-
MARINE ECOLOGY PROGRESS SERIESMar Ecol Prog Ser
Vol. 505: 4964, 2014doi: 10.3354/meps10772
Published May 28
INTRODUCTION
In marine systems, primary production is largelydriven by
physical and chemical parameters (Falk -owski & Raven 2007,
Napolon et al. 2012). Manystudies have focused on the relationship
betweenprimary production and nutrients (Lippemeier et al.1999,
Behrenfeld et al. 2004, Claquin et al. 2010) ortemperature (Davison
1991, Claquin et al. 2008),while others have focused on the
relationship be -tween primary production and incident light
(Anning
et al. 2000). Biological parameters such as the struc-ture of
the phytoplankton assemblage can majorlyinfluence the variability
in primary production andproductivity (Behrenfeld & Falkowski
1997, Videauet al. 1998, Jouenne et al. 2005, 2007, Duarte et
al.2006, Claquin et al. 2010). The dynamics in the phyto -plankton
assemblage are mainly controlled by sea-sonal changes in light and
nutrient concentrations(Huisman & Weissing 1995), but physical
and chemi-cal parameters can also influence the relative
abun-dances of picophytoplankton and microphytoplank-
Inter-Research 2014 www.int-res.com*Corresponding author:
[email protected]
Dynamics of phytoplankton diversity structure andprimary
productivity in the English Channel
Camille Napolon1,2,3, Liliane Fiant3, Virginie Raimbault1,2,
Philippe Riou3, Pascal Claquin1,2,*
1Universit de Caen Basse-Normandie, UMR BOREA, 14032 Caen,
France2UMR BOREA, CNRS-7208, IRD-207, MNHN, UPMC, UCBN, 14032 Caen,
France
3IFREMER, Laboratoire Environnement Ressources de Normandie,
Avenue du Gnral de Gaulle, 14520 Port-en-Bessin, France
ABSTRACT: The dynamics of the phytoplankton assemblage, the
physical, chemical and biologi-cal parameters, and primary
productivity and production were monitored in the central
EnglishChannel along a transect between Ouistreham and Portsmouth
from January to December 2010.The spatial patterns of the
phytoplankton assemblage were controlled by the hydrological
charac-teristics of the water masses, and the annual structure of
the phytoplankton assemblage was char-acteristic of the central
English Channel and was controlled by seasonality. The spring bloom
wasdominated by a single species, Chaetoceros socialis, and
associated with low microphytoplanktonevenness and Shannon-Wiener
indices, whereas the evenness index was high from late spring
towinter and associated with the proliferation of pico- and
nanophytoplankton cells. We identified 2species responsible for
harmful algal blooms, Phaeocystis globosa, which dominated the
commu-nity in the northern part of the Seine Bay in May, and
Lepidodinium chlorophorum, which domi-nated the community near the
French coast in September. We examined the relationship
betweenmicrophytoplankton diversity and maximum primary production
and productivity. We found anegative parabolic relationship between
the diversity indices (evenness and Shannon-Wiener)and maximum
primary production, and a positive parabolic relationship between
the number oftaxa (richness) and maximum primary production.
However, we found no relationship betweenmaximum productivity and
the evenness or richness indices. High levels of productivity
weremeasured during the increasing abundance of pico and
nanophytoplankton cells, highlighting theimportance of taking the
dominant functional group into account, rather than the degree of
diver-sity, when explaining the level of productivity.
KEY WORDS: Phytoplankton diversity Primary production
Productivity English Channel
Resale or republication not permitted without written consent of
the publisher
This authors' personal copy may not be publicly or
systematically copied or distributed, or posted on the Open Web,
except with written permission of the copyright holder(s). It may
be distributed to interested individuals on request.
-
Mar Ecol Prog Ser 505: 4964, 2014
ton cells. For example, high temperatures and oligo-trophic
waters stimulate the development of pico-phytoplankton (Agawin et
al. 2000), while microphy-toplankton tend to dominate coastal and
eutrophicwaters (Pannard et al. 2008).
It is well known that the structure and diversity ofthe
phytoplankton assemblage drive productivityand hence carbon input
into marine systems (Mittel-bach et al. 2001, Gamfeldt &
Hillebrand 2011). Somestudies have focused on the relationship
betweenproductivity and the biodiversity of ecosystems, butthe
shape of the relationship is variously reported asa negative linear
relationship, a positive linear rela-tionship, a unimodal
relationship, or no relationshipat all (Abrams 1995, Waide et al.
1999, Jouenne et al.2007, Chase 2010, Claquin et al. 2010).
Limited species diversity can reduce productivityand this can
explain the positive linear diversityproductivity relationship,
which is the one most fre-quently found. Mechanisms that might
explain thisrelationship include: (1) an increasing level of
diver-sity increases the probability that a highly produc-tive
species could be present in a phytoplanktonassemblage, and (2) the
complementarity of speciescould lead to higher productivity in
systems charac-terized by high diversity (Tilman et al. 1997,
Loreau1998). The unimodal diversityproductivity relation-ship can
be explained by competitive exclusionoccurring as productivitiy
increases and resourceavailability decreases (Huston &
Deangelis 1994,Duarte et al. 2006). The negative linear
relationshipis observed when high production is associated withlow
biodiversity due to the domination by one or fewspecies which
exclude other taxa from the ecosys-tem. The different
diversityproductivity relation-ships described in the literature
indicate that thelevel of diversity that triggers productivity is
still notclear. The complexity of (and variability in)
environ-mental factors may explain the heterogeneity of
thediversityproductivity relationship, as may the dif-ferent
methodologies used to describe the degree ofdiversity.
In this context and in order to improve our under-standing of
the diversityproductivity relationship,the dynamics in, and
diversity of, phytoplankton as -semblages need to be further
monitored and de -scribed in parallel with environmental (physical
andchemical) parameters. The English Channel (north-western Europe)
is an epicontinental sea understrong anthropogenic pressure.
Napolon et al.(2012) describe 4 distinct hydrological areas along
atransect that transverses the central region of theEnglish
Channel. The functioning of each hydrologi-
cal area depends mainly on nutrient inputs fromrivers and on
offshore influences (Napolon et al.2012). It is consequently useful
to study the dynamicsof the community structure, diversity and
primaryproduction in this highly variable system.
In the present study we monitored the dynamics ofthe
phytoplankton assemblage, the physical, chemi-cal, and biological
parameters, and primary produc-tivity and production, in the
central English Channel,along a transect between Ouistreham and
Ports mouth,over a period of 1 yr. Our objectives were to (1)
studythe influence of the physical, chemical, and biologi-cal
parameters on the dynamics in the phytoplanktonassemblage, (2)
monitor the spatiotemporal variabil-ity in the microphytoplankton
diversity at 2 scales(intra-station and inter-station) and identify
commonpatterns between seasons, and (3) identify
possiblerelationships between phyto plankton biomass, phyto
-plankton dynamics, and primary production.
MATERIALS AND METHODS
Sampling strategy
Monthly measurements were made from Januaryto December 2010 in
the central region of the EnglishChannel (except in April and
November). Data werecollected in daylight on board the
Normandie-Brit-tany ferries during their daily 175 km crossing be
-tween Ouistreham (France, 49 17 27 N, 0001445 W) and Portsmouth
(Great Britain, 50 48 49 N,001 05 29 W) (Fig. 1). Physical
parameters (tempera-ture, salinity and incident light) were
recorded every500 m, photosynthetic parameters were measuredevery 5
km and biological (chl a, phytoplankton spe-
50
12
34
5
Ouistreham
Portsmouth
100 km
GREAT BRITAIN
FRANCE
ENGLISHCHANNEL
51N
48
50
49
5 W 4 3 2 1 0 1 2 E
N
67
89
10
Fig. 1. The English Channel, with the location of the sam-pling
transect and the 10 stations at which complete data
sets were obtained
Aut
hor c
opy
-
Napolon et al.: Phytoplankton dynamics in the English
Channel
cies, suspended particular matter) and chemicalparameters
(dissolved inorganic nitrogen, phosphateand silicate) every 15 km.
The data set is thus com-plete for 10 sampling stations (Fig. 1).
Water sampleswere collected by using the difference in
pressurebetween the seawater (1.4 bar) and the ship (1 bar)through
a pipe let down from the front of the ship toa depth of 4 m.
Sampling stopped in the vicinity ofthe harbour to limit possible
contamination by pol-luted waters. Supplementary data, including
timeand position (latitude, longitude) were provided bythe
crew.
Chl a, physical, and chemical parameters
The chl a concentration was measured using themethod of
Welschmeyer (1994; see also the descriptionin Napolon et al. 2012).
Temperature and salinitywere recorded with a YSI 6600 V2
multi-parameterprobe, and light was measured on deck with a 2PAR
sensor LI-192 connected to a LI-1400 data logger(LI-COR). Dissolved
inorganic nitrogen (DIN), phos-phate (DIP) and silicate (DSi)
concentrations weredetermined in the laboratory using an AxFlow
AA3autoanalyzer, following the method of Aminot &Krouel (2007).
Concentrations of suspended partic-ulate matter (SPM) were measured
using the methodof Aminot & Chaussepied (1983).
Species composition
Microphytoplankton. Immediately after sampling,1 l of water was
preserved using acid Lugols solution(2 ml l1). The Utermhl (1931)
method was used forthe analysis of the composition and
concentration ofmicrophytoplankton. After homogenisation, a 10
mlwater sample was poured into a sedimentation cham-ber and left to
settle for at least 8 h. The phytoplank-ton cells on the chamber
bottom were identified andcounted using an inverted microscope.
Organismswere identified to the lowest taxonomic level possi-ble,
depending on the skill of the operator (a singleoperator was
involved for all taxonomic analysis).The strategy used for each
species was to count thewhole bottom of the chamber, half the
bottom, oralong a diagonal, depending on the abundance of
thespecies. The same magnification (400) was used inall cases and
the counts are expressed in cells l1.
Pico- and nanophytoplankton. Analyses of pico-and
nanophytoplankton samples and processing offlow cytometric data
(FACSCanto II flow cytometer,
BD-Biosciences) were performed at the LaboratoireNational
dAnalyse en Cytomtrie en Flux, CNRSINSU, Observatoire Ocanologique
de Banyuls surmer, France. The samples were fixed with
glu-taraldehyde at a final concentration of 1%, frozen inliquid
nitrogen, stored at 80C, and were thenthawed at room temperature
immediately beforecytometric analysis (Vaulot et al. 1989). A
blueargon laser (excitation at 488 nm) was used to dis-tinguish and
count autotrophic and heterotrophiccells. Phototrophic cells were
enumerated accordingto their right-angle light scatter properties
(SSC,roughly related to cell size), and the orange (585/42mm BP)
and red (670 nm LP) fluorescence from phy-coerythrin and
chlorophyll pigments, respectively.Data were acquired using
FACSDiva software (BD-Biosciences). Fluorescent 1.002 m beads
(Poly-sciences) were systematically added to each ana-lysed sample
to normalize cell fluorescence andlight scatter emission, thus
making it possible tocompare the results. To estimate cell
abundancesaccurately, the flow rate of the sample was
routinelymeasured every 10 samples using BD Trucounttubes (Cat.
340334; Lot 822525).
Productivity and primary production
We used the maximum primary production (PPmaxthat we transformed
from mg C l1 h1 to mg C m2
d1) data of Napolon & Claquin (2012) and calcu-lated maximum
productivity rates (PBmax) using:
PBmax = PPmax / [chl a] (1)
where PBmax is expressed in mg C mg1 chl a h1, PPmaxin mg C l1
h1 and [chl a] in mg chl a l1.
Diversity indices
To characterise the species richness of the micro-phytoplankton
community, we counted the numberof taxa (S) in each sample. The
ShannonWienerindex (H ) of the microphytoplankton was
calculatedusing:
H = Si=1S pi ln(pi) (2)
and the evenness index (J ) was calculated followingthe widely
used formula of Pielou (1966):
J = [Si=1S pi ln(pi)] / ln(S) (3)
where pi is the proportion of the microphytoplanktonspecies
i.
51A
utho
r cop
y
-
Mar Ecol Prog Ser 505: 4964, 2014
Statistical analyses
Canonical correspondence analysis (CCA) wasperformed using R
v.2.11.1, to examine the relation-ship between physical, chemical,
and biologicalparameters and the structure of the
phytoplanktonassemblage. For this analysis, a matrix was
builtcontaining the physical and chemical parameters,the biological
parameters, and the abundance ofeach microphytoplankton species in
the samples.Microphytoplankton species abundance data (cellsl1)
were log-transformed [log10(x + 1)] as this vari-able may have an
asymmetric distribution due toexponential growth when conditions
are favourable(Ter Braak & Smilauer 2002). Physical,
chemical,and biological data were centered by the mean ofthe
variable and reduced by the variance. CCA isan efficient ordination
technique when a Gaussianrelationship between species and the
environmentalgradients is expected (Ter Braak 1986). This
con-strained analysis extracts the best environmentalgradients that
ex plain the maximum variability inspecies data. Biological
variables (chl a, diatom,dinoflagellate, Synechococcus and
picoeukaryoteconcentrations, PPmax, PBmax, species richness
S,Shannon-Wiener index H , and the microphyto-plankton evenness
index J ) were added as supple-mentary variables to the CCA, and
were thus corre-lated with the canonical axis (which is a
linearcombination of environmental parameters) on theplot (Klein et
al. 2011).
To resolve the space and time variability in thestructure of the
microphytoplankton community, par-tial triadic analysis (PTA) was
applied to the data setusing the ADE-4 package (Chessel et al.
2004, Dray& Dufour 2007) with the R v.2.11.1 software. The
datawere organised in sub-matrices. A sub-matrix con-taining the
composition of microphytoplankton spe-cies recorded for all
sampling dates was built for eachstation. The data (cells l1) were
log-transformed[log10(x + 1)] to obtain a normal distribution.
ThePTA analysis compares the structures shared by thesubmatrices
and identifies stations with a similartemporal structure. Wards
cluster analysis based onthe vector correlation coefficients of the
PTA wasperformed to distinguish groups of stations accordingto
their microphytoplankton composition (Ward 1963).
To study the relationship between microphyto-plankton richness
and PPmax, as well as between themicrophytoplankton evenness index
and PPmax,quadratic polynomial regression analyses were car-ried
out on the data set using SigmaPlot v.11.0 (SystatSoftware).
To identify inter-site and intra-site variability, weused the
double principal coordinate analysis (DPCoA)developed by Pavoine et
al. (2004). This analysismakes it possible to break down total
inertia into theinertia of species around stations (intra-station
diver-sity) and the inertia between stations
(inter-stationdiversity). The intra-station diversity is the
inertia(variance) of species weighted by their relative abun-dance
at the station concerned, within the space ofthe DPCoA. Conversely,
the inter-station variabilityis the inertia of all the stations
weighted by theweight vector of each station within the space of
theDPCoA. DPCoA were performed with R v.2.11.1using the ADE-4
package (Pavoine et al. 2004). A sin-gle matrix was built
containing the frequencies ofmicrophytoplankton species at each
station and ateach sampling date, with species listed in the
columnsand the station/date in the rows.
RESULTS
Spatiotemporal variability in biological parameters
The phytoplankton biomass (chl a, data fromNapolon et al. 2012)
and the number of diatom cellsshowed the same pattern, but the
pattern varied con-siderably over time and in space (Fig. 2A,B).
Thehighest values were observed from the French coastto the
northern part of the Seine Bay, between theend of winter and June.
The highest chl a concentra-tion (7.2 g l1) was observed in March
and the high-est number of diatom cells (955 800 cells l1) was ob
-served in May. A weaker winter/spring proliferationwas observed
near the English coast (i.e. Stns 9 and10) from January to May,
with a maximum chl a con-centration of 3.3 g l1 and a maximum
number ofdiatom cells of 742 500 cells l1, recorded in spring.
Compared to the concentrations of diatoms, con-centrations of
dinoflagellates remained low through-out the year of our study
(Fig. 2C). Dinoflagellatesproliferated later than diatoms, i.e.
between July andSeptember, with values ranging between 2400 and139
000 cells l1, near the French coast.
The highest concentrations of cryptophytes wererecorded between
May and July on both coasts, thehighest value being 1642 cells ml1
recorded in Maynear the French coast (Fig. 2D).
The concentrations of picoeukaryotes (Fig. 2E) andSynechococcus
(Fig. 2F) showed the same spatio -temporal pattern over the year
(rPicovSyne = 0.802). Thehighest values were recorded between the
Englishcoast and the centre of the English Channel, espe-
52A
utho
r cop
y
-
Napolon et al.: Phytoplankton dynamics in the English Channel
53
Number of Synechococcus cells
Chl a0.250.400.631.001.582.513.98
0.40
0.40
0.40
0.40
0.40
0.63
0.63
0.40
1.00
1.00
0.63
0.63
0.63
0.63
0.63
1.001.00
1.00 1.00
1.58
1.58
2.51
1.58
1.58
1.58
1.58
3.98
2.51
2.51
1.00
1.001.58
A Number of diatom cells31621000031623100000316228
3162
3162
3162
3162
3162
10000
10000
10000
10000
10000
10000
10000
31623
31623
31623
31623100000
100000
316228
31623
31623
100000
100000
B
Number of dinoflagellate cells103210031610003162
32
32
32100
10032
32
100
100
316
316
316
316
316
100
1000
1000
1000
1000
3162
3162
3162
10001000
3162
316
1000
1000
1000
316
100
100
32
32
1000
316
316
10
10
316
316
C Number of cryptophyte cells4063100158251398
63
40
40
63
63
100
63100
100
100
100
158
158
251
158
158
158
251
158
251
251
251
158
398
251
251
251
251
251
158
158
158
158
100
100
100
63
63
158
100
D
Number of picoeukaryote
cells19952512316239815012631079431000012589
3162
3162
3981
3981
5012
3981
3981
3981
3162
3162
3162
5012
5012
5012
5012
5012
5012
5012
3162
3162
3981
79436310
6310
6310
6310
6310
6310 7943
7943
7943
7943
10000
10000
10000
10000
39813162
1258912589
3981
3981
10000
10000
7943
7943
6310
3981
3981
5012
3981
3981
5012
3981
3162
3162
2512
3162
3162
3981
2512
3162
3162
2512
1995
3981
E1585251239816310100003981
3981
6310
6310
6310
63106310
6310
3981
3981
3981
3981
10000
10000
10000
10000
10000
10000
2512
2512
6310
3981
1585
1585
1585
3981
6310
2512
2512
2512
2512
2512
2512
1585
1585
F
Date DateFeb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Feb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
Latit
ude
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Feb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Feb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Feb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Feb 10 Apr 10 Jun 10 Aug 10 Oct 10 Dec 10
50.6N
50.4
50.2
50.0
49.8
49.6
49.4
Fig. 2. Latitude-time distribution of (A) chl a biomass (g chl a
l1) (data from Napolon et al. 2012), (B,C) abundance (cells l1)of
(B) diatoms and (C) dinoflagellates, and (DF) abundance (cells ml1)
of (D) cryptophytes, (E) picoeukaryotes, and
(F) Synechococcus spp.
Aut
hor c
opy
-
Mar Ecol Prog Ser 505: 4964, 2014
cially in the centre of the English Channel betweenJune and
August. The overall highest values, 32 835cells ml1 for
picoeukaryotes and 55 067 cells ml1 forSynechococcus, were recorded
in July at latitude50.0N.
Microphytoplankton S (Fig. 3A) varied over timeand in space. A
decreasing southnorth gradient wasobserved along the transect, with
the highest numberof taxa (33 taxa sample1) observed in May at
lati-tude 49.4N. Microphytoplankton H (Fig. 3B) and J (Fig. 3C)
showed the same spatiotemporal variabilityover the year except in
NovemberDecember whereH dropped and J remained high. Minimum
valueswere recorded between the end of winter and theend of spring
from the French coast to the centre ofthe English Channel. The
lowest H (0.25) and thelowest J (0.09) were recorded in May at
latitude49.9N.
PPmax showed the same spatiotemporal pattern
asmicrophytoplankton S except for the relatively highvalues
recorded on the English coast over the year(Fig. 3D). The highest
PPmax value, 28.7 mg C m2 d1,was measured in June in the centre of
the EnglishChannel.
PBmax (Fig. 3E) remained low near the French coastthroughout the
year of study. High values wererecorded between May and December
between theEnglish coast and latitude 49.8N with a maximumvalue of
10.6 mg C mg1 chl a h1 recorded in July atlatitude 50.0N.
Phytoplankton assemblage dynamics
We used CCA to link the variability in the structureof
phytoplankton assemblage to physical, chemical,and biological
parameters (Fig. 4). The first 2 axes ofthe CCA explained more than
59% of the variance ofthe data set (Axis 1: 37.74%; Axis 2:
21.60%). MonteCarlo permutation tests showed that all the
canonicalaxes (p < 0.001) were statistically significant. As
pre-viously reported in Napolon et al. (2012), physical,chemical,
and biological parameters revealed tempo-ral uncoupling due to the
seasonality of the parame-ters (Fig. 4A). The high concentrations
of diatom cellswere positively linked to high concentrations of chl
a,PPmax, S, and irradiance, and negatively linked to J ,H , and
concentrations of DSi. Conversely, dinofla-gellate concentrations
were linked to high PBmax andhigh temperatures, and low DIN and chl
a concentra-tions. Synechococcus concentrations were
positivelylinked to PBmax and negatively linked to high
nutrientconcentrations, and positively to picoeukaryote con-
centrations, which, in turn, were positively linked toPPmax.
A clear seasonal structure was apparent in thephytoplankton
assemblage throughout the year ofstudy (Fig. 4B). On the left part
of the CCA (Fig. 4B,C),the summer and autumn communities were
charac-terised by dinoflagellates, while diatoms were ob -served
throughout the year (Fig. 4B,C) with the high-est concentrations in
spring (Fig. 4A,B). The springdiatom peak near the French coast was
mainly dom-inated by diatoms of the genus Chaetoceros,
particu-larly C. socialis (C_s) in May (880 900 and 846 000cells l1
at Stns 1 and 2, respectively). In contrast, thecommunity near the
English coast was characterisedby Skeletonema spp., (Sk) (449 000
cells l1) and 3species of Thalassiosira (T. levanderi [T_l], T.
minima[T_m] and T. nordenskioeldii [T_no], total of 294 500cells
l1), with the highest concentrations recorded inMarch at Stn 10.
The summer/autumn peak of dino-flagellates was characterised by
Lepidodinium chlo -ro phorum (L_c) (Fig. 4C), with a maximum
concen-tration of 135 800 cells l1 recorded in September atStn 2. A
high concentration of Phaeocystis globosa(P_g) (444 400 cells l1)
was recorded in May at Stn 4.
Spatial variability
The PTA interstructure analysis enabled us todetect similarities
in the structure of the communityof microphytoplankton between
stations over theyear of study. The first eigenvalue of the PTA
analy-sis represents more than 31% of the total inertia andis
isolated from the others (Fig. 5A). This suggests aclose link
between stations, which in turn indicates astrong common temporal
structure of the microphy-toplankton assemblage between stations.
The sec-ond eigenvalue represents more than 11% of totalinertia
(Fig. 5A) and highlights the differences be -tween stations (Fig.
5B). Based on Wards clusteranalysis (Fig. 5C), the transect between
Ouistrehamand Portsmouth can be divided into 3 groups of sta-tions:
Stns 1 to 3;, Stns 4 to 8, and Stns 9 and 10.
Diversity
A significant quadratic polynomial relationshipwas found between
microphytoplankton S and PPmax(Fig. 6A), between microphytoplankton
J and PPmax(Fig. 6B) and between H and PPmax (data not shown,R2 =
0.066, y = 0.0010x2 + 0.0088x + 1.7590). Therewas thus a positive
link between S and PPmax (p
