

0

1

2

3

4

5

A

C D

PC1P

C2

-4 -3 -2 -1 0 1 2-4

-3

-2

-1

0

1

2

3

PBPT1PBPT2PBPT3PBPT4PBPT5PBPT6PBPT7PBPT8PBPT9PBPT10PBPT11

B

Fig. 9. Graphical representation of principal component analysis for environmental variables. Correlation circle in the planeby (A) the 2 first principal components, PC1 vs. PC2, and (C) PC3 vs. PC4. Projection of the stations composing the photosyn-

thetic and bio-optical phytoplankton types (PBPT) for (B) PC1 vs. PC2 and (D) PC3 vs. PC4

Segura et al.: Phytoplankton types in the Argentine Sea

both photosynthetic and bio-optical characteristics.By way of comparison, the satellite model of Hirata etal. (2008) (one of the few that includes continentalmargins) is based only on phytoplankton cell size.Although there are some theoretical bases relatingthese properties (e.g. Aiken et al. 2007), these rela-tionships do not seem to hold for the Argentine Seabecause of the heterogeneity in environmental con-ditions and the wealth of phytoplankton taxonomicgroups present. This complexity resulted in the dis-tribution of more than 1 PBPT for a given size class(Fig. 10). A few field studies in well-characterizedhydrographic sites have been able to represent themain groups of phytoplankton present using pigmentcomposition alone (Carreto et al. 2003, Moreno et al.2012). Nevertheless, we found that it is not easy tomodel the distribution of phytoplankton types usingremote sensing in the whole Argentine Sea. This isdue to the diversity of taxa present, from ultraphyto-plankton (coccal, prymnesiophytes, diatoms) tomicrophytoplankton (dinoflagellates, diatoms), andtheir physiological flexibility, resulting in a high vari-ability in bio-optical and photosynthetic properties,e.g. aBph(440) varied between 0.015 and 0.067 m2 (mgchl a)−1, and PBm varied between 0.68 and 10.05 mg C(mg chl a)−1 h−1.

Concluding remarks

The lack of a clear relationship between bio-opticalvariables and cell size, as found in this study, sug-gests that it is not always possible to model phyto-plankton functional types using remote sensingbased on considerations of cell size alone. Satellitemodels that identify phytoplankton types in this way(e.g. Alvain et al. 2005, Hirata et al. 2008) may notalways be appropriate for highly heterogeneousregions such as the Argentine Sea.

In most of the PBPT, photosynthetic propertiesshowed a higher degree of variability than bio-opti-cal properties. Hence, it would be difficult to assignglobal photosynthetic parameters based on bio-opti-cal characteristics derived from satellite models.Although we used data from extensive cruises during3 seasons, this may nevertheless represent a some-what limited number of data points to describe thevariability found in the Argentine Sea. Additionalfield studies aimed at measuring photosynthetic ratesshould be performed in this zone of the southwesternAtlantic if we want to develop better biogeochemicalmodels for the region that capture both the bio-opti-cal as well as the physiological characteristics of thishighly productive area.

29

Fig. 10. Distribution of the phytoplankton size classes (pico-, nano- and micro-phytoplankton), modelled according to Hirata et al. (2008), for the different peri-ods analyzed based on satellite images from MODIS. Overlapped on the mapsare the distributions of the photosynthetic and bio-optical phytoplankton types

distinguished in this study (see symbols in Fig. 6)

Mar Ecol Prog Ser 491: 15–31, 2013

Acknowledgements. We are grateful to D. Cucchi Colleoni,D. Hernandéz, R. Reta, M. Sabatini, M. Carignan, N. Mon-toya, Y. Collos, C. Campagna and J. Carreto for providinginstruments and useful advice to develop this work; A.Bianchi and A. Piola for providing hydrographic data; andNASA and CONAE for satellite data. This work wasfinanced by INIDEP and grants from Fundación Antorchas(13900-12), PNUDARG02/018 (GEF −BB46, −BB12, −BB61),CONICET-PIP11220090100698 and CONICET-ANPCyT-PICT-08-0146. We thank V. Stuart for the final revision ofthe text. The manuscript highly benefited from suggestionsby 3 anonymous reviewers. This research is part of the doc-toral thesis work of V.S. This is INIDEP contribution # 1802.

LITERATURE CITED

Aiken J, Fishwick JR, Lavender SJ, Barlow RG and others(2007) Validation of MERIS reflectance and chlorophyllduring the BENCAL cruise October 2002: preliminaryvalidation of new demonstration products for phyto-plankton functional types and photosynthetic parame-ters. Int J Remote Sens 28: 497−516

Alvain S, Moulin C, Dandonneau Y, Bréon FM (2005)Remote sensing of phytoplankton groups in case 1waters from global SeaWiFS imagery. Deep-Sea Res I 52: 1989−2004

Balech E (1988) Los dinoflagelados del Atlántico sudocci-dental. Publ Espec Inst Esp Oceanogr 1, Ministerio deAgricultura, Pesca y Alimentación, Madrid

Bouman HA, Platt T, Sathyendranath S, Stuart V (2005)Dependence of light-saturated photosynthesis on tem-perature and community structure. Deep-Sea Res I 52: 1284−1299

Brainerd KE, Gregg MC (1995) Surface mixed and mixinglayer depths. Deep-Sea Res I 42: 1521−1543

Brewin RJW, Hardaman-Mountford NJ, Lavender SJ, Rait-sos DE and others (2011) An intercomparison of bio-opti-cal techniques for detecting dominant phytoplanktonsize class from satellite remote sensing. Remote SensEnviron 115: 325−339

Brown CW, Podestá GP (1997) Remote sensing of coccolitho-phore blooms in the western South Atlantic Ocean.Remote Sens Environ 60: 83−91

Carreto JI, Negri RN, Benavides HR (1981) Fitoplancton,pigmentos y nutrientes. Resultados campañas III y VI delB/I ‘Shinkai Maru’, 1978. Rev Invest Desarrollo Pesq 383: 181−201

Carreto JI, Montoya NG, Benavides HR, Guerrero R, Carig-nan MO (2003) Characterization of spring phytoplanktoncommunities in the Río de la Plata maritime front usingpigment signatures and cell microscopy. Mar Biol 143: 1013−1027

Cermeño P, Marañón E, Rodríguez F, Fernández E (2005)Large-sized phytoplankton sustain higher carbon-spe-cific photosynthesis than smaller cells in a coastaleutrophic ecosystem. Mar Ecol Prog Ser 297: 51−60

Claustre H, Babin M, Merien D, Ras J and others (2005)Toward a taxon-specific parameterization of bio-opticalmodels of primary production: a case study in the NorthAtlantic. J Geophys Res 110:C07S12, doi: 10.1029/2004JC002634

Falkowski PG (1980) Light-shade adaption in marine phyto-plankton. In: Falkowski PG (ed) Primary productivity inthe sea. Plenum Press, New York, NY, p 99–119

Fernández CI, Raimbault P, Garcia N, Rimmelin P (2005) Anestimation of annual new production and carbon fluxesin the northeast Atlantic Ocean during 2001. J GeophysRes 110:C07S13, doi: 10.1029/2004JC002616

Garcia VMT, Garcia CAE, Mata MM, Pollery RC and others(2008) Environmental factors controlling the phytoplank-ton blooms at the Patagonia shelf-break in spring. Deep-Sea Res I 55: 1150−1166

Geider RJ, Platt T, Raven JA (1986) Size dependence ofgrowth and photosynthesis in diatoms: a synthesis. MarEcol Prog Ser 30: 93−104

Gómez MI, Piola AR, Kattner G, Alder VA (2011) Biomass ofautotrophic dinoflagellates under weak vertical stratifi-cation and contrasting chlorophyll levels in subantarcticshelf waters. J Plankton Res 33: 1304−1310

Gregg WW, Casey NW, McClain C (2005) Recent trends inglobal ocean chlorophyll. Geophys Res Lett 32: L03606,doi: 10.1029/2004GL021808

Hama T, Miyazaki T, Ogawa Y, Iwakuma T, Takahashi M,Otsuki A, Ichimura S (1983) Measurement of photosyn-thetic production of a marine phytoplankton populationusing a stable 13C isotope. Mar Biol 73: 31−36

Hasle GR (1978) Using the inverted microscope. In: SourniaA (ed) Phytoplankton manual. Monographs on oceano-graphic methodology 6. UNESCO, Paris, p 191−196

Hirata T, Aiken J, Hardman-Mountford N, Smyth TJ, BarlowRG (2008) An absorption model to determine phyto-plankton size classes from satellite ocean colour. RemoteSens Environ 112: 3153−3159

Hoepffner N, Sathyendranath S (1992) Bio-optical charac-teristics of coastal waters: absorption spectra of phyto-plankton and pigment distribution in the western NorthAtlantic. Limnol Oceanogr 37: 1660−1679

Holm-Hansen O, Lorenzen CJ, Holmes RW, Strickland DH(1965) Fluorometric determination of chlorophyll. J ConsInt Explor Mer 30: 3−15

Kiefer DA (1973) Fluorescence properties of natural phyto-plankton populations. Mar Biol 22: 263−269

Le Quéré C, Harrison SP, Prentice IC, Buitenhuis ET andothers (2005) Ecosystem dynamics based on planktonfunctional types for global ocean biogeochemistry mod-els. Glob Change Biol 11: 2016−2140

Lee Z, Carder KL, Arnone R (2002) Deriving inherent opticalproperties from water color: a multi-band quasi-analyti-cal algorithm for optically deep waters. Appl Opt 41: 5755−5772

Lutz VA, Carreto JI (1991) A new spectrofluorometricmethod for the determination of chlorophylls and degra-dation products and its application in two frontal areas ofthe Argentine Sea. Cont Shelf Res 11: 433−451

Lutz VA, Segura V, Dogliotti AI, Gagliardini DA, Bianchi A,Balestrini CE (2010) Primary production in the ArgentineSea during spring estimated by field and satellite mod-els. J Plankton Res 32: 181−195

Mitchell BG (1990) Algorithms for determining the absorp-tion coefficient of aquatic particulates using the quantita-tive filter technique. Proc SPIE Ocean Optics X: 137−148

Moreno DV, Marrero JP, Morales J, García CL, Úbeda MJ,Rueda MJ, Llinás O (2012) Phytoplankton functionalcommunity structure in Argentinian continental shelfdetermined by HPLC pigment signature. Estuar CoastShelf Sci 100: 78−81

Nair A, Sathyendaranath S, Platt T, Morales J and others(2008) Remote sensing of phytoplankton functional types.Remote Sens Environ 112: 3366−3375

30

http://dx.doi.org/10.1016/j.rse.2008.01.021http://dx.doi.org/10.1117/12.21440http://dx.doi.org/10.1093/plankt/fbp117http://dx.doi.org/10.1016/0278-4343(91)90052-8http://dx.doi.org/10.1364/AO.41.005755http://dx.doi.org/10.1111/j.1365-2486.2005.1004.xhttp://dx.doi.org/10.1007/BF00389180http://dx.doi.org/10.1093/icesjms/30.1.3http://dx.doi.org/10.4319/lo.1992.37.8.1660http://dx.doi.org/10.1016/j.rse.2008.03.011http://dx.doi.org/10.1007/BF00396282http://dx.doi.org/10.1093/plankt/fbr031http://dx.doi.org/10.3354/meps030093http://dx.doi.org/10.1016/j.dsr.2008.04.011http://dx.doi.org/10.1029/2004JC002616http://dx.doi.org/10.1104/pp.66.4.592http://dx.doi.org/10.1029/2004JC002634http://dx.doi.org/10.3354/meps297051http://dx.doi.org/10.1007/s00227-003-1147-zhttp://dx.doi.org/10.1016/S0034-4257(96)00140-Xhttp://dx.doi.org/10.1016/j.rse.2010.09.004http://dx.doi.org/10.1016/0967-0637(95)00068-Hhttp://dx.doi.org/10.1016/j.dsr.2005.01.008http://dx.doi.org/10.1016/j.dsr.2005.06.015http://dx.doi.org/10.1080/01431160600821036

Segura et al.: Phytoplankton types in the Argentine Sea

Negri RN, Carreto JI, Benavides HR, Akselman R, Lutz VA(1992) An unusual bloom of Gyrodinium cf. aureolum inthe Argentine sea: community structure and condition-ing factors. J Plankton Res 14: 261−269

Omachi CY, Garcia VMT, Garcia CAE, Campos EJ (2010)Distribuicao espaco-temporal da estrutura da comu-nidade fitoplanctonica na Patagonia. In: AOCEANO. IIICongresso Brasileiro de Oceanografia-CBO’2010, RioGrande, 17 a 21 de maio 2010, p 161−164

Platt T, Gallegos CL (1980) Modeling primary production.In: Falkowski PG (ed). Primary productivity in the sea.Plenum Press, New York, NY, p 339−362

Platt T, Gallegos CL, Harrison WG (1980) Photoinhibitionand photosynthesis in natural assemblages of marinephytoplankton. J Mar Res 38: 687−701

Platt T, Rao DVS, Irwin B (1983) Photosynthesis ofpicoplankton in the oligotrophic ocean. Nature 301: 702−704

Platt T, Sathyendranath S, Ulloa O, Harrison WG, HoepffnerN, Goes J (1992) Nutrient control of phytoplankton pho-tosynthesis in the western North Atlantic. Nature 356: 229−231

Pope RM, Fry ES (1997) Absorption spectrum (380-700 nm)of pure water. II. Integrating cavity measurements. ApplOpt 36: 8710−8723

Raitsos DE, Lavender SJ, Maravelias CD, Haralabous J,Richardson A, Reid PC (2008) Identifying four phyto-plankton functional types from space: an ecologicalapproach. Limnol Oceanogr 53: 605−613

Rivas AL, Dogliotti AI, Gagliardini DA (2006) Seasonal vari-ability in satellite-measured surface chlorophyll in thePatagonian shelf. Cont Shelf Res 26: 703−720

Sabatini M, Akselman R, Reta R, Negri MR and others (2012)Spring plankton communities in the southern Patagonianshelf: hydrography, mesozooplankton patterns andtrophic relationships. J Mar Syst 94: 33−51

Sathyendranath S, Platt T (1988) The spectral irradiancefield at the surface and in the interior of the ocean: a

model for applications in oceanography and remotesensing. J Geophys Res 93: 9270−9280, doi: 10.1029/JC093iC08p09270

Sathyendranath S, Lazzara L, Prieur L (1987) Variations inthe spectral values of specific absorption of phytoplank-ton. Limnol Oceanogr 32: 403−415

Sathyendranath S, Watts LJ, Devred E, Platt T, Caverhill C,Maass H (2004) Discrimination of diatoms from otherphytoplankton using ocean-colour data. Mar Ecol ProgSer 272: 59−68

Schloss IR, Ferreyra ME, Almandoz GO, Codina R and oth-ers (2007) Role of plankton communities in sea-air varia-tion in pCO2 in the SW Atlantic Ocean. Mar Ecol ProgSer 332: 93−106

Silva RI, Negri R, Lutz V (2009) Summer succession of ultra-phytoplankton at the EPEA coastal station (northernArgentina). J Plankton Res 31: 447−458

Stuart V, Sathyendranath S, Head EJH, Platt T, Irwin B,Maass H (2000) Bio-optical characteristics of diatoms andprymnesiophyte populations in the Labrador Sea. MarEcol Prog Ser 201: 91−106

Subramaniam A, Carpenter EJ, Karentz D, Falkowski PG(1999) Optical properties of the marine diazotrophiccyanobacteria Trichodesmium spp. I. Absorption andspectral photosynthetic characteristics. Limnol Oceanogr44: 608−617

Throndsen J (1978) Productivity and abundance of ultra-and nanoplankton in Oslofjorden. Sarsia 63: 273−284

Tomas CR (1997) Identifying marine phytoplankton. Aca-demic Press, San Diego, CA

Uitz J, Claustre H, Morel A, Hooker S (2006) Vertical distri-bution of phytoplankton communities in open ocean: anassessment based on surface chlorophyll. J Geophys Res111: C08005, doi: 10.1029/2005JC003207

Uitz J, Huot Y, Bruyant F, Babin M, Claustre H (2008) Relat-ing phytoplankton photophysiological properties to com-munity structure on large scales. Limnol Oceanogr 53: 614−630

31

Editorial responsibility: Christine Paetzold, Oldendorf/Luhe, Germany

Submitted: October 8, 2012; Accepted: June 27, 2013Proofs received from author(s): September 9, 2013

http://dx.doi.org/10.1016/j.ecss.2012.01.007http://dx.doi.org/10.4319/lo.2008.53.2.0614http://dx.doi.org/10.1029/2005JC003207http://dx.doi.org/10.4319/lo.1999.44.3.0608http://dx.doi.org/10.3354/meps201091http://dx.doi.org/10.1093/plankt/fbn128http://dx.doi.org/10.3354/meps332093http://dx.doi.org/10.3354/meps272059http://dx.doi.org/10.4319/lo.1987.32.2.0403http://dx.doi.org/10.1029/JC093iC08p09270http://dx.doi.org/10.1016/j.jmarsys.2011.10.007http://dx.doi.org/10.1016/j.csr.2006.01.013http://dx.doi.org/10.4319/lo.2008.53.2.0605http://dx.doi.org/10.1364/AO.36.008710http://dx.doi.org/10.1038/356229a0http://dx.doi.org/10.1038/301702a0http://dx.doi.org/10.1093/plankt/14.2.261

cite43: cite28: cite5: cite14: cite42: cite3: cite27: cite13: cite41: cite26: cite39: cite12: cite40: cite25: cite38: cite24: cite37: cite10: cite8: cite36: cite6: cite22: cite35: cite4: cite21: cite34: cite19: cite2: cite20: cite33: cite18: cite17: cite45: cite31: cite9: cite44: cite29: cite7: cite30: cite15: