Introduction
Marine phytoplankton form a taxonomically and functionally diverse group,
where communities are structured by a variety of factors, including nutrient
and light availability, predation and competition for resources
(Litchman and Klausmeier, 2008). Such environmental
heterogeneity creates biogeographical patterns of abundance, composition,
traits, and diversity of phytoplankton communities in the global ocean
(Barton et al., 2013;
Follows et al., 2007; Hays et al., 2005). Phytoplankton communities within a
biogeographical region are subject to similar environmental conditions, such
as temperature (Bouman et al., 2003), nutrient
concentration (Browning et al., 2014), and irradiance
(Arrigo et al., 2010). These
environmental factors, along with phytoplankton community composition itself
(Bouman et al., 2005), affect the
overall photo-physiological response and bulk rates of primary production.
The biogeography of phytoplankton communities and their photophysiological
characteristics, consequently, directly impact the structure of marine
ecosystems due to their functional roles in biogeochemical cycling and the
transfer of energy to higher trophic levels. For example, distinct
phytoplankton assemblages influence particulate
(Martiny
et al., 2013a, b; Smith and Asper, 2001) and dissolved elemental
stoichiometry (C : N : P) (Weber and Deutsch, 2010), the
drawdown of gases (Arrigo,
1999; Tortell et al., 2002) and the efficiency of carbon export
(Guidi
et al., 2009; Le Moigne et al., 2015) in different ways. Patterns of
phytoplankton stoichiometry may be consistent phylogenetically within higher
taxonomic levels (Ho et al., 2003; Quigg et al.,
2003); however, stoichiometry also varies according to nutrient supply
ratios
(Bertilsson
et al., 2003; Rhee, 1978) and phenotypically within species of the same
population (Finkel et al., 2006).
The subarctic North Atlantic is a complex system with contrasting
hydrography that structures plankton communities within distinct
biogeographical provinces
(Fragoso
et al., 2016; Head et al., 2003; Li and Harrison, 2001; Platt et al., 2005;
Sathyendranath et al., 1995, 2009). Biogeographical regions of the Labrador
Sea shape phytoplankton community composition
(Fragoso et al.,
2016), bio-optical properties
(Cota,
2003; Lutz et al., 2003; Platt et al., 2005; Sathyendranath et al., 2004;
Stuart et al., 2000), and the seasonality of phytoplankton blooms
(Frajka-Williams and
Rhines, 2010; Lacour et al., 2015; Wu et al., 2007, 2008). Phytoplankton
blooms, for example, occur first (April to early May) on the shelves due to
haline stratification driven by the input of Arctic-related waters,
in addition to rapid sea ice melt on the Labrador Shelf near Canada
(Frajka-Williams and Rhines, 2010;
Wu et al., 2007). The central Labrador bloom occurs later in the season
(late May to June) as a result of thermal stratification
(Frajka-Williams and Rhines, 2010).
Fragoso et
al. (2016) showed that the biogeography of phytoplankton communities in the
Labrador Sea during spring and early summer is shaped by distinct species
found in Atlantic or Arctic waters, which may have distinct influences on
biogeochemical cycles and the transfer of energy to upper trophic levels.
However, these authors focused on taxonomy and only investigated relatively
large phytoplankton (> 4 µm). The photophysiological and
biogeochemical signatures, such as particulate matter stoichiometry (C : N
ratio) of these different spring phytoplankton communities occurring in
distinct sectors of the Labrador Sea have not been investigated.
Quantification of marine phytoplankton community composition, for a large
numbers of samples, is challenging due to small cells (< 4 µm)
being difficult to identify and accurately count using light microscopy, in
addition to being a very time-consuming method. To overcome these problems,
quantification and analyses of phytoplankton pigments by high-performance
liquid chromatography (HPLC) has been widely used to monitor phytoplankton
community distributions over large temporal and spatial scales
(e.g.
Aiken et al., 2009; Peloquin et al., 2013; Platt et al., 2005). The
interpretation of the pigment data is not always straightforward, since some
pigments are shared by several algal groups and can vary according to local
nutrient and light conditions
(e.g.
DiTullio et al., 2007; van Leeuwe and Stefels, 1998, 2007). The
chemotaxonomic tool, CHEMTAX (CHEMical TAXonomy), provides a valuable
approach to estimate phytoplankton group abundances when used in conjunction
with microscopic information
(Irigoien et al.,
2004; Mackey et al., 1996; Wright et al., 1996). CHEMTAX has the advantage
of providing more information about phytoplankton groups than individual
diagnostic pigments or ratios and has been used widely to investigate
phytoplankton biogeography on regional scales
(e.g.
Muylaert et al., 2006; Wright and Van den Enden, 2000) and globally
(e.g. Swan et al., 2015).
Here, we investigated the multi-year (2005–2014) distributions of late
spring and early summer (May to June) phytoplankton communities in the
various hydrographic settings across the shelves, slopes, and deep basin of
the Labrador Sea based on phytoplankton pigments. In addition, we also
examined the overall photophysiological and biogeochemical traits associated
with these different phytoplankton communities. The purpose of this study
was to answer the following questions: were there distinct phytoplankton
communities in the Labrador Sea and if so, what were their main
constituents? How did spatial and temporal variability in environmental
factors explain the phytoplankton community distribution and composition?
What were the linkages between community composition and variability in both
particulate matter stoichiometry (i.e. C : N ratios) and photophysiological
traits (parameters of the photosynthesis versus irradiance relationships)
across the Labrador Sea?
Our results provide a geographical description of the phytoplankton
community structure in spring and early summer surface waters of the
Labrador Sea based on pigment data and CHEMTAX analysis from over a decade
of sampling (2005–2014). We show that several distinct phytoplankton
communities exist, which vary between the different hydrographic zones of
the Labrador Sea, and that they present variable patterns in terms of C : N
ratios and photophysiological responses to environmental conditions.
Methods
Study area
The Labrador Sea is a high-latitude marginal sea located in the northwestern
part of the Atlantic Ocean, and is an important transition zone between
Arctic and subarctic ecosystems (Fig. 1). It is bounded by Davis Strait to the
north, a line from Cape St Francis in Newfoundland (47∘45′ N,
52∘27′ W) to Cape Farewell (southern tip of Greenland) to the
southeast, and the coast of Labrador and Newfoundland to the west (Fig. 1)
(International Hydrography Organization, 1953). The bathymetry
of the Labrador Sea can be subdivided into the wide continental shelf and
relatively gentle continental slope on its western side (the Labrador Shelf,
> 500 km wide and < 250 m deep) and the narrow shelf and
steep continental slope on the eastern side (the Greenland Shelf and Slope,
< 100 km wide and < 2500 m deep).
Map showing stations along the AR7W transect and additional
stations sampled during late spring and early summer (2005–2014). The
station positions are superimposed on a composite image of sea surface
temperature for the last 3 weeks of May 2006 collected by the NOAA
satellite (AVHRR). White patches represent ice (Labrador and Greenland
coasts). Circulation elements – colder currents (Labrador Current, Arctic
outflows, and West Greenland Current, blue solid arrows), warmer currents
(Irminger Current (IC) and extension, dark red and light solid arrows,
respectively). The extended branch of the IC is a modified (cooled and
freshened) water mass caused by lateral and vertical mixing along the
Labrador slope.
The upper Labrador Sea (< 200 m) is comprised of waters originating
from the North Atlantic and the Arctic (Yashayaev, 2007).
Atlantic-influenced waters occur mostly in the central Labrador Sea, where
waters are relatively warm, salty, and mainly identified as the Irminger
Current (IC). Cold, low-salinity waters originate from the Arctic via the
surrounding shelves and are mainly identified as the Labrador Current (LC)
and the West Greenland Current (WGC) (Fig. 1). Circulation in the central
basin of the Labrador Sea is complex, often showing a gyre-like flow system
that alternates in direction (Palter et al., 2016; Wang et al., 2016).
The inshore branch of the LC overlies the Labrador Shelf and includes Arctic
waters originating from Baffin Bay and the Canadian Arctic Archipelago via
Davis Strait and from Hudson Bay via Hudson Strait, together with inputs of
melting sea ice that originates locally or from further north. The main
branch of the LC flows along the Labrador slope from north to south and is
centred around the 1000 m depth contour. It is composed of a mixture of
Arctic water from Baffin Bay via Davis Strait and the branch of the WGC that
flows west across the mouth of Davis Strait. The WGC, which flows from south
to north over the Greenland Shelf and along the adjacent slope, is a mixture
of cold, low-salinity Arctic water exiting the Nordic Seas with the East
Greenland Current (EGC) (Yashayaev, 2007),
together with sea ice and glacial melt water (Fig. 1). The WGC often spreads
westwards, forming a “tongue” of buoyant freshwater, where accumulation of
low-salinity waters is driven by high eddy kinetic activity in the central
eastern Labrador Sea during spring (Frajka-Williams and Rhines, 2010). The
WGC often floats over the IC in the central-eastern part of the Labrador
Sea; however, the IC is usually observed in surface waters of the
central-western Labrador Sea during spring. More detailed descriptions of
the hydrography of the Labrador Sea can be found elsewhere (e.g. Fragoso et
al., 2016, Head et al., 2013; Yashayaev and Seidov; Yashayaev, 2007).
Sampling
Data used for this study were obtained along the AR7W Labrador Sea
hydrography line (World Ocean Circulation Experiment Atlantic Repeat 7-West
section, for details see
Fragoso et
al., 2016), which runs between Misery Point on the Labrador coast (through
Hamilton Bank on the Labrador Shelf) and Cape Desolation on the Greenland
coast. Stations were sampled during late spring and/or early summer, varying
within a 6-week window (see sampling dates in Table 1) over a period of 10 years (2005–2014) by scientists from the Bedford Institute of Oceanography
(BIO), Canadian Department of Fisheries and Oceans. Fixed stations (total of
28), as well as some additional non-standard stations, were sampled across
the shelves and central basin on the AR7W section, or slightly north or
south of this transect (Fig. 1).
Research cruises, sampling dates, and number of samples per cruise
(n) where pigment data were collected in the Labrador Sea during early
spring and late summer (2005–2014). HUD-Year-ID and JR302 refer to
expeditions carried out on board the CCGS Hudson (Canada) and RRS James Clark Ross (UK), respectively.
Cruise
Dates
Year
n
HUD-2005-016
29 May–3 June
2005
25
HUD-2006-019
23–31 May
2006
12
HUD-2007-011
11–21 May
2007
32
HUD-2008-009
22–29 May
2008
25
HUD-2009-015
18–23 May
2009
26
HUD-2010-014
14–24 May
2010
27
HUD-2011-009
11–17 May
2011
33
HUD-2012-001
3–11 June
2012
30
HUD-2013-008
9–21 May
2013
27
JR302
10–24 June
2014
16
Vertical profiles of temperature and salinity were measured with a Sea-Bird
CTD system (SBE 911). Seawater samples were collected using 10 L Niskin
bottles mounted on a rosette frame. Mixed-layer depths were calculated from
the vertical density (σO) distribution and defined as the depth
where σO changes by 0.03 kg m-3 from a stable surface
value (∼ 10 m) (Weller and Plueddemann,
1996). A stratification index (SI) was also calculated as the seawater
density difference (between 10 to 60 m) normalized to the equivalent
difference in depth.
Water samples from the surface layer (< 10 m) were collected
(0.5–1.5 L) for the determination of fluorometric chlorophyll a
(Chl af), accessory pigments, nutrients, particulate organic carbon (POC),
and nitrogen (PON) analysis, and for primary production measurements.
Filters for Chl af measurements were immediately put in scintillation
vials containing 10 mL of 90 % acetone, which were placed into a
-20 ∘C freezer and extracted in the dark for 24 h. Samples for
detailed pigment analysis were filtered onto 25 mm glass fibre filters (GF/F
Whatman Inc., Clifton, New Jersey), immediately flash frozen in liquid
nitrogen, and kept frozen in a freezer (at -80 ∘C) until analysis
in the BIO (2005–2013) or National Oceanography Centre (UK) (2014)
laboratories within 2–3 months of collection. Volumes of water sampled for
pigment analysis were adjusted, such that samples were filtered as quickly
as possible (< 10 min). Nutrient samples were kept refrigerated at
5 ∘C and analysed at sea (within 12 h of collection) on a SEAL
AutoAnalyser III. Samples for POC and PON were filtered (0.25–1 L) onto
25 mm pre-combusted (400 ∘C, 12 h) GF/F filters, frozen (-20 ∘C),
and returned to the BIO laboratory for later analysis.
Biogeochemical analysis
Chlorophyll a concentrations were determined fluorometrically after 24 h of
extraction in 90 % acetone on board using a Turner Designs fluorometer
(Holm-Hansen et al., 1965). Back in the
laboratory, POC / PON samples were oven dried (60 ∘C) for
8–12 h, stored in a dessicator, pelletized in pre-combusted tin foil
cups and analysed using a PerkinElmer 2400 series CHNS/O analyser, as
described in Pepin and Head (2009).
Pigment analysis
Pigments (chlorophyll a and accessory pigments) were quantified using
reverse-phase, high-performance liquid chromatography (HPLC). Methods for
2005–2013 (Hudson cruises), including information about the standards,
calibration, and quantification procedures are described in detail in
Stuart and Head (2005), known as the
“BIO method”. Methods for samples collected in 2014 (JR302 cruise) are
described in
Poulton et
al. (2006). Quality control of both methods was applied according to
Aiken et al. (2009). Precision of the instruments was
tested by running samples and standards and the coefficient of variation for
pigments were < 10 % of the mean. Limits of detection were
∼ 0.01 and 0.002 mg m-3 for carotenoids and chlorins,
respectively (E. J. H. Head, personal communication, 2016; Poulton et
al., 2006). Pigment concentrations below detection limits were not reported.
A list of pigments identified and quantified for this study is included in
Table 2.
List of phytoplankton pigments and their distributions in algae
groups; abbreviations and formulas.
Abbreviation
Name
Characteristic of the pigment
Present in/index of/formula
PSC
Photosynthetic carotenoid
Light harvesting
All algae
PPC
Photoprotective carotenoid
Photoprotection
All algae
PPP
Photosynthetic pigment
Light harvesting
All algae
But-fuco
19'-butanoyloxyfucoxanthin
PSC
Prymnesiophytes, chrysophytes, and dinoflagellates type 2a (lacking peridinin)
Hex-fuco
19'-hexanoyloxyfucoxanthin
PSC
Major in prymesiophytes and dinoflagellates type 2a (lacking peridinin)
Allo
Alloxanthin
PPC
Cryptophytes
α-Car
α-carotene
PPC
Dominant in prochlorophytes, rhodophyte, and cryptophyte
β-Car
β-carotene
PPC
Dominant in cyanobacteria, prochlorophytes, chlorophytes, prasinophytes, euglenophytes, and diatoms
Chl b
Chlorophyll b
PPP
Chlorophytes, prasinophytes, euglenophytes
Chl c1+c2
Chlorophyll c1+c2
PPP
Diatoms, prymnesiophytes, dinoflagellates, cryptophytes, chrysophytes, and raphidophytes
Chl c3
Chlorophyll c3
PPP
Prymnesiophytes, chrysophytes and dinoflagellates type 2a (lacking peridinin)
Chlide a
Chlorophyllide a
Degradation product of Chl aHPLC
Senescent phytoplankton
DD
Diadinoxanthin
PPC
Diatoms, prymnesiophytes, dinoflagellates, chrysophytes, and raphidophytes
DT
Diatoxanthin
PPC
Diatoms, prymnesiophytes, dinoflagellates, chrysophytes, and raphidophytes
Fuco
Fucoxanthin
PSC
Diatoms, prymnesiophytes, chrysophytes, pelagophytes, and dinoflagellates type 2a (lacking peridinin)
Chl aHPLC
HPLC-derived chlorophyll a
PPP
All phytoplankton except Prochlorococcus
Peri
Peridinin
PSC
Dinoflagellates type 1a
Pras
Prasinoxanthin
PPC
Prasinophytes type 1b
Viola
Violaxanthin
PPC
Chlorophytes, prasinophytes, and eustigmatophytes
Zea + Lut
Zeaxanthin + lutein
PPC
Cyanobacteria, Prochlorococcus, chlorophytes, and prasinophytes type 2b
TChl a
Total chlorophyll a derived from HPLC analysis
Chl aHPLC + Chlide a
TC
Total carotenoids
Include all carotenoids
But-fuco + Hex-fuco + Allo + α-Car + β-Car + DD + DT + Fuco + Peri + Pras + Viola + Zea + Lut
AP
Accessory pigments
Include all pigments except TChl a
TC + Chl b + Chl c1 + c2 + Chl c3
According to Jeffrey et al. (1997) or a Higgins et al. (2011) or
b Vidussi et al. (2004).
CHEMTAX analysis
The CHEMTAX software (Mackey et al., 1996) was used
to estimate the relative abundance of distinct micro-algal groups to total
chlorophyll a from in situ pigment measurements. The software utilizes a
factorization program that uses “best guess” ratios of accessory pigments
to chlorophyll a that are derived for different groups from the literature
available and marker pigment concentrations of algal groups that are known
to be present in the study area, as reported in Fragoso et al. (2016). The
program uses the steepest descent algorithm to obtain the best fit to the
data based on assumed ratios of pigment to chlorophyll a (for more detail, see
Mackey et al., 1996). Because CHEMTAX is sensitive to the seed values of the
initial ratio matrix (Latasa, 2007), we used a later
version (v1.95) to obtain the more stable output matrices. In this CHEMTAX
version, the initial matrices are optimized by generating 60 further pigment
ratio tables using a random function (RAND in Microsoft Excel) as described
in Wright et al. (2009).
The results of the six best output matrices (with the smallest residuals,
equivalent to 10 % of all matrices) were used to calculate the averages
of the abundance estimates and final pigment ratios.
One of the main assumptions of the CHEMTAX method is that information about
the phytoplankton taxonomy is used to assure that the pigment ratios are
applied and interpreted correctly (Irigoien et al., 2004). To
satisfy this requirement, initial pigment ratios were carefully selected and
applied to each cluster to adjust the pigments to the appropriate groups, according to previous microscopic observations
(Fragoso et al.,
2016) and literature information (see Table 3). Pigment ratio tables were
based on the literature in waters having comparable characteristics to the
Labrador Sea, such as Baffin Bay
(Vidussi et al., 2004), the Beaufort
Sea (Coupel et al., 2015), and the North
Sea
(Antajan
et al., 2004; Muylaert et al., 2006) or from surface (high light) field data
(Higgins et al., 2011) (Table 3). High light field ratios
were chosen because samples were collected from surface waters during May
and June, when photosynthetic active radiation (PAR) was high (daily
incident irradiance averaged per month > 30 mol PAR m-2 d-1, Harrison et al., 2013). The pigments chosen for CHEMTAX
analysis were: 19-butanoyloxyfucoxanthin (But-fuco),
19-hexanoyloxyfucoxanthin (Hex-fuco), alloxanthin (Allo), total chlorophyll
a derived from HPLC analysis (TChl a, see Table 2), chlorophyll b (Chl b),
chlorophyll c3 (Chl c3), fucoxanthin (Fuco), peridinin (Peri),
prasinoxanthin (Pras), and zeaxanthin + lutein (Zea + Lut). Zeaxanthin
(Zea) and lutein (Lut) are two different pigments that co-eluted as a single
peak by the methods of pigment analyses applied in this study.
Initial ratio matrix of accessory pigment to chlorophyll a for distinct algal groups for each cluster
group. * Rf refers to the literature where the pigment ratios were extracted.
See explanation of each group in the methods section.
Region
I and II (eastern Labrador Sea)
Group/pigment
Chl b
Chl c3
Fuco
Peri
Zea + Lut
Allo
But-fuco
Hex-fuco
Pras
TChl a
Rf*
Prasinophyte 1
0.512
0
0
0
0
0
0
0
0.075
1
2
Prasinophyte 2
0.738
0
0
0
0.008
0
0
0
0
1
2
CHLORO-1
0.339
0
0
0
0.047
0
0
0
0
1
4
Dinoflagellates
0
0
0
0.600
0
0
0
0
0
1
5
Cryptophytes
0
0
0
0
0
0.673
0
0
0
1
2
Phaeocystis
0
0.208
0.350
0
0
0
0
0
0
1
1
HAPTO-6
0
0.155
0.195
0
0
0
0.019
1.054
0
1
4
Chryso/pelagophyte
0
0.114
0.398
0
0
0
0.595
0
0
1
2
Cyanobacteria
0
0
0
0
0.232
0
0
0
0
1
3
Diatoms
0
0
1.229
0
0
0
0
0
0
1
2
Region
III and V (central Labrador Sea)
Group/pigment
Chl b
Chl c3
Fuco
Peri
Zea + Lut
Allo
But-fuco
Hex-fuco
Pras
TChl a
Rf*
Prasinophyte 1
0.512
0
0
0
0
0
0
0
0.075
1
2
Prasinophyte 2
0.738
0
0
0
0.008
0
0
0
0
1
2
CHLORO-1
0.339
0
0
0
0.047
0
0
0
0
1
4
Dinoflagellates
0
0
0
0.600
0
0
0
0
0
1
5
Dino-2
0
0.179
0.300
0
0
0
0.081
0.194
0
1
4
Cryptophytes
0
0
0
0
0
0.673
0
0
0
1
2
HAPTO-6
0
0.155
0.195
0
0
0
0.019
1.054
0
1
4
Chryso/Pelagophyte
0
0.114
0.398
0
0
0
0.595
0
0
1
2
Cyanobacteria
0
0
0
0
0.232
0
0
0
0
1
3
Diatoms
0
0
1.229
0
0
0
0
0
0
1
2
Region
IV (western Labrador Sea)
Group/pigment
Chl b
Chl c3
Fuco
Peri
Zea + Lut
Allo
But-fuco
Hex-fuco
Pras
TChl a
Rf*
Prasinophyte 1
0.512
0
0
0
0
0
0
0
0.075
1
2
Prasinophyte 2
0.738
0
0
0
0.008
0
0
0
0
1
2
CHLORO-1
0.339
0
0
0
0.047
0
0
0
0
1
4
Dino-2
0
0.179
0.300
0
0
0
0.081
0.194
0
1
4
Dinoflagellates
0
0
0
0.600
0
0
0
0
0
1
5
Cryptophytes
0
0
0
0
0
0.673
0
0
0
1
2
Prymnesiophyte 1
0
0.038
0.416
0
0
0
0
1.108
0
1
2
Chryso/Pelagophyte
0
0.114
0.398
0
0
0
0.595
0
0
1
2
Diatoms
0
0
1.229
0
0
0
0
0
0
1
2
1 Antajan et al. (2004), 2 Vidussi et al. (2004), 3 Muylaert et al. (2006), 4 Higgins et al. (2011), 5 Coupel et al. (2015)
The other main requirement of the CHEMTAX method is that pigment ratios
remain constant across the subset of samples that are being analysed
(Mackey et al., 1996). To satisfy this assumption, a priori
analysis was performed, where pigment data were sub-divided into groups
using cluster analysis (Bray–Curtis similarity; PRIMER-e V7, see Sect. 2.7) and each group was processed separately by the CHEMTAX program (Table 3;
for the final ratio matrix, see Supplement). This approach was
used as distinct phytoplankton communities have been previously observed in
the Labrador Sea
(Fragoso et al., 2016), so the ratio of accessory pigment to chlorophyll a likely varies
within different water masses across the Labrador Sea (LC, IC and WGC).
Absolute concentrations of selected pigments (But-fuco, Hex-fuco, Allo, Chl b, Chl c3, Fuco, Peri, Pras, and Zea + Lut) were fourth-root transformed
and standardized (converted to %) before being analysed. Due to the high
abundance of diatoms in the data, we decided to apply a fourth-root
transformation to increase the importance of less abundant groups, which
would allow us to better discern the spatial–temporal patterns of the
phytoplankton communities in the Labrador Sea.
An initial cluster analysis on the select pigment data identified five major
groups having 60 % similarity between samples. Clusters included stations
partially located: (1) on the shelves, where Fuco dominated at a few stations
(I); (2) in the eastern part of the Labrador Sea, where most stations had
high relative concentrations of Fuco and Chl c3 (II); (3) in the central
Labrador Sea, where a few stations had high proportions of Fuco, Hex-fuco,
and Peri (III); (4) on the western part of the section, where Chl b and Fuco
were the main pigments at most stations (IV); and (5) in the central Labrador
Sea, where most stations had a mixture of pigments (Fuco, Chl c3,
Hex-fuco, Chl b, Peri, and others) (V) (Fig. S1, Supplement).
Prasinophytes were separated into “prasinophyte type 1”, which contains
Pras, and “prasinophyte type 2”, such as Pyramimonas and Micromonas, with the latter previously
found lacking Pras and containing Zea + Lut in the North Water Polynya
(Canadian Arctic) (see Vidussi et
al., 2004). Both genera were observed in light microscope counts in Labrador
Sea samples (G. M. Fragoso, personal observation, 2015),
M. pusilla has been observed in the Beaufort Sea
(Coupel et al., 2015), and was found to
be one of the main pico-eukaryotes in the North Water Polynya from April to
July of 1998 (Lovejoy et al., 2002). Zea is not only found
in “prasinophytes type 2”, but is also the major accessory pigment of
cyanobacteria (such as Synechococcus spp.), which have been observed in the Labrador Sea
(particularly in Atlantic waters; Li et al.,
2006). Zea is also a minor pigment in chlorophytes, while Lut is often the
dominant carotenoid in this group
(MacIntyre et al., 2010;
Vidussi et al., 2004). Due to their association with the warmer Atlantic
waters, cyanobacteria were assumed to be absent from very cold waters, such
as the Labrador Current
(see Fragoso
et al., 2016). Prasinophytes contain Chl b, as well as chlorophytes
(Vidussi et al., 2004) which were
observed in large numbers with the microscope (G. M. Fragoso, personal observation, 2015).
Dinoflagellates were separated into species that contain Peri (Heterocapsa sp. and
Amphidium; Coupel et al., 2015; Higgins
et al., 2011), and those that do not (Gymnodinium spp.; herein defined as
“dinoflagellates type-2” (Dino-2) according to Higgins et al., 2011) and
may contain Chl c3, But-fuco, Hex-fuco, and Fuco. Dinoflagellates were
observed in lower concentrations in the eastern Labrador Sea (Fragoso et
al., 2016), so that “Dino-2” was assumed absent from this area (clusters I
and II in Table 3). Cryptophytes (Table 3) are the only group to contain
Allo.
Prymnesiophytes were divided into three groups: (1) Phaeocystis pouchetii, which was observed in
high concentrations in the eastern Labrador Sea
(Fragoso et al.,
2016) (clusters I and II, Table 3); (2) “Prymnesiophyte 1” (as
in Vidussi et al., 2004),
associated with Chrysochromulina spp. and observed in the western Labrador Sea (Labrador
Current, this study) (cluster IV, Table 3); and (3) “HAPTO-6” (as in
Higgins et al., 2011), which included the
coccolithophores, particularly Emiliania huxleyi associated with Atlantic waters
(central-eastern region of the Labrador Sea) (clusters I, II, III and V,
Table 3). Phaeocystis pouchetii occurred in waters having low Hex-fuco and But-fuco
concentrations and high Chl c3 and Fuco concentrations (cluster II, Fig. S1, Supplement). Similar pigment compositions were found in
Phaeocystis globosa blooms in Belgian Waters
(Antajan
et al., 2004; Muylaert et al., 2006) and high ratios of Chl c3 to
Chl aHPLC (HPLC-derived chlorophyll a) have been used to identify
Phaeocystis pouchetii in the Labrador Sea (see
Stuart et al., 2000). Thus, Chl c3 and Fuco were the only pigments that
could be used to represent Phaeocystis. In addition to Chl c3 and Fuco,
“Prymnesiophyte 1” included Hex-fuco, while “HAPTO-6” included Hex-fuco
and But-fuco as in Higgins et al. (2011). Chrysophytes
and pelagophytes (such as Dictyocha speculum) have high ratios of But-fuco to Chl aHPLC
(Coupel et al., 2015; Fragoso
and Smith, 2012), and finally diatoms were identified as containing high
Fuco: Chl aHPLC ratios (Vidussi et
al., 2004) (Table 3).
Photosynthesis versus irradiance incubations
Water samples were spiked with 14C bicarbonate and incubated in a light
box under 30 different irradiance levels (from 1–600 W m-2) at in situ
temperature for 2 to 3 h to measure parameters derived from
photosynthesis versus irradiance (P–E) curves as described by
Stuart et al. (2000).
Measurements were fitted to the equation of Platt and Gallegos
(1980) to determine photosynthetic efficiency (αB), the maximum
photosynthetic rate normalized to chlorophyll biomass (PmB), the
light intensity approximating the onset of saturation (Ek), the
saturation irradiance (Es), and the photoinhibition parameter (β).
Statistical analysis
Fragoso et al. (2016) found a significant linear relationship between
phytoplankton carbon, calculated from phytoplankton cell counts, and POC
data using results from 2011–2014 surveys in the Labrador Sea (i.e. POC = 1.01 POCphyto+ 240.92; r2=0.47; n=44; p<0.0001). To estimate phytoplankton-derived carbon (POCphyto)
concentrations (as opposed to total POC, which includes detritus and
heterotrophic organisms), regression analysis was performed using the carbon
calculated from cell counts (derived from
Fragoso et
al., 2016) and measurements of fluorometric chlorophyll a (Chl af). This
regression (POCphyto= 38.9 Chl af; r2=0.9; n=41;
p < 0.0001) was then applied to estimate POCphyto for stations
where phytoplankton cell counts were not available (2005–2010).
Phytoplankton community structure derived from pigment concentrations was
investigated using PRIMER-E (v7) software (Clarke and
Warwick, 2001). Chlorophyll a concentrations derived for each algal group
resulting from CHEMTAX analysis were standardized (converted to percentage
values) to obtain their relative proportions, which were fourth-root
transformed to allow the least abundant groups to contribute to the
analysis. Similarity matrices were generated from Bray–Curtis similarity for
cluster analysis. A SIMPER (SIMilarity PERcentages) routine with a cut-off
of 90 % cumulative contribution to the similarity was used to reveal the
contributions of each group to the overall similarity within clusters.
One-way ANOSIM was also applied to determine whether taxonomic compositions
of the clusters were significantly different.
A redundancy analysis (RDA) using the CANOCO 4.5 software (CANOCO,
Microcomputer Power, Ithaca, NY) was performed to analyse the effects of
different environmental factors on the Labrador Sea phytoplankton community
structure (see also
Fragoso et
al., 2016). Data were log-transformed and forward-selection (a posteriori analysis)
identified the subset of environmental variables that significantly
explained the taxonomic distribution and community structure when analysed
individually (λ1, marginal effects) or when included in a
model where other forward-selected variables were analysed together
(λa, conditional effects). A Monte Carlo permutation test (n=999, reduced model) was applied to test the statistical significance (p < 0.05) of each of the forward-selected variables.
Results
Environmental variables
Map with sampling stations and distances from a fixed reference
position (Northeast Gulf of St Lawrence) in the x axis shown by the star
(a). Values are given at individual stations sampled between 2005 and 2014
(y axis) for the following variables: date of sample collection (b),
temperature (c), salinity (d), stratification index (SI) (e), chlorophyll
a (f), nitrate (NO3-) (g), phosphate (PO43-) (h), silicate
(Si(OH)4) concentrations (i), ratios of particulate organic carbon
(POC) to particulate organic nitrogen (PON) (j), silicate to nitrate
(Si(OH)4 : NO3-) ratios (k), and nitrate to phosphate
(NO3- : PO43-) ratios (l). LSh, Labrador Shelf; LSl,
Labrador Slope; CB, Central Basin; GSl, Greenland Slope; GSh,
Greenland Shelf.
Sampling dates varied from May to June during this 10-year study, where
samples from 2007, 2011 and 2013 were collected in early May, as opposed to
samples from 2012 and 2104, which were collected later in the season (mid- to
late June) (Fig. 2b). Environmental parameters, as well as fluorometric
chlorophyll a (Chl af) concentrations varied noticeably along the
southwest–northeast section of the Labrador Sea (Fig. 2c–l). The shelf and
slope regions (LSh, LSl, GSl, GSh) had colder and fresher waters (< 3 ∘C and < 33.5, respectively) compared to the central
basin (CB), where surface waters were saltier (> 33.5) and warmer
(> 3 ∘C), particularly in 2005, 2006, 2012, and 2014
(> 5 ∘C) (Fig. 2c, d). Shelf waters that were the
coldest and freshest were also the most highly stratified ((stratification
index (SI) > 5 × 10-3 kg m-4), particularly on
the Labrador Shelf (SI > 15 × 10-3 kg m-4),
whereas waters from the CB were less stratified (SI < 5 × 10-3 kg m-4), apart from at stations collected later in the
season (Fig. 2b), where waters were slightly warmer than usual (> 5 ∘C) in 2005, 2012, and 2014 (Fig. 2e). Chl af concentrations
were highest (> 4 mg Chl af m-3) at stations where waters
were highly stratified, particularly on the shelves (Fig. 2f). Nitrate,
phosphate, and silicate concentrations were inversely related to Chl af
concentration, being lowest (< 5, 0.5, and 3 µmol L-1,
respectively) on the shelves, and during some years in the CB (e.g. 2012),
where blooms formed (Fig. 2f–i). POC : PON ratios were > 8 at most
stations in shelf and slope waters and at a few stations in the CB during
2009 and 2011 (Fig. 2j). Shelf waters mostly had higher silicate : nitrate
(Si(OH)4 : NO3-) ratios (> 1) than the CB,
particularly in the LSh (Fig. 2k). Labrador Sea surface waters usually had
nitrate : phosphate (NO3- : PO43-) less than 16, although
NO3- : PO43- were relatively higher in the CB than in the
shelf regions (> 10) (Fig. 2l).
CHEMTAX interpretation and group distributions
Relative contribution (%) of chlorophyll a from distinct phytoplankton groups at each
station from 2005 to 2014 along the section distance from Labrador coast
represented in Fig. 2a (star symbol). LSh, Labrador Shelf; LSl,
Labrador Slope; CB, Centre Basin; GSl,
Greenland Slope; GSh, Greenland Shelf. Note the distinct scales
for each group.
Diatoms were the most abundant phytoplankton group found in the Labrador
Sea, particularly at some stations on the shelves where they dominated
almost 100 % of the total phytoplankton community (Fig. 3a). Chlorophytes
and prasinophytes were common in the central-western part (Fig. 3b, c),
whereas Phaeocystis was highest at the eastern part of the Labrador Sea (Fig. 3d).
Dinoflagellates were abundant in the central region of the Labrador Sea
(Fig. 3e). Other prymnesiophytes, including coccolithophores and
Chrysochromulina, were also common in the central part of the Labrador Sea (Fig. 3f).
Overall, chrysophytes and pelagophytes were found in low abundances in the
Labrador Sea, except at the central region of the Labrador Sea during 2011
(Fig. 3g). Cyanobacteria were most abundant at the Labrador Slope and
Greenland Shelf, and during some years (2005 and 2012) in the central
Labrador Sea (Fig. 3h). Cryptophytes comprised less than 10 % of total
phytoplankton chlorophyll concentrations (data not shown).
Dendrogram showing clustering of samples (a) and the proportion of
chlorophyll a contributed by each phytoplankton group for each cluster (b).
Spatial distribution of distinct phytoplankton communities (cluster groups)
along the section, showing the distance from the star in Fig. 2a) (c).
Bubble size in (c) represents total chlorophyll a biomass (minimum 0.3 mg Chl af m-3 and maximum 25 mg Chl af m-3).
A cluster analysis of algal groups derived from CHEMTAX results revealed
clusters of stations at various similarity levels (Fig. 4). Pairwise
one-way analysis of similarity (ANOSIM) between clusters suggested that they were
significantly different in terms of algal pigment composition (p=0.001).
However, pairwise analysis of clusters C3a and C3b showed that these groups
were more similar in composition (R statistic of 0.33) than other clusters
(R statistic values approached 1) (see Clarke and Warwick, 2001). The first
division occurred at 61 %, separating three main clusters (A, B, and C)
(Fig. 4a). Cluster C was subdivided at 65 % resulting in clusters C1, C2,
and C3 (Fig. 4a). A third division (similarity of 73 %) occurred at
cluster C3 resulting in two other clusters C3a and C3b (Fig. 4a). Overall,
six functional clusters (A, B, C1, C2, C3a, and C3b) represented the distinct
phytoplankton communities occurring in the Labrador Sea (Fig. 4a). These
communities generally occupied different regions of the Labrador Sea, namely
the Labrador Shelf/Slope (west, Cluster C1 and, mainly, Cluster C3a), the
Central Basin (middle, mainly Clusters C2 or C3b), and the Greenland
Shelf/Slope (east, mainly Clusters C3a, A, B) (Fig. 4b, c).
Chl af concentrations were high at stations where diatoms were dominant
(Fig. 4b, c). Diatoms were the most abundant phytoplankton group in Labrador
Sea waters, particularly at stations on the shelves, where communities were
sometimes composed of almost 100 % diatoms (clusters A and C1) (Fig. 4b, c).
Diatoms were also abundant at (or near to) the Greenland Shelf, where
Phaeocystis was co-dominant (cluster B) and at (or near to) the Labrador Shelf in the
west section, where chlorophytes were the second most abundant group
(cluster C3a). Likewise, diatoms were dominant in the central Labrador Sea
in some years (2008, 2012, and 2014, cluster C2), where dinoflagellates were
also dominant (Fig. 4b, c). Most stations in the central basin had low
Chl af concentrations and high diversity of algal groups (cluster C3b),
with mixed assemblages of diatoms, dinoflagellates, and other flagellates
(Fig. 4b, c). The positions of oceanographic fronts, usually characterized by
sharp transitions in phytoplankton communities, varied from year to year but
were generally located near the continental slopes (Fig. 4c).
Phytoplankton distributions and environmental controls
Positions of individual stations in relation to temperature
(∘C) and salinity (a) and redundancy analysis (RDA) ordination
plot (b). The stations are colour-coded according to the cluster groups (see
details in Fig. 4). The TS plot (a) shows the approximate ranges of
potential temperature (∘C) and salinity of the Labrador Current
(LC), the West Greenland Current (WGC) and the Irminger Current (IC). Arrows
in (b) show the explanatory (environmental) variables used in the analysis.
Distributions of surface phytoplankton communities defined above varied
according to the water mass distributions across the shelves and central
basin of the Labrador Sea. Potential temperatures and salinities also varied
among these water masses (Fig. 5a). In general, a community dominated by
chlorophytes and diatoms (cluster C3a) was associated with the inshore
branch of the Labrador Current (LC) on the Labrador Shelf. Surface waters
from the LC were the coldest (temperature < 2 ∘C) and
least saline, with the lowest density (σO of most stations
approximately < 26.5 kg m-3) of all the surface water masses of
the Labrador Sea (Fig. 5a). Mixed assemblages (cluster C3b), as well as
blooms (chlorophyll average, 4 mg Chl af m-3) of dinoflagellates
and diatoms (cluster C2) were associated with the Atlantic water mass, the
Irminger Current (IC) (Fig. 5a). These were the warmest (temperature
> 3 ∘C), saltiest (salinity > 34), and
densest (σO of most stations > 27 kg m-3)
surface waters of the Labrador Sea (Fig. 5a). A community dominated by
diatoms and Phaeocystis (cluster B) occurred in waters of the West Greenland Current
(WGC), which had intermediate temperatures (mostly 0–4 ∘C) and
salinities (33–34.5) when compared to those of the LC and IC (Fig. 5a).
Redundancy analysis (RDA) was used to investigate the hydrographic variables
that explained the variance (explanatory variables) in the phytoplankton
communities identified from pigment analyses. The ordination diagram
revealed that stations from each distinct cluster are concentrated in
different quadrants (Fig. 5b), with the arrows in the ordination diagram
representing the environmental variables. Positive or negative correlations
indicate that the arrows are orientated parallel to the distribution of
cluster stations (same direction, positive; opposite direction,
negative correlations), with the strength of the correlation proportional to
the arrow length. Table 4a indicates that the first axis (x axis) of the
redundancy analysis explained most of the variance (83.5 % of
species–environment relationship; taxa–environmental correlation of 0.68).
Summed, the canonical axes explained 99.8 % of the variance (axis 1, p=0.002; all axes, p=0.002) (Table 4a), which indicates that the
environmental variables included in this analysis explained almost 100 %
of the variability. Forward selection showed that five of the six
environmental factors (silicate, temperature, salinity, nitrate, and
phosphate) included in the analysis best explained the variance in
phytoplankton community composition when analysed together (p< 0.05,
Table 4b). When all variables were analysed together (conditional effects,
referred to as λa in Table 4b), silicate concentration was the
most significant explanatory variable (λa=0.2, p=0.001), followed by temperature (λa=0.05, p=0.001),
salinity (λa=0.02, p=0.002), nitrate concentration
(λa=0.01, p=0.016) and phosphate concentration
(λa=0.02, p=0.002) (Table 4). Stratification index
(SI) was the only explanatory variable that had no statistical
significance in explaining the distribution of phytoplankton communities
(Table 4b).
Results of the redundancy analyses (RDA) with the eigen-values,
taxa–environmental correlations and percentages of variance explained used
in the analysis (a). Automatic forward selection (a posteriori analysis) was used to
determine the environmental variable(s) that best explain the variance of
the data (b). The subset of environmental variable(s) that significantly
explained phytoplankton distribution are referred to as marginal effects
(λ1) when analysed individually, or conditional effects
(λa) when analysed additively in the model (b). Explanatory
variables are temperature (∘C), salinity, nitrate
(NO3-; µmol L-1), phosphate (PO43-; µmol L-1), silicate (Si(OH)4; µmol L-1), and
stratification index (SI) (kg m-4). Significant p values (p < 0.05) represents the variables that explain the variation in the analyses.
(a) Axes
1
2
3
4
Total variance
Eigen-values
0.26
0.04
0.005
0
1
Taxa–environment correlations
0.68
0.4
0.321
0.25
Cumulative percentage variance
of species data
25.7
29.9
30.3
30.7
of species–environment relation
83.5
97.2
98.8
99.8
Sum of all eigenvalues
1
Sum of all canonical eigenvalues
0.31
(b) Marginal effects
Conditional effects
Variable
λ1
Variable
λa
p
F
Si(OH)4
0.2
Si(OH)4
0.2
0.001
61.7
NO3-
0.19
Temperature
0.05
0.001
17.3
PO43-
0.17
Salinity
0.02
0.002
6.94
Salinity
0.09
NO3-
0.01
0.016
4.31
Temperature
0.07
PO43-
0.02
0.002
7.22
SI
0.06
SI
0.01
0.153
1.72
The first axis (x axis) of the analysis, which explained most of the
variance, clearly shows that the phytoplankton communities are associated
with environmental parameters (Fig. 5b). Thus, stations in Arctic waters
were to the left of the y axis (low nutrients, temperatures, and salinity
values), while stations located in Atlantic waters were to the right
(opposite trend, Fig. 5b). A community dominated by diatoms and chlorophytes
(cluster C3a, upper left quadrant of Fig. 5b) was associated with lower
salinities and temperatures, and highly stratified waters. Another community
dominated by Phaeocystis and diatoms (cluster B, lower left quadrant of Fig. 5b) was
associated with waters where nutrient concentrations (mainly nitrate, but
also phosphate and silicate) were relatively low (average nitrate
concentration for cluster B < 3 µmol L-1, Table 5). In Atlantic
waters (upper and lower right quadrants (Fig. 5b)), the phytoplankton
community was composed of mixed taxa during May (orange circles), but became
dominated by diatoms and dinoflagellates during the bloom in June (red
circles), showing a clear temporal succession in these waters. Thus, mixed
assemblages (cluster C3b) were associated with higher nutrient
concentrations (pre-bloom conditions in Atlantic waters, upper right
quadrant), whereas dinoflagellates and diatoms (cluster C2) were associated
with warmer and saltier waters, resembling bloom conditions in Atlantic
waters induced by thermal stratification (lower right quadrant of Fig. 5b).
Average, standard deviations and number of observations (in
parentheses) of environmental and biological variables of each cluster
group. MLD, mixed-layer depth; SI, stratification index; NO3-, nitrate; PO43-, phosphate; Si(OH)4, silicate;
DT, diatoxanthin; DD, diadinoxanthin; POC, particulate organic carbon;
PON, particulate organic nitrogen; POCphyto, phytoplankton-derived
particulate organic carbon; αB, initial slope of the
photosynthesis-irradiance curve; PmB, maximum normalized
photosynthesis; Ek, onset saturation irradiance; Es, saturation
irradiance.
Cluster A
Cluster B
Cluster C3a
Cluster C3b
Cluster C2
Cluster C1
DIAT (> 99 %)
DIAT + PHAEO
DIAT + CHLORO
MIXED
DIATO + DINO
DIAT (> 93 %)
Temperature (∘C)
2.8 ± 2.4
(17)
2.0 ± 1.8
(46)
1.6 ± 1.9
(62)
3.4 ± 1.9
(92)
4.8 ± 1.5
(32)
1.4 ± 1.7
(4)
Salinity
33.4 ± 1.5
(17)
33.7 ± 0.8
(46)
33.1 ± 1.2
(62)
34.1 ± 1.0
(92)
34.4 ± 0.5
(32)
33.0 ± 1.6
(4)
MLD (m)
32.2 ± 43.8
(17)
32.6 ± 23.4
(46)
31.2 ± 28.5
(62)
59 ± 71.1
(92)
29.8 ± 17.0
(32)
16.0 ± 4.2
(4)
SI × 10-3 (kg m-4)
9.1 ± 6.3
(17)
6.3 ± 5.7
(46)
10.7 ± 8.5
(62)
5.0 ± 6.8
(92)
6.1 ± 4.5
(31)
6.6 ± 8.5
(4)
NO3- (µmol L-1)
2.9 ± 4.7
(17)
2.7 ± 3.5
(46)
3.4 ± 4.3
(58)
8.4 ± 4.1
(83)
3.7 ± 3.9
(32)
3.8 ± 6.8
(4)
Si(OH)4 (µmol L-1)
2.2 ± 2.7
(17)
2.8 ± 2.1
(46)
3.5 ± 2.4
(58)
5.4 ± 2.2
(83)
3.0 ± 2.2
(32)
2.3 ± 3.4
(4)
PO43- (µmol L-1)
0.3 ± 0.3
(17)
0.3 ± 0.2
(45)
0.4 ± 0.2
(55)
0.7 ± 0.2
(79)
0.3 ± 0.2
(32)
0.4 ± 0.3
(4)
Si(OH)4 : NO3-
6.0 ± 11.8
(14)
3.6 ± 7.9
(37)
8.5 ± 18.2
(54)
1.1 ± 1.5
(82)
1.6 ± 1.8
(32)
3.9 ± 4.4
(4)
NO3- : PO43-
8.2 ± 6.7
(11)
5.2 ± 5.0
(45)
5.9 ± 5.8
(55)
11.4 ± 4.1
(79)
8.7 ± 4.6
(32)
5.5 ± 7.1
(4)
Chlorophyll a (mg Chl af m-3)
3.8 ± 4.7
(17)
5.5 ± 4.8
(45)
7.7 ± 5.6
(59)
2.0 ± 1.7
(91)
4.0 ± 1.8
(31)
8.8 ± 9.6
(4)
DT : (DT+DD)
0.01 ± 0.03
(16)
0.02 ± 0.05
(44)
0.04 ± 0.05
(62)
0.10 ± 0.01
(92)
0.08 ± 0.07
(32)
0.02 ± 0.04
(4)
(DD+DT) : TChla
0.08 ± 0.07
(17)
0.03 ± 0.03
(46)
0.04 ± 0.02
(62)
0.07 ± 0.03
(92)
0.12 ± 0.03
(32)
0.07 ± 0.04
(4)
POC (mg C m-3)
245 ± 90
(4)
498 ± 198
(27)
533 ± 198
(45)
234 ± 145
(63)
512 ± 179
(15)
393 ± 418
(2)
PON (mg N m-3)
39 ± 16
(4)
65 ± 23
(27)
74 ± 30
(45)
38 ± 26
(64)
83 ± 33
(15)
42 ± 41
(2)
POCphyto (%)
23.0 ± 5.2
(4)
49.2 ± 29.5
(26)
60.9 ± 25.6
(44)
33.3 ± 10.1
(64)
36.0 ± 11.4
(15)
37.8 ± 1.3
(2)
POC : PON
6.5 ± 1.2
(4)
7.8 ± 2.1
(27)
7.5 ± 2.1
(45)
6.6 ± 1.3
(64)
6.2 ± 0.9
(15)
8.6 ± 1.6
(2)
αB×10-2 (mgC [mg Chl af] h-1 [W m-2]-1)
–
6.8 ± 6
(9)
9.2 ± 10
(10)
7.1 ± 4
(18)
7.1 ± 1.5
(4)
–
PmB (mgC [mg Chl af] h-1)
–
3.0 ± 1.2
(9)
2.3 ± 0.8
(10)
2.3 ± 0.6
(18)
3.3 ± 0.7
(4)
–
Ek (W m-2)
–
60 ± 33
(9)
29 ± 13
(10)
39 ± 14
(18)
46 ± 5
(4)
–
Es (W m-2)
–
62 ± 32
(9)
35 ± 18
(10)
43 ± 18
(18)
56 ± 8
(4)
–
β×10-4 (mgC [mg Chl af] h-1[W m-2]-1)
–
4 ± 7
(9)
16 ± 23
(10)
10 ± 16
(18)
29 ± 24
(4)
–
Phytoplankton distribution and elemental stoichiometry
Particulate organic carbon (POC) collected on filters can include organic
carbon from a variety of sources, such as phytoplankton, bacteria,
zooplankton, viruses, and detritus (Sathyendranath et al., 2009). Assuming
that phytoplankton-associated organic carbon, as estimated from
phytoplankton cell volumes (POCphyto) is strongly correlated with
chlorophyll a values, the proportion of POCphyto should increase in
eutrophic waters, which usually occurs with high chlorophyll a and POC
concentrations, and that it should be lower in oligotrophic waters. Indeed,
our results showed higher proportions of POCphyto (> 60 %) in waters with higher POC concentrations (Fig. 6a). However, there
were several stations where POC levels were high and where the contribution
of POCphyto was low, suggesting that there may have been other sources
of POC (e.g. detritus).
Relationship between particulate organic carbon (POC) and
particulate organic nitrogen (PON) in a logarithmic scale, with the points
(stations) as a function of phytoplankton-derived organic carbon content
(POCphyto/ POC, %) (a), POC : PON versus salinity (b), phytoplankton-derived
organic carbon content (POCphyto / POC, %) versus the POC : PON ratio (c). The
points (stations) in (b) and (c) are colour-coded according to the cluster
groups (see details in Fig. 4). Solid lines in (b) and (c) show the C : N
Redfield ratio of 6.6 and the dashed line in (c) shows where POCphyto
contributes 50 % of the total POC.
To investigate the influence of phytoplankton community structure on the
stoichiometry of particulate organic material of surface Labrador Sea
waters, the relationships between POCphyto (the estimated proportion of
POC from phytoplankton) and the ratio of POC to PON were examined. In
general, different phytoplankton communities had distinct relationships
between POCphyto and POC : PON. Stations in shelf regions, which have
higher inputs of Arctic and glacial melt waters (lower salinity values),
where diatoms co-dominated with chlorophytes in the west and east (cluster
C3a) or with Phaeocystis in the east (cluster B), had higher and more variable values
for POC : PON ratios than did stations influenced by Atlantic water (Fig. 6b).
Some shelf stations had relatively high proportions of POCphyto to
total POC, suggesting that phytoplankton community growth dominated by
diatoms and chlorophytes (cluster C3a) contributed to a high proportion of
the total POC (most stations from cluster C3a had POCphyto > 50 %) (Fig. 6b). On the other hand, some shelf stations, particularly the
one dominated by a community composed of diatoms and Phaeocystis (cluster B), had high
POC : PON ratios (> 10), with low POCphyto contributions,
suggesting an increased contribution of detritus to the total POC (Fig. 6c).
Stations influenced by Atlantic waters had generally lower contributions of
POCphyto compared to Arctic-related waters, with most stations having
POC : PON ratios < 6.6 (Fig. 6c).
Physiological patterns
The linear relationship of accessory pigment (AP) versus total chlorophyll a
(TChl a) was investigated, given that it is often used as an index of
quality control in pigment analysis (Aiken et al., 2009).
This relationship could also represent a response of phytoplankton
communities to light conditions, given that AP allows a broader range of
wavelengths to be absorbed (chromatic adaptation), whereas TChl a
concentrations would vary according to light intensities (light–shade
adaptation) (Boyton et al., 1983). The log–log
linear relationship of AP versus total chlorophyll a
(TChl a) from surface waters of the Labrador Sea varied with temperature
(Fig. 7a) and among the distinct phytoplankton communities (Fig. 7b).
Phytoplankton communities in cold waters (of Arctic origin), such as those
co-dominated by diatoms and Phaeocystis in the east and diatoms and chlorophytes in the
west, had a lower ratio of accessory pigments to TChl a (logAP : logTChl a) (slope
of 0.86 and 0.89, respectively) than communities from warmer waters
(Irminger Current from Atlantic origin), particularly those co-dominated by
diatoms and dinoflagellates (logAP : logTChl a, slope of 1.03) (Fig. 7b).
Slopes of the logAP to logTChl a relationships were not statistically
different among the different communities (ANCOVA, p > 0.05),
except for those communities co-dominated by diatoms and Phaeocystis (cluster B), which
had a slope that was statistically different from the others (ANCOVA, p=0.016).
Relationship between total accessory pigments (mg AP m-3) and
total chlorophyll (mg TChl a m-3) on a logarithmic scale, with the points
(stations) according to temperature (a) and colour-coded according to
phytoplankton community cluster group (see details in Fig. 4) (b).
Photosynthetic parameters differed among the different phytoplankton
communities. Photosynthetic efficiencies (αB) were the lowest
in communities dominated by Phaeocystis and diatom communities in the east of the
transect (near Greenland, cluster B) (average αB=6.8 × 10-2 mg C [mg Chl af] h-1 [W m-2]-1) and
the highest in communities dominated by diatoms and chlorophytes (cluster
C3a) typically found in the west (Labrador Current) (αB=9.2 × 10-2 mg C [mg Chl af] h-1 [W m-2]-1)
(Table 5). The light intensity approximating the onset of saturation
(Ek) had the opposite pattern: it was highest in communities dominated
by Phaeocystis and diatoms (average Ek= 60 ± 33 W m-2) and lowest at
stations dominated by diatoms and chlorophytes (Ek= 29 W m-2) (Table 5). Phaeocystis and diatom communities also showed little
photoinhibition (β=4 × 10-4 mg C [mg Chl af] h-1 [W m-2]-1). Phytoplankton communities in Atlantic waters
(clusters C3b and C2) had the highest levels of photoprotective pigments,
such as those used in the xanthophyll cycle (diadinoxanthin (DD) +
diatoxanthin (DT)) : TChl a > 0.07), particularly those communities
co-dominated by diatoms and dinoflagellates (cluster C2) from stratified
Atlantic waters (Table 5). These communities were the most susceptible to
photoinhibition (β= 29 × 10-4 mg C [mg Chl af] h-1 [W m-2]-1), had the highest ratios of
photoprotective pigments to TChl a ((DD+DT) : TChl a=0.12 ± 0.01), and
the highest maximum photosynthetic rates (PmB=3.3 ± 0.7 mg C [mg Chl af] h-1) (Table 5).
Discussion
Biogeography of phytoplankton communities in the Labrador Sea
In this study, our assessment of phytoplankton pigments from surface waters
of the Labrador Sea during spring/early summer are based on a decade of
observations and show that the distribution of phytoplankton communities
varied primarily within distinct waters masses in surface waters (Labrador,
Irminger and Greenland Currents). However, a temporal succession of
phytoplankton communities from the central region of the Labrador Sea was
also observed as waters became thermally stratified from May to June. Major
blooms (Chl af concentrations > 3 mg Chl af m-3)
occurred on or near the shelves in shallower mixed layers (< 33 m,
Table 5). Diatoms were abundant in these blooms; however, they often
co-dominated with (1) chlorophytes in the west (mostly in the Labrador
Current) and (2) Phaeocystis in the east in the West Greenland Current. A more diverse
community with low chlorophyll a values (average Chl afconcentrations
∼ 2 mg Chl af m-3, Table 5) was found earlier in the
season (May) in deeper mixed layers (> 59 m, Table 5) in the
central basin. Once these waters of the central basin became
thermally stratified (June), a third bloom co-dominated by diatoms and
dinoflagellates occurred, revealing an ecological succession from mixed
flagellate communities. These patterns are similar to those seen in other
shelf and basin regions of Arctic/subarctic waters (e.g. Coupel et al.,
2015; Fujiwara et al., 2014; Hill et al., 2005).
It is well known that diatoms tend to dominate in high-nutrient regions of
the ocean due to their high growth rates, while their low surface area to
volume ratios mean that they do not do as well as smaller nano- or
picoplankton in low-nutrient conditions (Gregg et al., 2003; Sarthou et
al., 2005). The Labrador Sea is a high-nutrient region during early spring
due to deep winter mixing (200–2300 m) that provides nutrients to the
surface layers. Thus, high nutrient concentrations may have supported the
blooms dominated by diatoms once light became available, as observed in
previous studies (Fragoso et al., 2016; Harrison et al., 2013; Yashayaev and
Loder, 2009).
Chlorophytes were the second most abundant phytoplankton group in this
study, particularly in the central-western part of the Labrador Sea, but
occasionally occurring in the east as well. Chlorophytes are thought to
contribute 1–13 % of total chlorophyll a in the global ocean
(Swan et al., 2015) and to inhabit
transitional regions, where nutrient concentrations become limiting for
diatoms but are not persistently low enough to prevent growth due to
nutrient limitation, as occurs in the oligotrophic gyres
(Gregg et al., 2003; Gregg and Casey, 2007;
Ondrusek et al., 1991). The Labrador Shelf is a dynamic region during
springtime, where melting sea ice in May provides a local freshwater input
(Head et al., 2003). Melting
sea ice provides intense stratification and shallow mixed layers for the
phytoplankton, with increased access to light, which promotes rapid
growth of cold Arctic/ice-related phytoplankton near the sea ice shelf
(Fragoso et al.,
2016). It is possible that the rapid nutrient exhaustion in highly
stratified ice-melt waters might have stimulated the growth of chlorophytes
as a succession from large diatoms to smaller phytoplankton forms.
Chlorophytes, as well as prasinophytes, such as Pyramimonas, a genus found in high
abundances in surface Labrador Shelf waters, has been previously associated
with land-fast (Palmer et al.,
2011) and melting sea ice, given that they have been found blooming
(chlorophyll a concentration ∼ 30 mg Chl af m-3) in low-salinity melt waters (salinity, 9.1) under the Arctic pack ice
(Gradinger, 1996).
Dinoflagellates, in this study, were associated with the Irminger Current,
where they were occasionally found blooming with diatoms in the warmer,
stratified Atlantic waters of the central basin. These blooms dominated by
dinoflagellates and Atlantic diatom species, such as Ephemera planamembranacea and Fragilariopsis atlantica, start later in
the season (end of May or June) as thermal stratification develops in the
central Labrador Sea (Frajka-Williams and
Rhines, 2010; Fragoso et al., 2016). Transition from diatoms to
dinoflagellates has been well documented in the North Atlantic between
spring and summer, and occurs mainly as dinoflagellates can use mixotrophic
strategies to alleviate nutrient limitation as waters become warmer, highly
stratified and nutrient depleted
(Barton
et al., 2013; Head et al., 2000; Head and Pepin, 2010; Henson et al., 2012;
Leterme et al., 2005). The North Atlantic Oscillation index (NAO) and sea
surface temperatures (Zhai et al., 2013) appear to
influence the relative proportions of diatoms and dinoflagellates as well as
the variability in the start date of the North Atlantic bloom. A negative
winter phase of NAO is associated with weaker northwest winds over the
Labrador Sea and reductions in the depth of winter mixing and supply of
nutrients to the upper layers (Drinkwater and Belgrano, 2003).
Vertical stability, thermal stratification, and the initiation of the spring
bloom tend to occur earlier under negative NAO conditions and the proportion
of dinoflagellates in the warmer, more nutrient-limited waters may be higher
(Zhai et al., 2013). Unfortunately, it was not possible to
investigate the influence of NAO on the relative contribution of
dinoflagellates and diatoms in the Labrador Sea section of the North
Atlantic in this study, given that the sampling period varied from
early/mid-May to mid-/late June. However, abundances of dinoflagellates
appeared to be higher in warmer waters (> 5 ∘C),
suggesting that the communities were shifting from diatoms to
dinoflagellates as the water became stratified and nutrient concentrations
decreased.
In this study, a community dominated by Phaeocystis and diatoms was observed blooming
together in waters of the WGC, in the eastern central part of the Labrador
Sea. The occurrence of Phaeocystis in these waters has been observed before by several
authors
(Fragoso
et al., 2016; Frajka-Williams and Rhines, 2010; Harrison et al., 2013; Head
et al., 2000; Stuart et al., 2000; Wolfe et al., 2000). The eastern part of
the Labrador Sea is a region with high eddy kinetic energy during spring
(Chanut et al., 2008;
Frajka-Williams et al., 2009; Lacour et al., 2015), which causes the
accumulation of low-salinity surface waters from the West Greenland Current.
This buoyant freshwater layer contains elevated levels of algal biomass of
both Phaeocystis and diatoms
(this study,
Fragoso et al., 2016). Mesoscale eddies may stimulate growth of
Phaeocystis and diatoms by inducing partial stratification at irradiance levels that
are optimal for their growth, but too low for their competitors (blooms in
these eddies usually start in April). Lacour et al. (2015)
showed that irradiance levels estimated from satellite-derived PAR and mixed-layer depth climatologies are similar for thermally and haline-stratified
spring blooms in the Labrador Sea. Nonetheless, these authors recognize the
need for in situ measurements to confirm whether Labrador Sea spring blooms,
presumably composed of distinctive phytoplankton communities, respond in the
same manner to light-mixing regimes. The ability of Phaeocystis to grow under dynamic
light irradiances explains why they are often found in deeper mixed layers,
such as those found in Antarctic polynyas
(Arrigo, 1999; Goffart et al., 2000), although
this genus can also occur in shallow mixed layers, such as those found close
to ice edges
(Fragoso
and Smith, 2012; Le Moigne et al., 2015).
Mesoscale eddies are also often associated with elevated zooplankton
abundances
(Frajka-Williams
et al., 2009; Yebra et al., 2009). In the Labrador Sea, lower grazing rates
have been observed in blooms dominated/co-dominated by colonial
Phaeocystis, which are often located in these eddies and which may, in turn, explain
why this species is dominant
(Head and
Harris, 1996; Wolfe et al., 2000). Although the exact mechanism that
facilitates Phaeocystis growth in the northeastern region of the Labrador Sea is not
clear, it is evident that blooms of this species are tightly linked to
mesoscale eddies, and that this relationship needs further investigation to
better explain their regular reoccurrence in these waters.
Phytoplankton composition and related biogeochemistry
Particulate organic carbon (POC) and nitrogen (PON) concentrations, as well
as the molar ratio of POC : PON varied within distinct hydrographic zones,
indicating the presence of different biogeochemical provinces in the
Labrador Sea. A canonical Redfield ratio of 6.6 for POC : PON appears to
represent the global average (Redfield, 1958), although regional
variations on the order of 15 to 20 % have also been reported
(Martiny et al., 2013b). The POC : PON
appears to be closer to the Redfield ratio of 6.6 in productive
subarctic/Arctic waters, such as the northern Baffin Bay
(Mei et al., 2005), the northeastern Greenland Shelf
(Daly et al., 1999), and in Fram Strait and the Barents
Sea (Tamelander et al., 2012). Crawford et
al. (2015), however, recently reported very low POC : PON ratios in
oligotrophic Arctic waters of the Beaufort Sea and Canada Basin, where
depth-integrated values of the POC : PON ratio were ∼ 2.65, much
lower than those in more productive domains, such as the subarctic central
Labrador Sea (POC : PON ∼ 4).
In this study, highly productive surface waters of Arctic origin (near or
over the shelves) had higher phytoplankton-derived particulate organic
carbon (POCphyto > 43 % of total POC, Fig. 6c), as well
as higher and more variable POC : PON ratios (average > 6.9, Fig. 6b) compared with stations influenced by Atlantic water (average POC : PON
< 6.3, POCphyto > 35 %, Fig. 6b). Diatoms have
been suggested to contribute to higher phytoplankton-derived POC in
Arctic/subarctic waters (Crawford et al., 2015). The
Labrador Shelf region, where blooms are generally dominated by large
Arctic/ice-related diatoms
(Fragoso et al.,
2016), had relatively high contributions of POCphyto (> 50 %) to the total POC, even though smaller phytoplankton forms, such as
chlorophytes, were also abundant. Low POC : PON ratios, as well as low
POCphyto concentrations, were associated with Atlantic waters, which had
higher contributions of flagellates (particularly before bloom initiation).
Similar findings were reported by Crawford et al. (2015),
where low POCphyto was associated with larger contributions of
flagellates (< 8 µm) in oligotrophic Arctic waters, such as the
Beaufort Sea and Canada Basin. Crawford et al. (2015)
also considered that POC : PON ratios might have been reduced by the presence
of heterotrophic microbes (bacteria, flagellates and ciliates) since these
microorganisms have POC : PON ratios much lower than the canonical Redfield
ratio of 6.6
(Lee and
Fuhrman, 1987; Vrede et al., 2002). Bacteria and other heterotrophic
organisms were not quantified in our study, although
Li and Harrison (2001) have shown that bacterial
biomass from surface waters was 62 % greater (average from 1989 to 1998 of 13.8 mg C m-3) in the central region than in shelf areas of the
Labrador Sea.
Changes in POC : PON may also be related to the physiological status of
phytoplankton and/or community structure. In the North Water Polynya (Baffin
Bay), POC : PON ratios during phytoplankton blooms increased between spring
(5.8) and summer (8.9) as phytoplankton responded to nitrate starvation by
producing N-poor photoprotective pigments (Mei et
al., 2005). Daly et al. (1999) also found high POC : PON
ratios (∼ 8.9) in Arctic surface waters dominated by diatoms
on the northeastern Greenland Shelf, which were attributed to nutrient
limitation. Atlantic waters appear to have an excess of nitrate compared
with Arctic waters
(Harrison et al.,
2013), which could explain why phytoplankton from Atlantic Waters had lower
POC : PON ratios in our study (Fig. 6c). Conversely, Arctic-influenced
waters on or near the shelves had higher Si(OH)4 : NO3- and
lower NO3- : PO43- than those in the central basin in this
study (Fig. 2k and l), which could also have contributed to the observed
high POC : PON ratios.
A few stations in shelf waters of the Labrador Sea also had remarkably high
POC : PON ratios (> 10), and low POCphyto contributions,
suggesting high contributions of detritus. These waters probably receive
higher inputs of Arctic and glacial ice melt, which could introduce POC from
external sources. Hood et al. (2015) showed that POC export
from glaciers is large, particularly from the Greenland Ice Sheet and it
occurs in suspended sediments derived from glacier meltwater. High POC : PON
ratios (> 10), particularly in waters where Phaeocystis were abundant, may
also be linked to the mucilaginous matrix of the Phaeocystis colonies
(Palmisano et al., 1986). The mucopolysaccharide appears to
contain excess carbon, particularly when nutrients start to become depleted
and colonies become senescent (Alderkamp et
al., 2007; Wassmann et al., 1990).
Physiological parameters of distinct phytoplankton communities
Accessory pigments (AP) are assumed to have a ubiquitous, global, log–log
linear relationship with chlorophyll a in aquatic environments
(Trees et al., 2000). This linear relationship is often used as
an index of quality control in pigment analysis, which is required due to
uncertainties of the quantitative comparability of data among different
surveys, and may be related to differences in analytical procedures and
sample storage methods used in different laboratories. In the current study,
the slope of AP to total chlorophyll a (TChl a) on a logarithm scale (Fig. 7)
passed the quality control criteria of slopes ranging from 0.7 to 1.4 and
r2> 0.90 as applied in previous studies
(e.g.
Aiken et al., 2009; Peloquin et al., 2013; Thompson et al., 2011) and were
within the range observed throughout worldwide aquatic systems (slope from
0.8 to 1.3 compared to 0.86 to 1.03 observed in our study)
(Trees et al., 2000). An interesting trend was also found where
phytoplankton pigment ratios varied clearly within distinct communities in
the Labrador Sea. According to our data, phytoplankton communities found in
colder waters (of Arctic origin) had lower accessory pigments ratios to
total chlorophyll a ratio (logAP : logTChl a) (slope, 0.86) when compared to
communities from warmer waters (Irminger Current from Atlantic origin)
(slope, 1.03). Changes in the ratios of logAP : logTChl a as a function of
phytoplankton community composition has also been previously observed by
Stramska et al. (2006). These authors showed a higher slope
of logAP : logTChl a when dinoflagellates were dominant during summer in
northern polar Atlantic waters as opposed to lower ratios associated with
flagellates in spring. Trees et al. (2000) and
Aiken et al. (2009) also reported lower
logAP : logTChl a (slope < 1.00) in oligotrophic waters dominated by
picoplankton as opposed to higher ratios in upwelling waters where
microplankton, particularly diatoms, were dominant.
Environmental parameters, such as nutrients and light conditions, have also
been suggested to influence logAP : logTChl a, regardless of community
composition (Trees et al., 2000). However, in our study, these
parameters, analysed as nitrate and silicate concentrations and
stratification index, did not vary with logAP : logTChl a (data not shown) as
opposed to temperature. Phytoplankton community distributions varied clearly
according to temperature with Phaeocystis occurring in colder Arctic waters and
dinoflagellates in warmer Atlantic waters. Although both communities were
co-dominated by diatoms (relative abundance > 70 % of total
chlorophyll a), the ratio logAP : logTChl a varied considerably, suggesting that
diatom species from both Arctic and Atlantic waters varied intrinsically in
pigment composition, as observed by the distinct Fuco to TChl a ratios of
shelf (Arctic) versus central (Atlantic) waters (Table S1, Supplement). Fragoso et al. (2016) have previously observed that the diatom
species from Arctic and Atlantic waters of the Labrador Sea during spring
varied in terms of species composition. According to the study by Fragoso et
al. (2016), the diatoms Ephemera planamembranacea and Fragilariopsis atlantica were typically found in Atlantic waters,
whereas polar diatoms, including Thalassiosira species (T. hyalina, T. nordenskioeldii, for example), in addition to
Bacterosira bathyomphala, Fossula arctica, Nitzschia frigida, and Fragilariopsis cylindrus, were all found in Arctic-influenced waters. It is possible that
the distinct composition of diatoms from these biogeographical regions might
have influenced the pigment composition in these waters. Likewise, it is
possible that temperature had a strong physiological effect on the
logAP : logTChl a ratio. Many environmental factors, such as turbulence and
coloured dissolved organic matter (CDOM) concentrations, could have
contributed to the variance of chlorophylls (light–shade adaptation) and
AP (chromatic adaptation) observed due to changes they cause
in spectral light absorption by phytoplankton. Turbulence and CDOM, however,
were not measured in this study and a direct physiological
temperature-induced effect or taxonomic effect on logAP : logTChl a is currently
unknown.
The variation in photosynthetic parameters in the distinct phytoplankton
biogeographical provinces demonstrated how each phytoplankton community
responds to environmental conditions. Harrison and Platt (1986)
found that the photophysiology of phytoplankton from the Labrador Sea is
influenced by temperature and irradiance. Nonetheless, phytoplankton
composition may also influence the values of the photosynthetic parameters.
Light-saturated photosynthetic rates and saturation irradiances, for
instance, were higher at stations where diatoms were dominant (> 70 %), as opposed to stations where flagellates were more abundant (from
40 up to 70 %). Similar findings were reported by
Huot et al. (2013), who observed that
light-saturated photosynthetic rates in the Beaufort Sea (Arctic Ocean) were
higher for communities composed of large cells, presumably diatoms, compared
to smaller flagellates.
Polar phytoplankton communities from shelf waters (east versus west) observed in
this study had distinctive photophysiological characteristics. Comparing
these blooms, diatom/chlorophyte communities (west) had higher
photosynthetic efficiency (αB= 9.2 × 10-2 mg C [mg Chl af] h-1 [W m-2]-1), lower onset light-saturation
irradiance (Ek= 29 W m-2), and higher photoinhibition
(β= 16 × 10-4 mg C [mg Chl af] h-1 [W m-2]-1) than communities from the east. This suggests that the
community located in the Labrador Shelf waters (west) was more
light-stressed compared to the community observed in the east
(diatom/Phaeocystis). Haline-stratification due to the influence of Arctic waters
occurs in both regions during spring, contributing to the shallow mixed-layer depths (< 33 m) observed (Table 5). However, waters from the
Labrador Shelf (west, Cluster C3a) were more stratified than the Greenland
Shelf (cluster B, see SI values, Table 5) because of
the local sea ice melt observed in this area, which contributes to increased
stratification in this region. The diatom species observed on the Labrador
Shelf were mostly sea ice related (Fragilariopsis cylindrus, Fossula arctica, Nitzschia frigida) compared to pelagic species
observed in the Greenland Shelf waters (Thalassiosira gravida, for example) (Fragoso et al.,
2016). Sensitivity of sea-ice-related diatoms to irradiances > 15 µmol photons m-2 s-1 has been reported
(Alou-Font et al., 2016), which could help explain why
phytoplankton communities from the west were photoinhibited.
The communities dominated by Phaeocystis/diatoms located near Greenland (east) had the
inverse pattern: low photosynthetic efficiency (average αB= 6.8 × 10-2 mg C [mg Chl af] h-1 [W m-2]-1)
and high onset light-saturation irradiances (Ek= 60 W m-2). This pattern in diatom/Phaeocystis-dominated communities mean that
photosynthetic rates were relatively low at high light intensities, although
photoinhibition was low (β= 4 × 10-4 mg C [mg Chl af] h-1 [W m-2]-1). Phaeocystis antarctica, widespread in Antarctic waters, relies
heavily on photodamage recovery, such as D1 protein repair (Kropuenske et
al., 2009), which could explain how these communities overcome
photoinhibition. Stuart et al. (2000), however, found a high photosynthetic
efficiency (αB) for a population dominated by Phaeocystis near Greenland
and attributed this to the small cell size of Phaeocystis. In addition to the exposure
of ice-related diatoms to high light levels due to increased stratification,
the high concentration of chlorophytes and prasinophytes, which are also
small in cell size, might also explain the higher αB observed
in the Labrador Shelf waters (west, cluster C3a), when compared to values
from a community dominated by diatom/Phaeocystis blooms (east, cluster B).
Phytoplankton communities from Atlantic waters (co-dominated by diatoms and
dinoflagellates) were highly susceptible to photoinhibition (β= 29 × 10-4 mg C [mg Chl af] h-1 [W m-2]-1)
compared with the other communities in the Labrador Sea. Days are longer and
solar incidence is higher in June compared to May at these latitudes
(Harrison et al., 2013). Dinoflagellates were found to bloom in the central
Labrador Sea in June as a consequence of increased thermal stratification.
To cope with high light levels and potential photodamage, this
phytoplankton community appeared to increase the levels of photoprotective
pigments, such as those used in the xanthophyll cycle (diadinoxanthin (DD)
+ diatoxanthin (DT)). These communities also had high diatoxanthin levels
compared with the other phytoplankton communities in this study, suggesting
that the community was experiencing higher light intensities
(Moisan et
al., 1998). Increases in photoprotective pigments, including
(DD+DT)/TChl a, have also been reported to occur in Arctic phytoplankton
communities from spring to summer, presumably as a response to higher
irradiance (Alou-Font et al., 2016). Thus, photoprotective
capacity can be a key determinant for phytoplankton survival and may also be
related to the taxonomic segregation observed in Arctic and Atlantic
phytoplankton communities.
Phytoplankton communities assessed by HPLC and CHEMTAX methods
A number of studies have used CHEMTAX methods to determine phytoplankton
community structure in Arctic/subarctic waters
(e.g.
Coupel et al., 2012, 2015; Lovejoy et al., 2007; Piquet et al., 2014;
Vidussi et al., 2004; Zhang et al., 2015). Spring phytoplankton communities
from the Labrador Sea have already been investigated in detail
(Fragoso et al.,
2016), although the analysis did not include nano- and pico-flagellates
(except cryptophytes and Phaeocystis pouchetii) and was done over only 4 years (2011–2014) at
selected stations along the L3 (= AR7W) transect. Here, we have combined
phytoplankton information from
Fragoso et
al. (2016) with additional pigment analyses. Although cross-comparison between
these two techniques (carbon biomass estimated from microscopic counts
versus algal group chlorophyll a estimated from CHEMTAX) should not be expected to
give exactly equivalent results, given that most flagellates observed in the
pigment analysis were not counted under the microscope, some comparability
should be possible, at least for the larger cells (e.g. diatoms).
Phaeocystis (r2=0.79) and diatom (r2=0.74) biomasses were well
correlated when carbon biomasses estimated from microscopic counts when
compared with CHEMTAX-derived algal chlorophyll a biomass (data not shown).
Diatoms are the group that usually show the best agreement between the two
methods of biomass estimations (Vidussi et al., 2004; Coupel et al., 2015;
Mendes et al., 2012). For Phaeocystis, a positive relationship between the two methods of
biomass estimation (CHEMTAX and microscopy) confirms that using chlorophyll
c3 was appropriate for detecting and quantifying Phaeocystis biomass in the Labrador Sea.
Similar associations have been observed for Phaeocystis from boreal waters (e.g. P. pouchetii and
P. globosa), which lacks or has low 19-hexanoyloxyfucoxanthin
(Antajan
et al., 2004; Muylaert et al., 2006; Stuart et al., 2000; Wassmann et al.,
1990). Conversely, 19-hexanoyloxyfucoxanthin is a characteristic pigment
marker of Phaeocystis from austral waters (P. antarctica)
(Arrigo
et al., 2010, 2014; Fragoso, 2009; Fragoso and Smith, 2012). Dinoflagellates
gave a poor correlation between biomass estimates made using the two methods
(r2=0.12, data not shown). A lack of or weak relationship between
both biomass estimations for dinoflagellates has been previously reported in
Arctic waters (Vidussi et al., 2004; Coupel et al., 2005). The argument for this
inconsistency is that some heterotrophic dinoflagellates, which usually lack
photosynthetic pigments, unless they ingest a prey that contains them, might
have been included in the microscopic counts, and it is possible that the
same occurred in
Fragoso et
al. (2016). Cryptophyte biomass estimates from both methods were not related
(data not shown), likely as the biomass of this group was underestimated in
microscopic counts. Inconsistences between CHEMTAX and microscopy methods of
estimating biomasses have also been observed in nanoflagellates and this is
assumed to be because of the low accuracy of visual microscopic counts
(Coupel et al., 2015; Gieskes
and Kraay, 1983).