Introduction
Oceanic concentration of CO2 has increased by ∼ 42 %
over pre-industrial levels, with a continuing annual increase of
∼ 0.4 %. The current CO2 level has reached ∼ 400 µatm
and has been predicted to rise to > 700 µatm by the end of this century (IPCC, 2013), with
estimates exceeding 1000 µatm
(Matear
and Lenton, 2018; Raupach et al., 2007; Raven et al., 2005). With increasing
atmospheric CO2, the oceans continue to absorb CO2 from the
atmosphere, which results in a shift in oceanic carbonate chemistry
resulting in a decrease in seawater pH, or ocean acidification (OA). The
projected increase in atmospheric CO2 and corresponding increase in
ocean uptake is predicted to result in a decrease in global mean surface
seawater pH of 0.3 units below the present value of 8.1 to 7.8
(Wolf-Gladrow et al., 1999). Under this scenario, the shift in
dissolved inorganic carbon (DIC) equilibria has wide ranging implications
for phytoplankton photosynthetic carbon fixation rates and growth
(Riebesell, 2004).
Concurrent with OA, elevated atmospheric CO2 and other climate active
gases have warmed the planet by ∼ 0.6 ∘C over the
past 100 years (IPCC, 2014). Atmospheric temperature has been predicted to
rise by a further 1.8 to 4 ∘C by the end of this century
(Alley et al.,
2007). Phytoplankton metabolic activity may be accelerated by increased
temperature (Eppley, 1972), which can vary depending on the
phytoplankton species and their physiological
requirements (Beardall
et al., 2009; Boyd et al., 2013). Long-term data sets already suggest that
ongoing changes in coastal phytoplankton communities are likely due to
climate shifts and other anthropogenic influences
(Edwards et al.,
2006; Smetacek and Cloern, 2008; Widdicombe et al., 2010). The response to
OA and temperature can potentially alter the community composition,
community biomass and photophysiology. Understanding how these two factors
may interact, synergistically or antagonistically, is critical to our
understanding and for predicting future primary productivity
(Boyd and Doney, 2002; Dunne, 2014).
Laboratory studies of phytoplankton species in culture and studies on
natural populations in the field have shown that most species exhibit
sensitivity, in terms of growth and photosynthetic rates, to elevated
pCO2 and temperature individually. To date, only a few studies have
investigated the interactive effects of these two parameters on natural
populations
(e.g. Coello-Camba et al., 2014; Feng et al., 2009; Gao et al., 2017; Hare et al.,
2007). Most laboratory studies demonstrate variable results with
species-specific responses. In the diatom Thalassiosira weissflogii, for example, pCO2 elevated to
1000 µatm and +5 ∘C temperature synergistically
enhanced growth, while the same conditions resulted in a reduction in growth
for the diatom Dactyliosolen fragilissimus (Taucher et
al., 2015). Although there have been fewer studies on dinoflagellates,
variable responses have also been reported
(Errera
et al., 2014; Fu et al., 2008). In natural populations, elevated pCO2
has stimulated the growth of pico- and nanophytoplankton
(Boras
et al., 2016; Engel et al., 2008) while increased temperature has reduced
their biomass
(Moustaka-Gouni
et al., 2016; Peter and Sommer, 2012). In a recent field study on natural
phytoplankton communities, elevated temperature (+3 ∘C above
ambient) enhanced community biomass but the combined influence of elevated
temperature and pCO2 reduced the biomass
(Gao et al., 2017).
Phytoplankton species composition, abundance and biomass has been measured
since 1992 at the time series station L4 in the western English Channel
(WEC) to evaluate how global changes could drive future shifts in
phytoplankton community structure and carbon biogeochemistry. At this
station, sea surface temperature and pCO2 reach maximum values during
late summer and start to decline in autumn. During October, mean seawater
temperatures at 10 m decrease from 15.39 ∘C (±0.49 SD) to
14.37 ∘C (± 0.62 SD). Following a period of CO2
oversaturation in late summer, pCO2 returns to near-equilibrium at
station L4 in October when mean pCO2 values decrease from 455.32 µatm (± 63.92 SD) to 404.06 µatm (± 38.55 SD) (Kitidis et
al., 2012).
From a biological perspective, the autumn period at station L4 is
characterised by the decline of the late summer diatom and dinoflagellate
blooms (Widdicombe et al., 2010) when their biomass approaches values close
to the time series minima (diatom biomass range: 6.01 (± 6.88 SD)–2.85 (± 3.28 SD) mgCm-3;
dinoflagellate biomass range: 1.75
(± 3.28 SD)–0.66 (± 1.08 SD) mgCm-3). Typically, over
this period nanophytoplankton becomes numerically dominant and biomass
ranges from 20.94 (± 33.25 SD) to 9.38 (± 3.31 SD) mgCm-3,
though there is considerable variability in this biomass.
Based on the existing literature, the working hypotheses of this study are
that (1) community biomass will increase differentially under individual
treatments of elevated temperature and pCO2, (2) elevated pCO2
will lead to taxonomic shifts due to differences in species-specific
CO2 concentrating mechanisms and/or RuBisCO specificity, (3) photosynthetic
carbon fixation rates will increase differentially under
individual treatments of elevated temperature and pCO2, (4) elevated
temperature will lead to taxonomic shifts due to species-specific thermal
optima, and (5) temperature and pCO2 elevated simultaneously will have
synergistic effects.
The objective of the study was therefore to investigate the combined effects
of elevated pCO2 and temperature on phytoplankton community structure,
biomass and photosynthetic carbon fixation rates during the autumn
transition from diatoms and dinoflagellates to nanophytoplankton at station L4 in the WEC.
Materials and methods
Perturbation experiment, sampling and experimental setup
Experimental seawater containing a natural phytoplankton community was
sampled at station L4 (50∘15′ N, 4∘13′ W) on
7 October 2015 from 10 m depth (40 L). The experimental seawater was
gently pre-filtered through a 200 µm Nitex mesh to remove
mesozooplankton grazers, into two 20 L acid-cleaned carboys. While grazers
play an important role in regulating phytoplankton community structure
(e.g. Strom, 2002), our experimental goals considered only the
effects of elevated temperature and pCO2, though the mesh size used
does not remove microzooplankton. In addition, 320 L of seawater was
collected into sixteen 20 L acid-cleaned carboys from the same depth for use
as experimental media. Immediately upon return to the laboratory the media
seawater was filtered through an in-line 0.2 and 0.1 µm filter
(Acropak™, Pall Life Sciences) and then stored in the dark at 14 ∘C
until use. The experimental seawater was gently and thoroughly
mixed and transferred in equal parts from each carboy (to ensure
homogeneity) to sixteen 2.5 L borosilicate incubation bottles (four sets of
four replicates). The remaining experimental seawater was sampled for initial
(T0) concentrations of nutrients, Chl a, total alkalinity, dissolved
inorganic carbon, particulate organic carbon (POC) and nitrogen (PON) and
was also used to characterise the starting experimental phytoplankton
community. The incubation bottles were placed in an outdoor simulated
in situ incubation culture system and each set of replicates was linked to
one of four 22 L reservoirs filled with the filtered seawater media. Neutral
density spectrally corrected blue filters (Lee Filter no. 061) were placed
between polycarbonate sheets and mounted to the top, sides and ends of the
incubation system to provide ∼ 50 % irradiance,
approximating PAR measured at 10 m depth at station L4 on the day of
sampling prior to starting experimental incubations (see Fig. S1 in the
Supplement for time course of PAR levels during the experiment).
The media was aerated with CO2 free air and 5 % CO2 in air
precisely mixed using a mass flow controller (Bronkhorst UK Limited) and
used for the microcosm dilutions as per the following experimental design:
(1) control (390 µatm pCO2, 14.5 ∘C matching station
L4 in situ values), (2) high temperature (390 µatm pCO2, 18.5 ∘C),
(3) high pCO2 (800 µatm pCO2, 14.5 ∘C) and
(4) combination (800 µatm pCO2, 18.5 ∘C).
Initial nutrient concentrations (0.24 µM nitrate + nitrite, 0.086 µM
phosphate and 2.14 µM silicate on 7 October 2015)
were amended to 8 µM nitrate + nitrite and 0.5 µM phosphate.
Pulses of nutrient inputs frequently occur at station L4 from August to
December following heavy rainfall events and subsequent riverine inputs to
the system (e.g. Barnes et al., 2015). Our
nutrient amendments simulated these in situ conditions and were held
constant to maintain phytoplankton growth. Previous pilot studies
highlighted that, if these concentrations are not maintained, the
phytoplankton population crashes (Keys, 2017). As the
phytoplankton community was sampled over the transitional phase from diatoms
and dinoflagellates to nanophytoplankton, the in situ silicate concentration
was maintained to reproduce the silicate concentrations typical of this time
of year (Smyth et al., 2010). Nutrient
concentrations were measured at time point T0 only.
Media transfer and sample acquisition was driven by peristaltic pumps.
Following 48 h acclimation in batch culture, semi-continuous daily
dilution rates were maintained at between 10 and 13 % of the incubation
bottle volume throughout the experiment. CO2-enriched seawater was
added to the high-CO2 treatment replicates every 24 h, acclimating
the natural phytoplankton population to increments of elevated pCO2
from ambient to ∼ 800 µatm over 8 days followed by
maintenance at ∼ 800 µatm as per the method described
by Schulz et al. (2009). Adding CO2-enriched seawater is the
preferred protocol, since some phytoplankton species are inhibited by the
mechanical effects of direct bubbling
(Riebesell et al., 2010; Shi et al.,
2009), which causes a reduction in growth rates and the formation of
aggregates (Love et al., 2016). pH was monitored
daily to adjust the pCO2 of the experimental media (+/-) prior to
dilutions to maintain target pCO2 levels in the incubation bottles. The
seasonality in pH and total alkalinity (TA) are fairly stable at station L4
with high pH and low dissolved inorganic carbon during early summer
and low pH, high DIC throughout autumn and winter (Kitidis et
al., 2012). By maintaining the carbonate chemistry over the duration of the
experiment, we aimed to simulate natural events at the study site.
To provide sufficient time for changes in the phytoplankton community to
occur and to achieve an ecologically relevant data set, the incubation
period was extended well beyond short-term acclimation. Previous pilot
studies using the same experimental protocols highlighted that, after
∼ 20 days of incubation, significant changes in community
structure and biomass were observed (Keys, 2017). These results
were used to inform a more relevant incubation period of 30+ days.
Analytical methods
Chlorophyll a
Chl a was measured in each incubation bottle.
Triplicate samples of 100 mL from
each replicate were filtered onto 25 mm GF/F filters (nominal pore size 0.7 µm),
extracted in 90 % acetone overnight at -20 ∘C and
Chl a concentration was measured on a Turner Trilogy™
fluorometer using the non-acidified method of Welschmeyer
(1994). The fluorometer was calibrated against a stock Chl a standard
(Anacystis nidulans, Sigma Aldrich, UK), the concentration of which was determined with a
PerkinElmer™ spectrophotometer at wavelengths 663.89 and
750.11 nm. Samples for Chl a analysis were taken every 2–3 days.
Carbonate system
Samples of 70 mL for total alkalinity (TA) and dissolved inorganic carbon
analysis were collected from each experimental replicate, stored in amber
borosilicate bottles with no head space and fixed with 40 µL of
supersaturated Hg2Cl2 solution for later determination (Apollo
SciTech™ Alkalinity Titrator AS-ALK2; Apollo
SciTech™ AS-C3 DIC analyser, with analytical precision of 3 µmolkg-1).
Duplicate measurements were made for TA and
triplicate measurements for DIC. Carbonate system parameter values for media
and treatment samples were calculated from TA and DIC measurements using the
programme CO2SYS (Pierrot et al., 2006) with dissociation
constants of carbonic acid of Mehrbach et al. (1973)
refitted by Dickson and Millero (1987).
Samples for TA and DIC were taken for analysis every 2–3 days throughout the
experiment.
Phytoplankton community analysis
Phytoplankton community analysis was performed by flow cytometry (Becton
Dickinson Accuri™ C6) for the 0.2 to 18 µm size
fraction following Tarran et al. (2006) and inverted
light microscopy was used to enumerate cells > 18 µm (BS
EN 15204, 2006). For flow cytometry, 2 mL samples fixed with glutaraldehyde
to a final concentration of 2 % were flash frozen in liquid nitrogen and
stored at -80 ∘C for subsequent analysis. Phytoplankton data
acquisition was triggered on both chlorophyll fluorescence and forward light
scatter (FSC) using prior knowledge of the position of Synechococcus sp. to set the lower
limit of analysis. Density plots of FSC vs. Chl fluorescence, phycoerythrin
fluorescence vs. Chl fluorescence and side scatter vs. Chl
fluorescence were used to discriminate between Synechococcus sp., picoeukaryote phytoplankton
(approx. 0.5–3 µm), coccolithophores, cryptophytes, Phaeocystis sp. single
cells and nanophytoplankton (eukaryotes > 3 µm, excluding
the coccolithophores, cryptophytes and Phaeocystis sp. single cells; for further
information on flow cytometer calibration for phytoplankton size
measurements, see the Supplement). For inverted light microscopy,
140 mL samples were fixed with 2 % (final concentration) acid Lugol's
iodine solution and analysed by inverted light microscopy
(Olympus™ IMT-2) using the Utermöhl counting technique
(Utermöhl, 1958; Widdicombe et al.,
2010). Phytoplankton community samples were taken at T0, T10, T17, T24 and
T36.
Phytoplankton community biomass
The smaller size fraction identified and enumerated through flow cytometry;
picophytoplankton, nanophytoplankton, Synechococcus, coccolithophores and cryptophytes
were converted to carbon biomass (mgCm-3) using a spherical model to
calculate mean cell volume,
43⋅π⋅r3,
and a carbon conversion factor of 0.22 pgCµm-3 (Booth,
1988). A conversion factor of 0.285 pgCµm-3 was used for
coccolithophores (Tarran et al., 2006) and cell a
volume of 113 µm3 and carbon cell-1 value of 18 pg applied
for Phaeocystis spp. (Widdicombe et al., 2010). Phaeocystis spp. were identified
and enumerated by flow cytometry separately to the nanophytoplankton class
due to high observed abundance in the high-pCO2 treatment. Mean cell
measurements of individual species/taxa were used to calculate cell
bio-volume for the 18 µm + size fraction according to Kovala and
Larrance (1966) and converted to biomass according to the equations of
Menden-Deuer and Lessard (2000).
POC and PON
Samples for particulate organic carbon (POC) and particulate organic
nitrogen (PON) were taken at T0, T15 and T36. Samples of 150 mL were taken from
each replicate and filtered under gentle vacuum pressure onto pre-ashed
25 mm
glass fibre filters (GF/F, nominal pore size 0.7 µm). Filters were
stored in acid-washed petri-slides at -20 ∘C until further
processing. Sample analysis was conducted using a Thermoquest elemental
analyser (Flash 1112). Acetanilide standards (Sigma Aldrich, UK) were used
to calibrate measurements of carbon and nitrogen and also used during the
analysis to account for possible drift in measured concentrations.
Chl fluorescence-based photophysiology
Photosystem II (PSII) variable chlorophyll fluorescence parameters were
measured using a fast repetition rate fluorometer (FRRf) (FastOcean sensor
in combination with an Act2Run laboratory system, Chelsea Technologies, West
Molesey, UK). The excitation wavelengths of the FRRf's light-emitting diodes
(LEDs) were 450, 530 and 624 nm. The instrument was used in single turnover
mode with a saturation phase comprising 100 flashlets on a 2 µs pitch
and a relaxation phase comprising 40 flashlets on a 50 µs pitch.
Measurements were conducted in a temperature-controlled chamber at 15 ∘C.
The minimum (Fo) and maximum (Fm) Chl fluorescence
were estimated according to
Kolber et al. (1998).
Maximum quantum yields of PSII were calculated as
Fv/Fm=(Fm-Fo)/Fm.
PSII electron flux was calculated on a volume basis (JVPSII; mole-m-3d-1)
using the absorption algorithm
(Oxborough et al., 2012) following spectral correction
by normalising the FRRf LED emission to the white spectra using FastPRO 8
software. This step required inputting the experimental phytoplankton
community fluorescence excitation spectra values (FES). Since we did not
measure the FES of our experimental samples, we used mean literature values
for each phytoplankton group calculated proportionally (based on percentage
contribution to total estimated biomass per phytoplankton group) as
representative values for our experimental samples. The JVPSII rates
were converted to chlorophyll-specific carbon fixation rates (mgC (mg Chl a)-1 m-3h-1),
calculated as follows:
JVPSII×φE:C×MWC/Chla,
where φE:C is the electron requirement for carbon uptake
(molecule CO2 (molelectrons)-1), MWC is the molecular weight
of carbon and Chl a is the Chl a measurement specific to each sample.
Chl a-specific JVPSII-based photosynthesis–irradiance (PE) curves were conducted
in replicate batches between 10:00 and 16:00 to account for variability over
the photoperiod at between 8 and 14 irradiance intensities. The maximum
intensity applied was adjusted according to ambient natural irradiance on
the day of sampling. Maximum photosynthetic rates of carbon fixation
(PmB), the light-limited slope (αB) and the light
saturation point of photosynthesis (Ek) were estimated by fitting the
data to the model of Webb et al. (1974):
PB=(1-e×(-α×I/PmB)).
Due to instrument failure during the experiment, samples for FRRf
fluorescence-based light curves were taken at T36 only.
(a) MODIS weekly composite Chl a image of the western English Channel
covering the period 30 September–6 October 2015 (coincident
with the week of phytoplankton community sampling for the present study);
processing courtesy of NEODAAS. The position of coastal station L4 is marked
with a white diamond. (b) Profiles of weekly nutrient and Chl a concentrations
from station L4 at a depth of 10 m over the second half of 2015 in the
months prior to phytoplankton community sampling (indicated by the black arrow
and text).
Statistical analysis
To test for effects of temperature, pCO2 and possible time dependence
of the measured response variables (Chl a, total biomass, POC, PON,
photosynthetic parameters and biomass of individual species), generalised
linear mixed models with the factors pCO2, temperature and time (and
all interactions) were applied to the data between T0 and T36. Analyses were
conducted using the lme4 package in R (R Core Team, 2016).
Results
Chl a concentration in the WEC at station L4 from 30 September to 6 October 2015
(when sea water was collected for the experiment) varied
between 0.02 and 5 mgm-3, with a mean concentration of ∼ 1.6 mgm-3 (Fig. 1a).
Over the period leading up to phytoplankton
community sampling, increasing nitrate and silicate concentrations coincided
with a Chl a peak on 23 September (Fig. 1b). Routine net trawl (20 µm)
sample observations indicated a phytoplankton community dominated
by the diatoms Leptocylindrus danicus and L. minimus with a lower presence of the dinoflagellates
Prorocentrum cordatum, Heterocapsa spp. and Oxytoxum gracile.
Following decreasing nitrate concentrations, there was a P. cordatum
bloom on 29 September, during the week before the experiment started
(data not shown).
Calculated values of partial pressure of CO2 in seawater
(pCO2) (a) and pH (b) from direct measurements of total alkalinity and
dissolved inorganic carbon. (For full carbonate system values see Table S1 in
the Supplement).
Experimental carbonate system
Equilibration to the target high pCO2 values (800 µatm) within
the high-pCO2 and combination treatments was achieved at T10 (Fig. 2a, b).
These treatments were slowly acclimated to increasing levels of
pCO2 over 7 days (from the initial dilution at T3), while the control
and high-temperature treatments were acclimated at the same ambient
carbonate system values as those measured at station L4 on the day of
sampling. Following equilibration, the mean pCO2 values within the
control and high-temperature treatments were 394.9 (± 4.3 SD) and
393.2 (± 4.8 SD) µatm respectively, while in the high-pCO2
and combination treatments mean pCO2 values were 822.6 (± 9.4)
and 836.5 (± 15.6 SD) µatm respectively. Carbonate system
values remained stable throughout the experiment (for full carbonate system
measured and calculated parameters, see Table S1 in the Supplement).
Experimental temperature treatments
Mean temperatures in the control and high-pCO2 treatments were 14.1
(± 0.35 SD) ∘C, and in the high-temperature and combination
treatments the mean temperatures were 18.6 (± 0.42 SD) ∘C,
with a mean temperature difference between the ambient and high-temperature
treatments of 4.46 (± 0.42 SD) ∘C (Supplement,
Fig. S2a, b).
Results of generalised linear mixed model testing for effects of
time, temperature, pCO2 and all interactions on Chl a, phytoplankton
biomass, and particulate organic carbon and nitrogen. Significant results are
in bold; * p<0.05, ** p<0.01, *** p<0.001.
Response variable
n
df
z value
p
Significance
Chl a (mgm-3)
High temp
516
507
0.412
0.680
High pCO2
516
507
0.664
0.507
Time
516
507
3.815
< 0.001
***
High temp × high pCO2
516
507
1.100
0.271
Time × high temp
516
507
-0.213
0.831
Time × high CO2
516
507
-0.011
0.991
Time × high temp × high CO2
516
507
0.340
0.734
Estimated biomass (mgCm-3)
High temp
80
71
0.092
0.927
High pCO2
80
71
2.102
0.036
*
Time
80
71
2.524
0.012
*
High temp × high pCO2
80
71
1.253
0.210
Time × high temp
80
71
1.866
0.062
Time × high CO2
80
71
4.414
< 0.001
***
Time × high temp × high CO2
80
71
-1.050
0.294
POC (mgm-3)
High temp
48
38
-0.977
0.328
High pCO2
48
38
-0.866
0.386
Time
48
38
-0.203
0.839
High temp × high pCO2
48
38
-0.29
0.772
Time × high temp
48
38
3.648
< 0.001
***
Time × high CO2
48
38
4.333
< 0.001
***
Time × high temp × high CO2
48
38
0.913
0.361
PON (mgm-3)
High temp
48
38
-0.640
0.522
High pCO2
48
38
-0.479
0.632
Time
48
38
0.202
0.84
High temp × high pCO2
48
38
0.667
0.505
Time × high temp
48
38
1.674
0.094
Time × high CO2
48
38
2.037
< 0.05
*
Time × high temp × high CO2
48
38
-0.141
0.730
POC : PON mol C : mol N
High temp
48
38
0.222
0.824
High pCO2
48
38
0.029
0.977
Time
48
38
0.184
0.854
High temp × high pCO2
48
38
0.990
0.322
Time × high temp
48
38
2.377
0.017
*
Time × high CO2
48
38
2.748
0.005
**
Time × high temp × high CO2
48
38
-0.215
0.829
Chlorophyll a
Mean Chl a in the experimental seawater at T0 was 1.64 (± 0.02 SD) mgm-3
(Fig. 3a). This decreased in all treatments between T0 and T7 to
∼ 0.1 (± 0.09, 0.035 and 0.035 SD) mgm-3 in the
control, high-pCO2 and combination treatments, while in the
high-temperature treatment at T7 Chl a was 0.46 mgm-3 (± 0.29 SD) (z = 2.176,
p<0.05). From T7 to T12 Chl a increased in all treatments,
which was highest in the combination (4.99 mgm-3 ± 0.69 SD) and
high-pCO2 treatments (3.83 mgm-3 ± 0.43 SD).
Overall, Chl a was significantly influenced by experimental time, independent of
experimental treatments (Table 1). At T36 Chl a concentration in the
combination treatment was higher (6.87 (± 0.58 SD) mgm-3) than
all other treatments while the high-temperature treatment concentration was
higher (4.77 (± 0.44 SD) mgm-3) than the control and high-pCO2 treatment. Mean concentrations for the
control and high-pCO2
treatment at T36 were not significantly different at 3.30 (± 0.22 SD)
and 3.46 (± 0.35 SD) mgm-3 respectively (pairwise comparison t = 0.78, p = 0.858).
Phytoplankton biomass
The starting biomass in all treatments was 110.2 (± 5.7 SD) mgCm-3 (Fig. 3b).
The biomass was dominated by dinoflagellates
(∼ 50 %) with smaller contributions from nanophytoplankton
(∼ 13 %), cryptophytes (∼ 11 %), diatoms
(∼ 9 %), coccolithophores (∼ 8 %),
Synechococcus (∼ 6 %) and picophytoplankton (∼ 3 %).
Total biomass was significantly influenced in all treatments over time
(Table 1), and at T10 it was significantly higher in the high-temperature
treatment when biomass reached 752 (± 106 SD) mgCm-3 (z = 2.769,
p<0.01). Biomass was significantly higher in the elevated
pCO2 treatment (interaction of time × high pCO2) (Table 1),
reaching 2481 (± 182.68 SD) mgCm-3 at T36, ∼ 6.5-fold
higher than the control (z = 3.657, p<0.001). Total
biomass in the high-temperature treatment at T36 was significantly higher
than the combination treatment and ambient control (z=2.744, p<0.001), which were 525 (± 28.02 SD) mgCm-3 and 378 (± 33.95 SD) mgCm-3
respectively. Reaching 1735 (± 169.24 SD) mgCm-3, biomass in the high-temperature treatment was
∼ 4.6-fold higher than the control.
POC followed the same trends in all treatments between T0 and T36 (Fig. 3c)
and was in close range of the estimated biomass (R2=0.914, Fig. 3d).
POC was significantly influenced by the interaction of time × high
pCO2 and time × high temperature (Table 1). At T36 POC was
significantly higher in the high-pCO2 treatment (2086 ± 155.19 SDmgm-3)
followed by the high-temperature treatment
(1594 ± 162.24 SDmgm-3) –
∼ 5.4-fold and 4-fold higher than the
control respectively, whereas a decline in POC was observed in the control
and combination treatment. PON followed the same trend as POC over the
course of the experiment, though it was only significantly influenced by the
interaction between time and high pCO2 (Fig. 3e, Table 1). At T36
concentrations were 147 (± 12.99 SD) and 133 (± 15.59 SD) mgm-3
in the high-pCO2 and high-temperature treatments respectively,
while PON was 57.75 (± 13.07 SD) mgm-3 in the combination
treatment and 47.18 (± 9.32 SD) mgm-3 in the control. POC : PON
ratios were significantly influenced by the interaction of time × high
pCO2 and time × high temperature (Table 1). The largest increase of 33 %,
from 10.72 to 14.26 (± 1.73 SD) molC : molN, was in the
high-pCO2 treatment (73 % higher than the control), followed by an
increase of 32 % to 9.83 (± 1.82 SD) molC : molN in the combination
treatment (19 % higher than the control) and an increase of 17 % to
12.09 (± 2.14 SD) molC : molN in the high-temperature treatment (46 %
higher than the control). In contrast, the POC : PON ratio in the control
declined by 20 % from T0 to T36, from 10.33 to 8.26 (± 0.50 SD) molC : molN (Fig. 3f).
Results of generalised linear mixed model testing for significant
effects of time, temperature, pCO2 and all interactions on
phytoplankton species biomass. Significant results are in bold; * p<0.05, ** p<0.01, *** p<0.001.
Response variable
n
df
z value
p
Significance
Diatoms (mgCm-3)
High temp
80
70
-0.216
0.829
High pCO2
80
70
-0.895
0.371
Time
80
70
2.951
0.003
**
High temp × high pCO2
80
70
1.063
0.288
Time × high temp
80
70
-1.151
0.250
Time × high CO2
80
70
0.560
0.576
Time × high temp × high CO2
80
70
0.368
0.713
Dinoflagellates (mgCm-3)
High temp
80
70
-0.018
0.986
High pCO2
80
70
0.487
0.627
Time
80
70
-2.347
0.019
*
High temp × high pCO2
80
70
-0.166
0.868
Time × high temp
80
70
1.857
0.063
Time × high CO2
80
70
1.009
0.313
Time × high temp × high CO2
80
70
2.207
0.027
*
Nanophytoplankton (mgCm-3)
High temp
80
70
-0.371
0.710
High pCO2
80
70
-2.108
0.035
*
Time
80
70
2.162
0.031
*
High temp × high pCO2
80
70
0.79
0.430
Time × high temp
80
70
1.695
0.090
Time × high CO2
80
70
3.563
< 0.001
***
Time × high temp × high CO2
80
70
-0.806
0.420
Synechococcus (mgCm-3)
High temp
80
70
3.333
< 0.001
***
High pCO2
80
70
2.231
0.026
*
Time
80
70
0.049
0.961
High temp × high pCO2
80
70
2.391
0.017
*
Time × high temp
80
70
4.076
< 0.001
***
Time × high CO2
80
70
-1.553
0.1204
Time × high temp × high CO2
80
70
5.382
< 0.001
***
Picophytoplankton (mgCm-3)
High temp
80
70
0.951
0.342
High pCO2
80
70
-0.472
0.637
Time
80
70
0.897
0.370
High temp × high pCO2
80
70
-1.188
0.235
Time × high temp
80
70
-0.219
0.827
Time × high CO2
80
70
1.411
0.158
Time × high temp × high CO2
80
70
2.792
0.005
**
Coccolithophores (mgCm-3)
High temp
80
70
-0.408
0.683
High pCO2
80
70
-0.308
0.758
Time
80
70
0.211
0.833
High temp × high pCO2
80
70
-0.319
0.750
Time × high temp
80
70
0.269
0.788
Time × high CO2
80
70
0.295
0.768
Time × high temp × high CO2
80
70
0.502
0.615
Continued.
Response variable
n
df
z value
p
Significance
Cryptophytes (mgCm-3)
High temp
80
70
0.207
0.836
High pCO2
80
70
0.256
0.798
Time
80
70
-5.289
< 0.001
***
High temp × high pCO2
80
70
-0.349
0.727
Time × high temp
80
70
1.885
0.059
Time × high CO2
80
70
0.167
0.867
Time × high temp × high CO2
80
70
1.694
0.090
Microzooplankton (mgCm-3)
High temp
80
70
0.138
0.890
High pCO2
80
70
-0.142
0.887
Time
80
70
0.418
0.676
High temp × high pCO2
80
70
0.314
0.753
Time × high temp
80
70
-0.930
0.352
Time × high CO2
80
70
-2.100
0.036
*
Time × high temp × high CO2
80
70
-1.996
0.046
*
Time course of Chl a (a), estimated phytoplankton biomass (b), POC (c),
regression of estimated phytoplankton carbon vs. measured POC (d), PON (e) and POC : PON (f).
Percentage contribution to community biomass by phytoplankton
groups/species throughout the experiment in the control (a), high-temperature (b),
high-CO2 (c) and combination treatments (d).
Response of individual phytoplankton groups to experimental
treatments.
Community composition
From T0 to T24 the community shifted away from dominance of dinoflagellates
in all treatments, followed by further regime shifts between T24 and T36 in
the control and combination treatments. At T36 diatoms dominated the
phytoplankton community biomass in the ambient control (Fig. 4a), while the
high-temperature and high-pCO2 treatments exhibited near-mono-specific
dominance of nanophytoplankton (Fig. 4b, c). The most diverse
community was in the combination treatment where dinoflagellates and
Synechococcus became more prominent (Fig. 4d).
Between T10 and T24 the community shifted to nanophytoplankton in all
experimental treatments. This dominance was maintained to T36 in the
high-temperature and high-pCO2 treatments whereas, in the ambient control and
combination treatment, the community shifted away from nanophytoplankton
(Fig. 5a). Nanophytoplankton biomass was significantly higher in the
high-pCO2 treatment (Table 2) with biomass reaching 2216 (± 189.67 SD) mgCm-3
at T36. This biomass was also high (though not significantly
throughout the experiment until T36) in the high-temperature treatment (T36:
1489 (± 170.32 SD) mgCm-3, z = 1.695, p = 0.09) compared to
the control and combination treatments. In the combination treatment,
nanophytoplankton biomass was 238 (± 14.16 SD) mgCm-3 at
T36, which was higher than the control, though not significantly (162 ± 20.02 SDmgCm-3). In addition to significant differences in
nanophytoplankton biomass amongst the experimental treatments,
treatment-specific differences in cell size were also observed. Larger
nanoflagellates dominated the control (mean cell diameter of 6.34 µm) and smaller nanoflagellates dominated the high-temperature and combination
treatments (mean cell diameters of 3.61 and 4.28 µm),
whereas Phaeocystis spp. dominated the high-pCO2 treatment (mean cell diameter
5.04 µm) and was not observed in any other treatment (Supplement, Fig. S3a–d).
At T0, diatom biomass was low and dominated by Coscinodiscus wailessii (48 %; 4.99 mgCm-3),
Pleurosigma (25 %; 2.56 mgCm-3) and Thalassiosira subtilis (19 %; 1.94 mgCm-3).
Small biomass contributions were made by Navicula distans, undetermined pennate
diatoms and Cylindrotheca closterium. Biomass in the diatom group remained low from T0 to T24 but
increased significantly through time in all treatments (Table 2), with the
highest biomass in the high-pCO2 treatment (235 ± 21.41 SDmgCm-3, Fig. 5b).
The highest diatom contribution to total community
biomass at T36 was in the ambient control (52 % of biomass; 198 ± 17.28 SDmgCm-3). In both the
high-temperature and combination
treatments, diatom biomass was lower at T36 (151 ± 10.94 SD and 124 ± 19.16 SDmgCm-3 respectively). In all treatments, diatom
biomass shifted from the larger C. wailessii to the smaller C. closterium, N. distans,
T. subtilis and Tropidoneis spp., the relative
contributions of which were treatment specific. Overall N. distans dominated diatom
biomass in all treatments at T36 (ambient control: 112 ± 24.86 SDmgCm-3,
56 % of biomass; high temperature: 106 ± 17.75 SDmgCm-3,
70 % of biomass; high pCO2: 152 ± 19.09 SDmgCm-3,
61 % of biomass; and combination: 111 ± 20.97 SDmgCm-3, 89 % of biomass; Supplement, Fig. S4a–d).
The starting dinoflagellate community was dominated by Gyrodinium spirale (91 %; 49 mgCm-3),
with smaller contributions from Katodinium glaucum (5 %; 2.76 mgCm-3),
Prorocentrum cordatum (3 %; 1.78 mgCm-3) and undetermined Gymnodiniales (1 %; 0.49 mgCm-3). At
T36 dinoflagellate biomass was significantly higher in the combination
treatment (90 ± 16.98 SDmgCm-3, Fig. 5c, Table 2) followed by
the high-temperature treatment (57 ± 6.87 SDmgCm-3, Table 2).
There was no significant difference in dinoflagellate biomass between the
high-pCO2 treatment and ambient control at T36 when biomass was low. In
the combination treatment, the dinoflagellate biomass became dominated by
P. cordatum, which contributed 59 (± 12.95 SD) mgCm-3 (66 % of biomass
in this group).
Synechococcus biomass was significantly higher in the combination treatment (reaching
59.9 ± 4.30 SDmgCm-3 at T36, Fig. 5d, Table 2) followed by
the high-temperature treatment (30 ± 5.98 SDmgCm-3, Table 2).
In both the high-pCO2 treatment and control Synechococcus biomass was low
(∼ 7 mgCm-3 in both treatments at T36), though an
initial significant response to high pCO2 was observed between T0 and T10
(Table 2). In all treatments and throughout the experiment, relative to
the other phytoplankton groups, biomass of picophytoplankton (Fig. 5e),
cryptophytes (Fig. 5f) and coccolithophores (Fig. 5g) remained low, though
there was a slight increase in picophytoplankton in the combination
treatment (11.26 ± 0.79 SDmgCm-3; Table 2).
Microzooplankton was dominated by Strombilidium spp. in all treatments throughout the
experiment, though biomass was low relative to the phytoplankton community
(Fig. 6). Following a decline from T0 to T10, microzooplankton biomass
increased in all but the high-CO2 treatment until T17 when biomass
diverged. The biomass trajectory maintained an increase in the control when
at T36 it was highest at ∼ 1.6 mgCm-3, 90 % higher
than the high-temperature treatment (0.83 mgCm-3). Microzooplankton
biomass was significantly lower in the high-CO2 treatment at T36 (z=-2.100, p=0.036) and undetected in the combination treatment at this time
point (Table 2).
FRRf-based photosynthesis–irradiance curve parameters for the
experimental treatments on the final day (T36).
Parameter
Control
SD
High temp
SD
High CO2
SD
Combination
SD
PmB
2.77
1.63
9.58
1.94
18.93
2.65
3.02
0.97
α
0.03
0.01
0.09
0.01
0.13
0.01
0.04
0.00
Ek
85.33
45.47
110.93
6.09
144.13
17.91
86.38
33.06
Results of generalised linear model testing for significant effects
of temperature, CO2 and temperature × CO2 on phytoplankton
photophysiology at T36; PmB (maximum photosynthetic rates),
α (light-limited slope) and Ik (light-saturated photosynthesis).
Significant results are in bold; * p<0.05, ** p<0.001,
*** p<0.0001.
Response variable
n
df
t value
p
Significance
PmB
High temp
12
8
7.353
< 0.0001
***
High pCO2
12
8
8.735
< 0.0001
***
High temp × high pCO2
12
8
-8.519
< 0.0001
***
α
High temp
12
8
13.03
< 0.0001
***
High pCO2
12
8
15.15
< 0.0001
***
High temp × high pCO2
12
8
-14.82
< 0.0001
***
Ek
High temp
12
8
2.018
0.0783
High pCO2
12
8
2.541
0.0347
*
High temp × high pCO2
12
8
-2.441
0.0405
*
Microzooplankton biomass (dominated by Strombilidium sp.) relative to total
phytoplankton biomass.
Fitted parameters of FRRf-based photosynthesis–irradiance curves for
the experimental treatments on the final experimental day (T36).
Chl a fluorescence-based photophysiology
At T36, FRRf PE parameters were strongly
influenced by the experimental treatments. PmB was significantly
higher in the high-pCO2 treatment (18.93 mgC (mg Chl a)-1m-3h-1),
followed by the high-temperature treatment
(9.58 mgC (mg Chl a)-1m-3h-1; Fig. 7, Tables 3 and 4). There was no
significant difference in PmB between the control and combination
treatments (2.77 and 3.02 mgC (mg Chl a)-1m-3h-1).
Light-limited photosynthetic efficiency (αB) also followed the same
trend and was significantly higher in the high-pCO2 treatment (0.13 mgC (mg Chl a)-1m-3h-1 (µmolphotonm-2s-1)-1)
followed by the high-temperature treatment (0.09 mgC (mg Chl a)-1m-3h-1 (µmolphotonm-2s-1)-1;
Tables 3 and 4). αB was low in both the
control and combination treatment (0.03 and 0.04 mgC (mg Chl a)-1m-3h-1 (µmolphotonm-2s-1)-1
respectively). The light saturation point of photosynthesis (Ek) was
significantly higher in the high-pCO2 treatment relative to all
treatments (144.13 µmolphotonm-2s-1), though
significantly lower in the combination treatment relative to both the high-pCO2 and
high-temperature treatments (Tables 3 and 4).
Discussion
Individually, elevated temperature and pCO2 resulted in the highest
biomass and maximum photosynthetic rates (PmB) at T36, when
nanophytoplankton dominated. The interaction of these two factors had little
effect on total biomass with values close to the ambient control, and no
effect on PmB. The combination treatment, however, exhibited the
greatest diversity of phytoplankton functional groups, with dinoflagellates
and Synechococcus becoming dominant over time.
Elevated pCO2 has been shown to enhance the growth and photosynthesis
of some phytoplankton species which have active uptake systems for inorganic
carbon (Giordano et al., 2005; Reinfelder,
2011). Elevated pCO2 may therefore lead to lowered energetic costs of
carbon assimilation in some species and a redistribution of the cellular
energy budget to other processes (Tortell et al.,
2002). In this study, under elevated pCO2 where the dominant group was
nanophytoplankton, the most abundant species was the haptophyte
Phaeocystis spp. Photosynthetic carbon fixation in Phaeocystis spp. is presently near saturation
with respect to current levels of pCO2 (Rost et al.,
2003). Dominance of this spp. under elevated pCO2 may be due to lowered
grazing pressure since microzooplankton biomass was lowest in the
high-CO2 treatment throughout the experiment. The increased biomass and
photosynthetic carbon fixation in this experimental community under elevated
pCO2 is due to the community shift to Phaeocystis spp. The increased biomass in
the high-temperature treatment (where microzooplankton biomass remained
stable between T17 and T36, though lower than the control) may be attributed
to enhanced enzymatic activities, since algal growth commonly increases with
temperature until after an optimal range
(Boyd
et al., 2013; Goldman and Carpenter, 1974; Savage et al., 2004). Optimum
growth temperatures for marine phytoplankton are often several degrees
higher than environmental temperatures (Eppley,
1972; Thomas et al., 2012). Nanophytoplankton also dominated in this
treatment and, while Phaeocystis spp. was not discriminated, no further classification
was made at a group/species level. Reduced biomass in the control from T24
onwards may be due to increased grazing pressure given the highest
concentrations of microzooplankton biomass were observed in the control.
Conversely, microzooplankton biomass declined significantly from T17 in the
combination treatment, indicating reduced grazing pressure while
phytoplankton biomass also declined from this time point. Nutrient
concentrations were not measured beyond T0 and we cannot therefore exclude
the possibility that differences in nutrient availability may have
contributed to observed differences between control, high-temperature and
high-CO2 treatments.
Chl a
Biomass in the control peaked at T25 followed by a decline to T36.
Correlated with this, Chl a also peaked at T25 in the control and declined to
3.3 mgm-3 by T27, remaining close to this value until T36. Biomass in
the combination treatment peaked at T20 followed by decline to T36, whereas
Chl a in this treatment declined from T20 to T25 followed by an increase at
T27 before further decline similar to the biomass. Chl a peaked in this
treatment again at T36 (6.8 mgm-3). We attribute the increase in Chl a
between T25 and T27 (coincident with an overall biomass decrease) to lower
species-specific carbon : Chl a ratios as a result of the increase in
dinoflagellates, Synechococcus and picophytoplankton biomass from T25. We speculate that
the decline in biomass under nutrient replete conditions in the combination
treatment was probably due to slower species-specific growth rates when
diatoms and dinoflagellates became more prominent in this treatment.
Carbon : Chl a ratios in diatoms and dinoflagellates have previously been demonstrated
to be lower than nano- and picophytoplankton
(Sathyendranath et al., 2009). This contrasts
the results reported in comparable studies as Chl a is generally highly
correlated with biomass
(e.g. Feng et al.,
2009). Similar results were reported however by
Hare et al. (2007), which indicates that Chl a may not always be a reliable proxy for biomass in mixed communities.
Biomass
This study shows that the phytoplankton community response to elevated
temperature and pCO2 is highly variable. pCO2 elevated to
∼ 800 µatm induced higher community biomass, similar to
the findings of Kim et al. (2006), whereas in other
natural community studies no CO2 effect on biomass was observed
(Delille et al.,
2005; Maugendre et al., 2017; Paul et al., 2015). A ∼ 4.5 ∘C
increase in temperature also resulted in higher biomass at T36
in this study, similar to the findings of
Feng et al. (2009)
and Hare et al. (2007), though elevated temperature has
previously reduced the biomass of natural nanophytoplankton communities in the
western Baltic Sea and Arctic Ocean
(Coello-Camba et al., 2014;
Moustaka-Gouni et al., 2016). When elevated temperature and pCO2 were
combined, community biomass exhibited little response, similar to the
findings of Gao et al. (2017),
though an increase in biomass has also been reported
(Calbet
et al., 2014; Feng et al., 2009). Geographic location and season also play
an important role in structuring the community and its response in terms of
biomass to elevated temperature and pCO2.
(Li et al., 2009; Morán et al.,
2010). This may explain part of the variability in responses observed from
studies on phytoplankton during different seasons and provinces.
Carbon : nitrogen
In agreement with others, the results of this experiment showed highest
increases in C : N under elevated pCO2 alone
(Riebesell et al.,
2007). C : N also increased under high temperature, consistent with the
findings of
Lomas
and Glibert (1999) and Taucher et al. (2015). It also increased when
pCO2 and temperature were elevated, albeit to a lesser degree, which
was also observed by Calbet et al. (2014), but contrasts other studies that have observed C : N being unaffected
by the combined influence of elevated pCO2 and temperature
(Deppeler and
Davidson, 2017; Kim et al., 2006; Paul et al., 2015). C : N is a strong
indicator of cellular protein content (Woods and Harrison, 2003)
and increases under elevated pCO2 and warming may lead to lowered
nutritional value of phytoplankton which has implications for zooplankton
reproduction and the biogeochemical cycling of nutrients.
Photosynthetic carbon fixation rates
At T36, under elevated pCO2, PmB was > 6 times
higher than in the control, but only one time point was measured, so we are
not able to make decisive conclusions.
Riebesell et al. (2007) and
Tortell et al. (2008) also reported an
increase in PmB under elevated pCO2. By contrast other
observations on natural populations under elevated pCO2 reported a
reduction in PmB
(Feng et al.,
2009; Hare et al., 2007). Studies on laboratory cultures have shown that
increases in temperature cause an increase in photosynthetic rates
(Feng et al.,
2008; Fu et al., 2007; Hutchins et al., 2007), similar to what we observed
in this study. In the combined pCO2 and temperature treatment, we found
no effect on PmB, which has also been observed in experiments on
natural populations
(Coello-Camba and
Agustí, 2016; Gao et al., 2017). This contrasts the findings of
Feng et al. (2009)
and Hare et al. (2007) who observed the highest
PmB when temperature and pCO2 were elevated simultaneously.
In this study, increases in αB and Ek under elevated
pCO2 and a decrease in these parameters when elevated pCO2 and
temperature were combined also contrast the trends reported by
Feng et al. (2009).
However, we should stress that, while our photophysiological measurements
support our observed trends in community biomass, they were made on a single
occasion at the end of the experiment. Future experiments should focus on
acquiring photophysiological measurements throughout.
Species-specific photosynthetic rates have been demonstrated to decrease
beyond their thermal optimum
(Raven and Geider, 1988), which
can be modified through photoprotective rather than photosynthetic pigments
(Kiefer and Mitchell, 1983). This may explain the
difference in PmB between the high-pCO2 and high-temperature
treatments (in addition to differences in nanophytoplankton community
composition in relation to Phaeocystis spp. discussed above), as the experimental
high-temperature treatment in this study was ∼ 4.5 ∘C
higher than the control.
There was no significant effect of combined elevated pCO2 and
temperature on PmB, which was strongly influenced by taxonomic
differences between the experimental treatments. Warming has been shown to
lead to smaller cell sizes in nanophytoplankton
(Atkinson
et al., 2003; Peter and Sommer, 2012), which was observed in the combined
treatment together with decreased nanophytoplankton biomass. Diatoms also
shifted to smaller species with reduced biomass, while dinoflagellate and
Synechococcus biomass increased at T36. Dinoflagellates are the only photoautotrophs with
form II RuBisCO (Morse et al., 1995), which has
the lowest carboxylation : oxygenation specificity factor among eukaryotic
phytoplankton (Badger et al., 1998), which may give
dinoflagellates a disadvantage in carbon fixation under present ambient
pCO2 levels. Phytoplankton growth rates are generally slower in surface
waters with high pH (≥ 9) resulting from photosynthetic removal of
CO2 by previous blooms and the associated nutrient depletion
(Hansen, 2002; Hinga, 2002). Though growth under high pH
provides indirect evidence that dinoflagellates possess carbon concentrating mechanisms (CCMs), direct
evidence is limited and points to the efficiency of CCMs in dinoflagellates
as moderate in comparison to diatoms and some haptophytes
(Reinfelder, 2011, and references therein). Given that
dinoflagellates accounted for just ∼ 20 % of biomass in the
combination treatment, exerting a minor influence on community
photosynthetic rates, further work is required to explain the lower
PmB under the combined influence of elevated pCO2 and
temperature compared to the individual treatment influences. We applied the
same electron requirement parameter for carbon uptake across all treatments,
though in nature and between species there can be considerable variation in
this parameter (e.g. 1.15 to 54.2 mole-(molC)-1;
Lawrenz et
al., 2013), which can co-vary with temperature, nutrients, Chl a, irradiance
and community structure. Better measurement techniques at quantifying this
variability are necessary in the future.
Community composition
Phytoplankton community structure changes were observed, with a shift from
dinoflagellates to nanophytoplankton which was most pronounced under single
treatments of elevated temperature and pCO2. Amongst the
nanophytoplankton, a distinct size shift to smaller cells was observed in
the high-temperature and combination treatments, while in the high-pCO2
treatment Phaeocystis spp. dominated. Under combined pCO2 and temperature from T24
onwards, however, dinoflagellate and Synechococcus biomass increased and nanophytoplankton
biomass decreased. An increase in pico- and nanophytoplankton has previously
been reported in natural communities under elevated pCO2
(Bermúdez
et al., 2016; Boras et al., 2016; Brussaard et al., 2013; Engel et al.,
2008) while no effect on these size classes has been observed in other
studies
(Calbet
et al., 2014; Paulino et al., 2007).
Moustaka-Gouni et al. (2016) also found
no CO2 effect on natural nanophytoplankton communities but increased
temperature reduced the biomass of this group.
Kim et al. (2006) observed a shift from
nanophytoplankton to diatoms under elevated pCO2 alone while a shift
from diatoms to nanophytoplankton under combined elevated pCO2 and
temperature has been reported (Hare et al.,
2007). A variable response in Phaeocystis spp. to elevated pCO2 has also been
reported with increased growth
(Chen et al., 2014; Keys et al.,
2017), no effect (Thoisen et al., 2015) and decreased growth
(Hoogstraten et
al., 2012) observed. Phaeocystis spp. can outcompete other phytoplankton and form
massive blooms (up to 10 gCm-3) with impacts on food webs, global
biogeochemical cycles and climate regulation (Schoemann et
al., 2005). While not a toxic algal species, Phaeocystis spp. are considered a harmful
algal bloom (HAB) species when biomass reaches sufficient concentrations to
cause anoxia through the production of mucus foam which can clog the feeding
apparatus of zooplankton and fish (Eilertsen and Raa, 1995).
Recently published studies on the response of diatoms to elevated pCO2
and temperature vary greatly. For example,
Taucher et al. (2015)
showed that Thalassiosira weissflogii incubated at 1000 µatm pCO2 increased growth by
8 % while for Dactyliosolen fragilissimus, growth increased by 39 %; temperature elevated by +5 ∘C
also had a stimulating effect on T. weissflogii but inhibited the growth
rate of D. fragilissimus; and when the treatments were combined growth was enhanced in T. weissflogii but
reduced in D. fragilissimus. In our study, elevated pCO2 increased biomass in diatoms
(time dependent), but elevated temperature and the combination of these
factors reduced the signal of this response. A distinct size shift in diatom
species was observed in all treatments, from the larger Coscinodiscus spp., Pleurosigma and
Thalassiosira subtilis to the smaller Navicula distans. This was most pronounced in the combination treatment
where N. distans formed 89 % of diatom biomass. Navicula spp. previously exhibited a
differential response to both elevated temperature and pCO2. At +4.5 ∘C and 960 ppm CO2
Torstensson et al. (2012) observed
no synergistic effects on the benthic Navicula directa. Elevated temperature increased
growth rates by 43 % while a reduction of 5 % was observed under
elevated CO2. No effects on growth were detected at pH ranging from 8 to 7.4
units in Navicula spp. (Thoisen et al., 2015), while there was
a significant increase in growth in N. distans along a CO2 gradient at a shallow
cold-water vent system (Baragi et al.,
2015).
Synechococcus grown under pCO2 elevated to 750 ppm and temperature elevated by 4 ∘C
resulted in increased growth and a 4-fold increase in
PmB (Fu et al., 2007) which is similar to
the results of the present study.
The combination of elevated temperature and pCO2 significantly
increased dinoflagellate biomass to 17 % of total biomass. This was due
to P. cordatum, which increased biomass by more than 30-fold from T0 to T30 (66 % of
dinoflagellate biomass in this treatment). Despite the global increase in
the frequency of HABs few studies have focussed on the response of
dinoflagellates to elevated pCO2 and temperature. In laboratory studies
at 1000 ppm CO2, growth rates of the HAB species Karenia brevis increased by
46 %; at 1000 ppm CO2 and +5 ∘C temperature its growth
increased by 30 % but was reduced under elevated temperature alone
(Errera
et al., 2014). A combined increase in pCO2 and temperature enhanced
both the growth and PmB in the dinoflagellate Heterosigma akashiwo, whereas in
contrast to the present findings only pCO2 alone enhanced these
parameters in P. cordatum (Fu et al., 2008).