Biodiversity of phytoplankton is important for ecosystem
stability and marine biogeochemistry. However, the large-scale patterns of
diversity are not well understood and are often poorly characterized in
terms of statistical relationships with factors such as latitude,
temperature and productivity. Here we use ecological theory and a global
trait-based ecosystem model to provide mechanistic understanding of patterns
of phytoplankton diversity. Our study suggests that phytoplankton diversity
across three dimensions of trait space (size, biogeochemical function and
thermal tolerance) is controlled by disparate combinations of drivers: the
supply rate of the limiting resource, the imbalance in different resource
supplies relative to competing phytoplankton demands, size-selective
grazing and transport by the moving ocean. Using sensitivity studies we
show that each dimension of diversity is controlled by different drivers.
Models including only one (or two) of the trait dimensions will have
different patterns of diversity than one which incorporates another trait
dimension. We use the results of our model exploration to infer the controls
on the diversity patterns derived from field observations along meridional
transects in the Atlantic and to explain why different taxa and size classes
have differing patterns.
Introduction
Phytoplankton are an extremely diverse set of microorganisms spanning more
than 7 orders of magnitude in cell volume (Beardall et al., 2008) and an
enormous range of cell morphologies, bio(geo)chemical functions, elemental
requirements and trophic strategies. This range of traits play a key role in
regulating the biogeochemistry of the ocean (e.g. Cermeño et al., 2008;
Fuhrman, 2009), including the export of organic matter to the deep ocean
(Falkowski et al., 1998; Guidi et al., 2009), which is critical in oceanic
carbon sequestration and contributes to modulation of atmospheric CO2
levels and climate. Biodiversity is also important for the stability of the
ecosystem structure and function (e.g. McCann, 2000; Ptacnik et al., 2008;
Cermeño et al., 2016), though the exact nature of this relationship is
still debated. Studies suggest that diversity loss appears to coincide with
a reduction in primary production rates and nutrient utilization efficiency
(Cardinale et al., 2011; Reich et al., 2012), thereby altering the
functioning of ecosystems and the services they provide. Diversity is
important, but what factors control diversity still remains an elusive
problem.
Numerous studies have attempted to understand or predict observed patterns
of biodiversity or species richness of marine phytoplankton by correlating
with factors such as temperature and latitude (see e.g. Hillebrand and
Azovsky, 2001; Hillebrand, 2004; Irigoien et al., 2004; Smith, 2007;
Rodriguez-Ramos et al., 2015; Powell and Glazier, 2017; Righetti et al., 2019).
The metabolic theory of ecology posits that temperature could control the
probability of mutation and speciation leading to more diversity at higher
temperatures (see e.g. Allen et al., 2007). However, a recent study suggests a
unimodal statistical relationship between diversity and temperature
(Righetti et al., 2019). Studies have also proposed a latitudinal dependence
of diversity (e.g. Chust et al., 2013), though the shape of that dependence is
unclear. Chaudhary et al. (2016) for instance suggest a bimodal distribution,
and a study of the Cenozoic fossil records suggests that the diversity of
diatoms may actually have increased towards the poles (Powell and Glazier,
2017). However, Rodriguez-Ramos et al. (2015) found little evidence of a
relationship between nano- and micro-phytoplankton species richness and
either temperature or latitude after enforcing consistency of datasets.
Additionally, there is evidence suggesting that increased dispersal (up to a
point) could increase diversity (Matthiessen and Hillebrand, 2006), and
diversity was related to mesoscale features in a study in the North
Atlantic (Mousing et al., 2016).
There has been a debate as to how productivity links to diversity (see e.g.
review by Smith, 2007). Again, by standardizing datasets to correct for
differences in sampling efforts, only a weak (or no) correlation between
phytoplankton diversity and productivity emerges from basin-scale datasets
(Cermeño et al., 2013; Rodriguez-Ramos et al., 2015), suggesting that
previously reported connections might be skewed by sampling biases (Cermeño
et al., 2013). However, it also appears that biotic factors can potentially
impact diversity: the importance of top-down control has been suggested by
the experiments of Worm et al. (2002). Multiple factors appear to be likely
important, but correlations with multiple co-occurring environmental factors
do not satisfactorily explain diversity patterns (e.g. Rodriquez-Ramos et al.,
2015). There remains no holistic understanding of phytoplankton diversity
and its drivers.
Recent theoretical work (e.g. Vallina et al., 2014b, 2017; Terseleer et al.,
2014) suggests that breaking diversity down into traits can be useful.
Vallina et al. (2017) also suggested that a variety of traits respond
differently to environmental factors. The importance of multiple
phytoplankton traits in setting community structure has previously been
expounded (e.g. Litchman et al., 2010; Acevedo-Trejos et al., 2015). Theory
and models have considered several different phytoplankton traits and
environmental drivers to explain diversity. In one study, different
temperature dependencies and nutrient affinity trade-offs allowed
phytoplankton to have similar lowest-subsistence nutrient requirements (as
described in Tilman, 1977, 1982) that allowed sustained coexistence (Barton
et al., 2010). Other studies explored the importance of top-down control
(Prowe et al., 2012; Vallina et al., 2014a; Ward et al., 2014). A positive
relationship between diversity and productivity was found when a model
captured only different size classes but no temperature differences (Ward
et al., 2012, 2014). A series of studies also showed that mesoscale features
and dispersal enhanced diversity (Lévy et al., 2014, 2015; Clayton et al.,
2013), also revealing that hot spots of diversity occurred in regions of
high mixing (Clayton et al., 2013).
In this study we will almost exclusively consider diversity in terms of
“richness”, the number of locally coexisting species. This definition is
often referred to as alpha diversity. We focus on richness here as the
ecological theories we will use explain coexistence rather than other
common metrics of diversity such as Shannon index or evenness. We also do
not consider species present at extremely low population densities, the
so-called rare biosphere.
In this study we seek to disentangle the multiple, sometimes conflicting,
results from models and observational studies and seek to explain at least
some of the controls on diversity. We employ ecological theories and a
trait-based global model. We use observed patterns of diversity along
meridional transects in the Atlantic as motivation and as illustration of
the utility of this study. By using a model and theory, we explore the
mechanistic drivers of the modelled diversity. We conduct sensitivity
experiments to test the intuition that the theoretical framework provides.
However, on a cautionary side, this study tells us about the diversity in
the model world. Though our model is complex, it still missed many of the
traits of the real ocean microbial communities.
This study synthesizes much of the understanding that we have gained through
previous modelling and theoretical studies (e.g. Dutkiewicz et al.,
2009, 2012, 2014; Ward et al., 2013, 2014; Lévy et al., 2014). What is unique
here is bringing these all together, addressing disparities in previous
works and providing insight into the multiple interacting mechanisms that
drive diversity. We find that this can only be done by acknowledging that
diversity along different axes of traits (e.g. size, biogeochemical
function, thermal norms) each has their own set of drivers. And this is turn
suggests that no single or combined set of environmental variables will be
able to explain patterns of diversity in the ocean.
As an illustrative example from field observations, we used data of species
composition, abundance, and cell size in the range of nano- and
micro-phytoplankton from samples collected in marine pelagic ecosystems. The
data come from transects sampled during September to October 1995 (AMT-1),
April to May 1996 (AMT-2), September to October 1996 (AMT-3) and April–May 1997 (AMT4). These cruises crossed the same regions of the Atlantic Ocean by
a similar route. At each station, two replicate seawater samples were
preserved, one with 1 % buffered formalin (to preserve calcite
structures) and the other with a 1 % final concentration Lugol iodine
solution. After sedimentation of a subsample for 24 h (Utermöhl's
technique), cells were measured and counted with an inverted microscope at
×187, ×375 and ×750 magnifications to cover the full ensemble of nano- and
micro-phytoplankton and identified to the lowest possible taxonomic level
(usually species level). The volume of water samples used for sedimentation
varied between 10 and 100 mL, according to the overall abundance of
phytoplankton as shown by the fluorometer. At least 100 cells of each of the
more abundant species were enumerated. Here diversity is determined as
richness, which here is defined as the number of species detected in
sample volumes in the range 10–100 mL. Results from the coccolithophore and
diatom species richness from this dataset have previously been shown in
Cermeño et al. (2008). Cell volume was calculated by assigning different
geometric shapes that were most similar to the real shape of each
phytoplankton species. A mean cell volume was assigned for each
phytoplankton species. Cells were separated into diatoms, coccolithophore
and dinoflagellate groups. Here these data are used to determine total
species richness (number of coexisting species) of all the nano- and
micro-eukaryotes (Fig. 1a) but also species richness within diatom,
dinoflagellate and coccolithophore groups (Fig. 1b), as well as number of
species in three size classes (2–10, 10–20, >20µm, Fig. 1c). Given how these data are compared to model output
(see below) we purposely neglect the rare biosphere, so we do not attempt any
techniques such as rarefaction to account for rare species.
Nano- and micro-eukaryote normalized richness in the
Atlantic. Left: richness (number of coexisting species) normalized to the
maximum along the Atlantic Meridional Transects (AMT) 1, 2, 3, 4 for microscopy
counts (see methods). Right: normalized annual mean richness from model.
(a, d) All diatoms, coccolithophores and dinoflagellates together; (b, e) each functional groups separately (red: diatoms, dark blue:
coccolithophores, purple: dinoflagellates); (c, f) three size classes (light
blue: 2–10 µm, black: 10–20 µm, green: <20µm).
In left panels, circles are means of the four transects (two in May, two in September) within 4∘ latitude bins; the vertical lines indicate the
range within each bin. The maximum number used to normalize the plots is
provided in each panel. Model pico-phytoplankton and diazotrophs are not
included in the model analysis as they were not included in the
observations. Maps show the cruise track of the AMTs, and the model
includes the annual mean normalized richness of the diatoms,
coccolithophores and dinoflagellates together.
Numerical model
The model follows from Dutkiewicz et al. (2015a) in terms of biogeochemistry,
plankton interactions, and transmission of light as described by the tables
and equations of that paper. However, the types of phytoplankton and
zooplankton differ in that they include greater diversity. Here we briefly
provide an overview of the model and detailed descriptions of the more
complex ecosystem. More details of pertinent parameterizations and
parameters can be found in Text S1, Figs. S1 and
S2, and Tables S1 and S2 in the Supplement; the full set of equations and
remainder of biogeochemical parameters can be found in Dutkiewicz et al. (2015a).
The biogeochemical/ecosystem model resolves the cycling of carbon,
phosphorus, nitrogen silica, iron, and oxygen through inorganic, living,
dissolved and particulate organic phases. The biogeochemical and biological
tracers are transported and mixed by the MIT general circulation model
(MITgcm, Marshall et al., 1997) constrained to be consistent with altimetric and
hydrographic observations (the ECCO-GODAE state estimates, Wunsch and
Heimbach, 2007). This three-dimensional configuration has a coarse resolution
(1∘× 1∘ horizontally) and 23 levels ranging from 10 m in
the surface to 500 m at depth. At this horizontal resolution, the model does
not capture mesoscale features such as eddies and sharp fronts, a limitation
of the model that must be kept in mind when considering the results.
We use a complex marine ecosystem that incorporates 350 phytoplankton types
that can be described in three “dimensions” of trait space (schematically
shown in Fig. 2): size, biogeochemical function and temperature tolerance.
Within the “size” dimension we include 16 size classes spaced uniformly in
log space from 0.6 to 228 µm equivalent spherical diameter
(ESD). Within the “biogeochemical function” dimension we resolve diatoms
(that utilize silicic acid), coccolithophores (that calcify), mixotrophs
(that photosynthesize and graze on other plankton), nitrogen-fixing
cyanobacteria (diazotrophs) and pico-phytoplankton. We resolve 4 size
classes of pico-phytoplankton (from 0.6 to 2 µm ESD), 5 size classes
of coccolithophores and diazotrophs (from 3 to 15 µm ESD), 11 size
classes of diatoms (3 to 155 µm ESD), and 10 mixotrophic
dinoflagellates (from 7 to 228 µm ESD). Additionally, we resolve a
“temperature norm” trait axis, where phytoplankton growth rates are
defined over a specific range of temperatures (Fig. 3) by an empirically
motivated function (e.g. Thomas et al., 2012; Boyd et al., 2013). We include
10 different norms. Thus for any size class within a functional group there
are 10 different unique phytoplankton types (as demonstrated schematically
in Fig. 2) with a different range of temperatures over which the cells will
grow. Warmer-adapted types are assumed to grow faster as suggested
empirically (Eppley, 1972; Bissenger et al., 2008) and from enzymatic
kinetics (Kooijman, 2000). In total we resolve 350 phytoplankton types
within 16 size classes, 5 biogeochemical functional groups, and 10
temperature norms.
Schematic of the three dimensions of trait space: size
classes, biogeochemical functional groups and thermal norms. There are 16
size classes, 5 functional groups and 10 thermal norms. In all there are 350
individual phytoplankton types. However, the 3 largest size classes go
extinct, and as such here (and in other figures) we show only 13 size
classes.
Growth as a function of temperature. Shown are the 10
thermal norms (unitless), each with a different colour. The function used
here is from Dutkiewicz et al. (2015b) and is discussed further in
the Supplement.
Phytoplankton parameters influencing maximum growth rate, nutrient affinity,
grazing and sinking are parameterized as a power function of cell volume:
aVb (following Ward et al., 2012; see Text S1.2 and Table S2). Thus many size classes can be described by just two coefficients
(a, b) per parameter. Maximum growth rate is parameterized (i.e. the a and b in the
above equation) as distinct between functional groups (as suggested by
observations in Fig. 4a; see also Buitenhuis et al., 2008; Sommer et al., 2017).
The smallest diatoms (3 µm) have the highest maximum growth rates.
Plankton smaller than 3 µm have an increase in growth rate with size,
and those larger than 3 µm have a decrease in growth rate with size.
This unimodal distribution has been observed (e.g. Raven, 1994; Bec et al.,
2008; Finkel et al., 2010; Marañón et al., 2013; Sommer et al., 2017) and
explained as a trade-off between replenishing cell quotas versus
synthesizing new biomass (Verdy et al., 2009; Ward et al., 2017). There are
also specific differences between functional groups in cell elemental
stoichiometry and palatability to grazers (we assume that the hard coverings of diatoms and coccolithophores
deter grazers; see e.g. Monteiro et al.,
2016; Pančić et al., 2019). The smallest phytoplankton have the
highest affinity for nutrients (Edwards et al., 2012) as a result of the
lower surface-to-volume ratio found in larger cells (Kiorboe, 1993; Raven,
1994).
Model parameter guide by laboratory studies.
Phytoplankton maximum growth rate (a) and R*(b) as a function of
cell size. In (a) small symbols indicate laboratory studies normalized to
20 ∘C and large symbols indicate the model size/functional groups. Colour of symbols denotes different functional groups: red circle, diatoms; purple
diamond, mixotrophic dinoflagellates; dark blue plus, coccolithophores;
light-blue cross, diazotrophs; green square, pico-phytoplankton. In (b),
R∗=kRMμmax-M, where M=0.5 1 d-1 (see Appendix A).
Data compilations of concurrent size and growth in (a) are from Tang (1995),
Marañón et al. (2013), Sarthou et al. (2005) and Buitenhuis et al. (2008).
Additional data are derived from separate measurement of size and growth:
these are shown as light lines centred at the mean and arms covering range.
These are for the pico-prokaryotes (green) Prochlorococcus and Synechococcus (Morel et al., 1993;
Johnson et al., 2006; Christaki et al., 1999; Moore et al., 1998; Agawin and
Agustí, 1997) and the diazotrophs (light blue) Crocosphaera and Trichodesmium (Garcia and
Hutchins, 2014; Webb et al., 2009; Wilson et al., 2017; Bergman et al., 2013;
Boatman et al., 2017; Breitbarth et al., 2008; Hutchins et al., 2007; Kranz et
al., 2010; Shi et al., 2012).
The model includes spectral irradiances, and each functional group has
different spectra for absorption (as a result of group-specific accessory
pigments) and scattering of light. The absorption spectra are flatter with
larger sizes following Finkel and Irwin (2000) to account for self-shading, and
scattering spectra are also influenced by size following Stramski et al. (2001) (see Text S1.3, Fig. S1). The simulation
uses Monod kinetics, and C : N : P : Fe stoichiometry is constant over time (though it differs between phytoplankton groups). Chl a for each phytoplankton type varies in time and space depending on light, nutrient and temperature conditions following Geider et al. (1998). Following empirical evidence,
mixotrophic dinoflagellates are assumed to have lower maximum photosynthetic
growth rates than other phytoplankton of the same size (Tang, 1995; Fig. 4a).
We resolve 16 size classes of zooplankton (from ESD 6.6 to 2425 µm) that graze on plankton (phyto- or zoo-) 5 to 20 times smaller than
themselves, but preferentially 10 times smaller (Hansen et al., 1997;
Kiorboe, 2008; Schartau et al., 2010). Maximum grazing rate is a function of
size (following Hansen et al., 1997), though the four smallest grazers are
assumed to have the same maximum grazing rates (Fig. S2). Here
the smallest grazers do not have a clear difference in grazing related to
size (following the data compilation of Taniguchi et al., 2014). Mixotrophic
dinoflagellates also graze on plankton with the same predator–prey ratios as
the zooplankton and have size-dependent maximum grazing rates. However,
they have lower maximum grazing rates than zooplankton of the same size
(Jeong et al., 2010, Fig. S2). We use a Holling III grazing
function (Holling, 1959). Sensitivity studies with a Holling II
parameterization show that the results here are not sensitive to this
choice.
We perform a “default” simulation (EXP-0) for 10 years. The ecosystem
quickly (within 2 years) reaches a quasi-steady state. Here we show results
from the 5th year of the simulation but note that the patterns of
biogeochemical and ecologically relevant output and diversity are not
significantly different if we instead used the 10th year. We also
conducted a series of sensitivity experiments, where we alter either
physical or ecosystem assumptions to provide evidence for the controls of
diversity (Table 1).
Table of sensitivity experiments.
EXP-0EXP-1EXP-2EXP-3Number grazers1611616Nutrient requirements of functional groupsDifferingDifferingSameDifferingHorizontal transport of planktonYesYesYesNo
As mentioned in the introduction, in this study we primarily discuss
diversity in term of “richness”, defined here as the number phytoplankton
types that coexist at any location above a biomass threshold. We, in
particular, look at the annual mean of the instantaneous surface richness
(though see Supplement for examples with depth). Technically, we use a
threshold value (10-5 mmol C m-3) to determine whether a
phytoplankton type is present or absent in a given community. This value
would convert to about 10 Prochlorococcus cells mL-1 (typical oligotrophic waters are above
103 cells mL-1), or only a tiny fraction (10-4) of a larger diatom cell per millilitre. Thus, this definition neglects rare species, often at abundances in
the order of individuals per litre, that would be difficult to separate from
numerical noise. This is why we do not account for the rare species in the
AMT observations discussed above. The value of richness can be altered
depending on the threshold chosen, but the patterns and results discussed
below remain robust. We also emphasize that the level of richness that the
model captures, though large for a model, is orders of magnitude lower than
the real ocean. Thus, this is not a fully comprehensive study of diversity
or species richness but does nevertheless provide a promising avenue for
understanding some of the controls on diversity.
ResultsDiversity observations along the AMT
The four Atlantic Meridional Transect (AMT) cruises provide a large-scale
consistent dataset of phytoplankton diversity including microscopic counts
of diatoms, coccolithophores and dinoflagellates. Such microscopic
measurements depict species richness patterns of abundant taxa but miss much
of the rare biosphere. This dataset shows distinct large-scale patterns (Fig. 1a), with high richness (as determined by number of coexisting species) on the northern edge of the Southern Ocean and in the Canary upwelling, low
richness in the subtropical gyres, and slightly elevated richness in the
equatorial region.
However, the patterns of richness are very different if we look only within
a single functional group (e.g. diatoms, Fig. 1b) or within a specific size
class (Fig. 1c). Diatoms exhibit higher diversity in the Southern Ocean than
the other functional groups, while the diversity of coccolithophores and
dinoflagellates is much more uniform across the transects. Among size
classes, the smallest size category (2–10 µm) has the highest diversity,
while there is lower and more regionally varying diversity in the larger
size categories, with some regions having none of the largest size class
(>20µm). This suggests that the controlling mechanism(s)
on, for instance, diatom diversity is different to those controlling
coccolithophore diversity, which also differs to what determines the
diversity within different size classes. Indeed, modelling and theoretical
work (e.g. Vallina et al., 2014b, 2017; Terseleer et al., 2014) have suggested
that breaking diversity down into traits can be insightful. Thus, a starting
point of our study is to separate out different dimensions of diversity.
Numerical model
Model development was guided by evaluating against a range of in situ and
satellite-derived observations (see Text S2 and Figs. S3–S8). We refer the reader to the fuller evaluation in the
Supplement text but provide a brief version here. The model captures the
patterns of low and high Chl a seen in the satellite estimate (Fig. S4), though it underestimates the Chl a in the subtropical gyres and
overestimates it in the high latitudes. However, we note that
satellite-estimated Chl a has large uncertainties (Moore et al., 2009)
especially in the Southern Ocean (e.g. Johnson et al., 2013). The coarse
resolution of the model does not capture important physical processes near
coastlines, and the lack of sedimentary and terrestrial supplies of nutrients
and organic matter leads to Chl a being too low in these regions. The
underestimation of Chl a in the gyres is also seen when comparing the model
to the observations of surface Chl a along the AMT (Figs. 5b,
S6b). The model does capture the drawdown of nutrients in the gyres and the
large increase in nutrient concentrations in the Southern Ocean (Figs. 5a,
S6a). However, the model overestimates the amount of
silicic acid in this ocean (seen also in the global evaluation, Supplement
Fig. S3), likely a reflection of Si : C of the model diatoms being too low in
the region.
Observations and model output along the Atlantic
Meridional Transect (AMT). (a, b) Nutrients (black: nitrate,
mmol N m-3; green: phosphate, ×16 mmol P m-3; light blue: silicic
acid, mmol Si m-3); (c, d) Chl a (mg Chl m-3); (e, f) phytoplankton biomass (mg C m-3; red: diatoms; blue: coccolithophores;
purple: dinoflagellates). Observations (left panels) are means (circles) for
the four AMT cruises (two in May, two in September; see transects in Fig. 1 maps) in
4∘ bins; the vertical lines show the range within each bin. Model
results are annual means along the AMT cruise track.
The model individual types have plausible ranges (four representative species
shown in Fig. S9) compared to distributions determined from
thermal niches (e.g. Thomas et al., 2012) and statistical techniques from
sparse observations (e.g. Barton et al., 2016). The model captures biomass in
almost all size classes (Figs. 6, S10a), though the largest
size classes are underestimated. Traits not included in the model (e.g.
buoyancy regulation, chain formation, symbiosis) are possibly more important
for maintaining these large size classes. The model has biomass in all
temperature norms (Figs. 6, S10c), though with lower biomass
in the coldest- and warmest-adapted ones, suggesting the model parameterization
covers an adequate range of norms. There are some interesting eliminations
(which match observations) such as coldest-adapted smallest
pico-phytoplankton and diazotrophs, as well as large warmest-adapted diatoms. The
phytoplankton are complemented by a range of size classes of zooplankton
(Fig. S11).
Model phytoplankton types biomass and range. (a) Global
integrated biomass (TgC); (b) areal extent of the type (1012 km2).
Types are arranged by functional group as indicated by colour and labels at
the top of the graph, by size classes (equivalent spherical diameter, ESD)
as labelled below the graph, and by thermal norms from cold adapted to warm
adapted from left to right in between vertical dotted lines. The text
(A, B, C, D) in panel (a) refers to representative types whose distributions
are shown in Fig. S9.
We evaluate the model's ability to capture the size distribution of
phytoplankton as derived from satellite products (Figs. 7a,
S4, S5). Here we capture the ubiquitous pico-phytoplankton and the
limitation of the larger size classes to the more productive regions. The
model pico-phytoplankton size class Chl a is potentially slightly too low
and the nano size class too high. Though we note that if we set the
pico/nano break at the fifth model size class (just under 3 µm)
instead at the fourth (2 µm) size class, the relative values are much
more in line with the satellite product. We suggest that the satellite
product division might not be that exact. The micro-size class matches in
location to the satellite product but is too low as discussed above.
Comparison to observations. (a) Sizes classes: Chl-a
concentration (mg Chl m-3) in pico-phytoplankton (<2µm), nano-phytoplankton (2–20 µm) and micro-phytoplankton
(>20µm) from (left) a satellite-derived estimate
(Ward, 2015) and (right) default model (0–50 m); and (b) functional group
(top) default model (0–50 m) and (bottom) data compilation (MAREDAT,
Buitenhuis et al., 2013) in carbon biomass (mg C m-3). Note the difference
in units for (a) and (b), chosen to match the appropriate observations. For
the MAREDAT databases: pico-phytoplankton (Buitenhuis et al., 2012);
coccolithophores (O'Brien et al., 2013); diazotrophs (Luo et al., 2012); diatoms
(Leblanc et al., 2012). There was no MAREDAT dataset for dinoflagellates.
We also compare the model functional group distribution to the compilation
of observations (Fig. 7b, MAREDAT, Buitenhuis et al., 2013, and references
therein). Though the observations are sparse, we do capture the ubiquitous
nature of the pico-phytoplankton, the limited domain of the diazotrophs
(including the observed lack of diazotrophs in the South Pacific Gyre), and the pattern of enhanced diatom biomass in high latitudes and low in subtropical
gyres. We overestimate the coccolithophore biomass relative to MAREDAT in
many regions, but we note that the conversion from cells to biomass in that
compilation was estimated to have uncertainties as much as several hundred percent (O'Brien et al., 2013). The MAREDAT compilation did not include a category
for dinoflagellates. We also compare the model biomass of diatoms,
coccolithophores and dinoflagellates along the AMT, though we note that the
conversion from cell counts to biomass in the observations has significant
uncertainties. The model captures the much lower biomass of diatoms in the
subtropical gyres than the other two functional groups and higher biomass in the Southern Ocean. Coccolithophore biomass is too low in the Southern Ocean in
the model, likely due to the modelled smallest diatom being parameterized as
too competitively advantaged, but it compares better in the rest of the
transect than the MAREDAT comparison above suggested.
In this study we mostly consider richness, the number of coexisting types,
as a metric of diversity. However, we do discuss the Shannon index (another
commonly used metric of diversity) later in the text. We will refer to
“total richness”, i.e. the number of coexisting phytoplankton types, out
of the 350 initialized in the model, at any location (Fig. 8a). Here we
specifically look at the annual mean richness in the surface layer, which is
a good indicator of the diversity within the mixed layer (Fig. S12). We find lowest richness in the subtropical gyres and highest
associated with the western boundary currents.
Model diversity measured as annual mean normalized
richness in the surface layer. Normalization is by the maximum value for
that plot (value noted in the bottom right of each panel). (a) Total
richness determined by number of individual phytoplankton types that
coexist at any location; (b) size class richness determined by number of
coexisting size classes; (c) functional richness determined by number of
coexisting biogeochemical functional groups; (d) thermal richness
determined by number of coexisting temperature norms. Total richness (a) is
a (complex) multiplicative function of the three sub-richness categories (b–d). At first glance it may seem that thermal norm diversity (d) is most
important, but this is only because our eyes are drawn to the hot spots. In
reality, total diversity patterns (a) are strongly impacted by all three
dimensions of diversity (see main text and sensitivity studies).
The model is designed to allow for richness within specific functional
groups and size classes. A unique feature about this study is a comparison
to the richness found in the AMT data (Figs. 1, S7, S8). The
model captures the low and high patterns of the AMT observations, though
it underestimates the diversity in the subtropical gyres. In these regions it
is likely that trait axes (e.g. symbiosis and colony formation) not
captured in the model provide additional means for phytoplankton to
coexist. Excitingly the model also captures the differences in the
diversity within functional groups and in size classes. Diatoms have much
larger diversity in the Southern Ocean than the other functional groups,
while coccolithophore and mixotrophic-dinoflagellate diversity is much
more uniform across the transect. The model captures the much higher
diversity within the smallest size category (2–10 µm) and the lower and
much more regionally varying diversity in the larger size category,
including the lack of diversity in the largest size class (>20µm) in the subtropical gyres.
It is instructive to also consider richness along each of the dimensions of
trait space. The number of size classes (irrespective of functional group or
thermal norm) that coexist in any location will be referred to as size
class diversity (Fig. 8b). We find that in high latitudes and along the
Equator many size classes are present, while in the subtropical gyres only
few small-sized classes survive (Figs. 7a, S10a). We find
that there are different patterns of richness when looking along the two
other axes of traits (Figs. 8c, d; S10b, c). Richness of
biogeochemical functional groups is highest in the mid-latitudes, strongly
linked to the distributions of diazotrophs (Figs. 7b, S10b).
In contrast, the diversity within temperature norms is maximum in the
western boundary currents, in particular the Gulf Stream and Kuroshio, and
high in coastal upwelling regions (e.g. off Peru and Canary) and along the
northern boundary of the Southern Ocean.
The total richness is a complex integral function (i.e. multiplicative) of
the three different trait dimensions. At first glance total diversity (Fig. 8a) may look most like the thermal norm diversity (Fig. 8d), but this is
mostly because our eyes are drawn to the hot spots. In reality, total
diversity patterns are strongly impacted by all three dimensions of
diversity as will be shown more clearly by the sensitivity experiments
discussed later. We find that some trait dimensions are more (or less)
important in different regions. For instance, thermal norm richness leads to
the total richness hot spots (Fig. 8a) in the western boundary currents and
coastal upwelling regions. While reduction in functional groups and thermal
norms counteracts the increase in size classes in the Southern Ocean, all
three dimensions together lead to the lowest total richness captured in the
middle of the subtropical gyres.
Understanding the dimensions of diversity: model and theoretical
framework
None of these three dimensions can, in isolation, explain the controls on
the total richness. Nor can we a priori understand the total richness. By
using ecological theories and a series of sensitivity experiment (Table 1),
we can begin to understand the mechanisms setting the different dimensions
of diversity individually. Here, we step through each of the dimensions.
The theoretical frameworks are presented in Appendix A and are informed
from the seminal work of Tilman (1977, 1982) and Armstrong (1994). Resource
competition theory (Tilman, 1977, 1982) has been extensively used in
theoretical and experimental studies (e.g. Sommer, 1986; Grover, 1991a, b;
Huisman and Weissing, 1994; Schade et al., 2005; Miller et al., 2005; Wilson et al.,
2007; Agawin et al., 2007; Snow et al., 2015) as well as in linking to numerical models (Dutkiewicz et al., 2009, 2012, 2014; Ward et al., 2013) to explain
aspects of community structure. The theoretical underpinnings of
size-selected grazing (Armstrong, 1994) have similarly been used in many
studies (e.g. Lampert, 1997; Kiorboe, 1993, 2008; Schartau et al., 2010; Ward
et al., 2014; Acevedo-Trejos et al., 2015). Appendix A and the insight we
develop in the rest of this section are in some sense a synthesis of many
prior studies. Here, these theories are specifically directed at
understanding diversity patterns, something that to our knowledge has not
been done before.
Size class diversity
We find that the richness of cell sizes increases with the supply rate of
the limiting nutrient (Fig. 9). Theoretical predictions and previous model
studies suggest that this should be the case when the resource requirements
of phytoplankton increase with increasing size (Appendix A, Armstrong, 1994;
Ward et al., 2014; Follows et al., 2018). In the nomenclature of resource
supply theory (Tilman, 1977), R* of a phytoplankton type is the minimum
resource concentration required for it to survive at a steady state. In the
absence of grazing, R∗=kRMμmax-M, where KR is
the resource half-saturation constant, μmax is the maximum growth rate
and M is a loss rate (see Appendix A). The phytoplankton with the lowest R*
will draw the nutrients down to this concentration and exclude all others.
In our model, the smallest pico-phytoplankton have the lowest R* and larger
phytoplankton have subsequent higher R* (Fig. 4b). In this formulation, the
smallest phytoplankton should outcompete all others. However, when we take
grazing by a zooplankton (Z) into account, R∗=kRgZμmax-gZ, where g is a per biomass grazing rate. Thus, R* increases with
increased grazing. When the grazing pressure is sufficiently strong on the
smallest type, the R* of the next smallest phytoplankton is reached and the
two phytoplankton can coexist. The smallest size class phytoplankton and
its grazer have their biomass capped, and any increase in biomass is now due
to the next size class (Armstrong, 1994). This process continues to more and
more size classes as we go from regions of low to high nutrient supply rates
(Fig. 9).
Model rate of supply of nutrients into the top 50 m. Supply
rate of (a) dissolved inorganic nitrogen (mol N m-2 yr-1), (b) phosphate (mol P m-2 yr-1), (c) iron (mol Fe m-2 yr-1) and (d) silicic acid
(mol Si m-2 yr-1). All transport, diffusion and remineralization terms are
included, and for iron dust supply is also included. In (a)–(c), contours are the size class
richness from the total phytoplankton community (Fig. 4b), and in (d) the contour
is for size classes within the diatom functional group alone. Since there are
multiple limiting nutrients (especially for the non-diatoms), patterns of
size diversity shown in (a)–(c) do not exactly match any single nutrient
supply rate. However, the link between size classes of diatoms and silicic
acid supply is clear in (d).
We note that the model is significantly more complex than the simple
theoretical framework, including multiple limiting nutrients, multiple
variants of one of those resources (NH4, NO2 and NO3) with
differing affinities, additional loss terms (e.g. sinking), and more
complicated grazing and food web (rather than food chain). However, this
framework still helps us understand the patterns of size diversity in the
model.
In the model, some regions have different limiting nutrients (e.g. iron
versus dissolved inorganic nitrogen), so the patterns of size diversity from
the total community are more complicated than considering only one nutrient
supply rate. However, this process is nicely shown by the number of size classes
within the diatom group alone increasing cleanly with the supply of silicic
acid (Fig. 9d). The fact that each size class is capped by grazing leads the
distributions of size classes to be relatively even, especially in the
highest nutrient regimes (shown by the Shannon index, Text S3,
Fig. S13).
To explore the importance of size-specific top-down control on diversity
suggested by this theoretical construct, we conduct a sensitivity experiment
(EXP-1, Table 1), where we allow only one grazer to prey on all
phytoplankton. We also do not allow for mixotrophy. We find that only the
smallest size class in each functional group survives (Figs. 10b,
S14): the 0.6 µm pico-phytoplankton and the 3 µm
diazotrophs, coccolithophores and diatoms. The dinoflagellates do not
survive without mixotrophy. The size diversity reduces to one in most
regions (Fig. 11). This experiment highlights that size diversity (Fig. 8b) is
controlled not only by the rate of supply of the limiting nutrients, but
also by size-specific grazing (Armstrong, 1994; Poulin and Franks, 2010; Ward
et al., 2012).
Sensitivity experiments: phytoplankton global biomass.
Global integrated biomass (TgC) for (a) default experiment (identical to Fig. 6a); (b) EXP-1 (experiment with single generalist grazer); (c) EXP-2
(experiment where all phytoplankton have same nutrient requirements); (d) EXP-3 (experiment where plankton are not transported). Types are arranged by
functional group as indicated by the colour and labels at the top of the
graph, by size classes (equivalent spherical diameter, ESD) as labelled
below the graph, and by thermal norms from cold adapted to warm adapted from
left to right in between vertical dotted lines.
Sensitivity simulations: model annual mean richness.
EXP-1 has no size-dependent loss rates (i.e. only one grazer); EXP-2 has no
nutrient requirement differences between functional groups; EXP-3 has no
transport of the plankton (all nutrients and non-living organic pools are
transported). Top row: total richness; second row: size class richness
determined by number of coexisting size classes; third row: functional
richness determined by number of coexisting biogeochemical functional
groups; bottom row: thermal richness determined by number of coexisting
temperature norms. The leftmost column is the same output as shown in Fig. 8a, b, c, d for the original (“default”) experiment, but with absolute values
(i.e. not normalized).
The thermal norm richness of EXP-1 is very similar to the original
default experiment (Fig. 8d), and thus richness of this dimension is not
(at least greatly) controlled by size-specific grazing. Functional group
richness decreases as the dinoflagellates are no longer viable without
mixotrophy. All other functional groups survive (Figs. 10b, S14), and there is coexistence at the functional level; however, the patterns
are different to the default experiment. In EXP-1 there are significant
changes to the biogeochemistry, including the primary production (lower) and
subsequent changes to nutrient supplies. It is these biogeochemical changes
that alter the functional richness patterns (discussed more below). However,
the total diversity reduces dramatically (Fig. 11, top row). Patterns of hot
spots are, however, still apparent, but the increases in diversity with higher nutrient supply are no longer apparent.
We have used steady-state theory to explain the coexistence of size
classes. We contend that when looking at annual average richness this theory
provides insight even in non-steady-state regions such as the highly
seasonal latitudes. However, we do acknowledge that the processes are more
complex in these regions. For instance, during times of resource-saturated
conditions (e.g. beginning of the spring blooms), the smallest diatoms,
which are the fastest-growing phytoplankton, will dominate (Dutkiewicz et
al., 2009; see Appendix A). However, as the grazer of the smallest diatom
increases, the phytoplankton net growth rate (growth minuses losses)
decreases until the next fastest-growing phytoplankton (whose net growth
rate is higher since it is not yet under grazer control) is able to grow
in (Fig. 12). Such a progression of size classes of diatoms has been observed
using Continuous Plankton Recorder (CPR) data (Barton et al., 2013) and
modelled for a coastal system (Terseleer et al., 2014). This process of
succession continues until nutrients are drawn down, allowing the
pico-phytoplankton and mixotrophs to dominate in this more-steady-state
lower-nutrient environment (as suggested by Margalef's mandala, Margalef,
1978). Given that annually there is an optimum condition for each of those
size classes, they do all coexist though at seasonally varying abundances
(i.e. they never go extinct locally).
Default model time series in the North Atlantic
(20∘ W, 45∘ N). Carbon biomass
(mg m-3) of (a) pico-phytoplankton functional group binned by size
class; (b) coccolithophores binned by size class; (c) diatoms binned by size
class; (d) mixotrophic dinoflagellates binned by size class; (e) zooplankton
binned by size class. Diazotrophs do not survive at this location. The line width goes from thinnest for the smallest size class of each phytoplankton functional group and gets thicker for each larger size class (i.e. thinnest line is for 0.6 µm picophytoplankton, 3 µm
for diatoms, etc.). For the zooplankton, however, the thickness of the line is linked to the preferential diatom prey size (i.e. 30 µm ESD zooplankton
for the thinnest line), to show the zooplankton–diatom interactions.
Functional group diversity
The size class and functional group classifications are not completely
orthogonal as the “pico-phytoplankton” group is entirely composed of the four
smallest size classes. We therefore use a similar explanation as to why
pico-phytoplankton can coexist with the other functional types in low
seasonality regions: the low R* of pico-phytoplankton allows them to survive
ubiquitously and other functional groups can only coexist where (or when)
grazing pressures on the pico-phytoplankton and resource supplies are high
enough.
We find that for the rest of the functional groups, coexistence is strongly
controlled by the differences in their resource requirements and the
imbalances in the supply rates of multiple resources (resource supply ratio
theory, Tilman, 1982; see Appendix A). For instance, slow-growing diazotrophs
can only coexist with faster-growing phytoplankton groups when there is an
excess supply of iron and phosphorus delivered relative to the
non-diazotroph N : P and N : Fe demands (Fig. 13a, b, c; Dutkiewicz et al., 2012,
2014; Ward et al., 2013; Follows et al., 2018). In such locations, the
non-diazotrophs are nitrogen limited, while the diazotrophs can fix their
own nitrogen, and the excess P and Fe not utilized by the non-diazotrophs is
available (Appendix A; Fig. 13b, c).
Coexistence of functional types defined by imbalance of
different nutrient supply rates. Left column depicts controls on diazotroph
distribution: (a) fraction of total biomass made up of diazotrophs; (b) ratio of iron to dissolved inorganic nitrogen (DIN) supply rates (see Fig. 9); (c) ratio of phosphate to DIN
supply rate. Colour scale is chosen such that purple indicates supply rate
ratios in excess of the non-diazotroph Fe : N and P : N requirements. Right column depicts the controls on the coexistence of diatoms and non-diatoms: (d) fraction of biomass
made up of non-diatoms; (e) ratio of iron to silicic acid supply rates; (f) ratio of phosphate to silicic acid supply rate; (g) ratio of DIN to silicic
acid supply rates. Colour scale is chosen such that purple indicates supply
rate ratios in excess of the diatom Fe : Si, P : Si and N : Si requirements.
Similar arguments explain where non-diatoms can coexist with the
fast-growing diatoms (Fig. 13d). In regions where there is excess supply of
dissolved inorganic nitrogen, phosphate, and iron relative to the diatom
Si : N, Si : Fe, and Si : P demands, there can be coexistence (Fig. 13e, f, g). In these
locations (or occasions), diatoms are limited by silicic acid, and any
excess N, P and Fe can be used by the other phytoplankton. When the excess
supply is significantly high, non-diatoms can dominate. The high silicic
acid supply in the Southern Ocean leads to lower diversity as the diatoms
win out in all but the low-nutrient summer months, when (in this simulation)
pico-phytoplankton are the only other functional group to survive. In other
seasonal regions, such as the northern North Atlantic (Fig. 12), diatoms
dominate at the beginning of spring, but coccolithophores can outcompete
later in the summer when the diatoms become limited by availability of
silicic acid.
The mixotrophs have two sources of resources: inorganic nutrients and other
plankton. They are parameterized to photosynthesize slower than other
phytoplankton (of the same size, as suggested by observations, Tang, 1994;
Fig. 4a) and graze slower than other grazers (of the same size, Jeong et al.,
2010; Fig. S2). They are advantaged over specialist autotrophs
and heterotrophs when competition for both inorganic nutrients and prey is
strong, and, by using both, their R* for each resource is lowered.
To demonstrate that differential nutrient requirements lead to much of the
functional group coexistence, we conduct another sensitivity experiment
(EXP-2, Table 1) where we force all functional groups to have the same
resource requirements (e.g. diatoms do not require silicic acid, diazotrophs
cannot fix nitrogen, dinoflagellates cannot graze on other phytoplankton)
and C : N : P : Fe ratios are the same for all types. All other growth and grazing
parameterizations remain the same as in the default experiment. In EXP-2,
the functional richness reduces dramatically (Fig. 11), and only
pico-phytoplankton and diatoms survive (Figs. 10c, S15). The
diatoms are the ultimate opportunists (r-strategists) in this model, with
the highest growth rate (Fig. 4a), and survive when nutrient supplies are
high enough. Without any differentiating nutrient requirements relative to
the other functional groups, they outcompete them. Pico-phytoplankton (the
gleaners, k-strategists) survive in regions of lowest nutrient supply, where
their low R* and low grazing allows them to exclude the diatoms. Size class
and thermal norm diversity change very little (Fig. 11). Total diversity is
reduced everywhere but mostly in the lower latitudes where the loss of
diazotrophs and coccolithophores has a high impact.
Thermal norm diversity
We find that thermal norm richness is highest in the regions of the western
boundary currents and other regions generally anticipated to have high
levels of mixing of different water masses. Clayton et al. (2013) identified
a link between hot spots of diversity and eddy kinetic energy and the
variance in sea surface temperature. Anticipating the role of currents and
mixing of water mass (Clayton et al., 2013; Lévy et al., 2014), in a third
sensitivity experiment (EXP-3, Table 1) we do not allow transport of
plankton between grid cells, though we do allow diffusion vertically in the
water column. Thus, this simulation is a collection of one-dimensional
models with regard to the plankton. However, nutrients and dissolved and
detrital organic matter are allowed to be transported as in the default
experiment. Thermal norm diversity decreases (Fig. 11), and there are no
longer hot spots. These results echo findings from Lévy et al. (2014) and clearly show the importance of mixing of water masses for maintaining
thermal norm diversity. When temperature is fluctuating, all phytoplankton
with different temperature norms can survive together provided their
respective temperature optimal occurs for long enough (Kremer and Klausmeier,
2017) or there is a constant supply of the types from upstream (Clayton et
al., 2013). This is different from resources or grazing control where
competition for limited resources is the main process controlling
coexistence (or lack thereof), and as such we find the greatest effect in EXP-3
on thermal norm diversity. Total diversity is reduced everywhere but most
dramatically in these hot-spot regions. Both Clayton et al. (2013) and Lévy et
al. (2014) showed the importance of eddies in enhancing this process of
transport-mediated diversity. Thus the hot spots in the default experiment
would likely be even higher in a model that did resolve the mesoscale.
We find in EXP-3 that the geographical size of almost all habitats (Figs. 14, S16) is reduced. In the case of thermal norms, the lack of
transport allows for very little coexistence. For functional diversity, the
pattern changes, but the maximum richness remains the same. This suggests
that the boundaries of functional group domains are expanded by transport
(see for instance the decrease expanse of diatoms in the gyres,
Figs. S16 versus S10), but transport per se is not the ultimate controller.
Domains for each size class also decrease (Figs. 14, S16),
but most dramatically for the larger size classes, and the two largest go
extinct in this experiment. This suggests that transport also plays a role
in maintaining the grazer/phytoplankton links and that for classes with
smaller domains and/or very low biomass this becomes more crucial. A few
types have an increase in range or in fact exist in EXP-3 and not in the
original experiment (Figs. 10d, 14, S16). These are almost
all the warmest-adapted types that in EXP-3 have very small biomass and
ranges. Thus, transport can also reduce domains of types with very small
potential niches as the constant influx of less-fit types from cooler
regions is sufficient to overcome any competitive advantage of the locally
superior warm-adapted types (see Appendix A).
EXP-3 difference in phytoplankton range geographic
extent. Change in areal extent of all phytoplankton types (1012 km2) between EXP-0 and EXP-3 (no horizontal transport of plankton).
Negative (red) indicates a decrease in the geographic domain of the
phytoplankton type. Types are arranged by functional group as indicated by
the coloured labels at the top of the graph, by size classes (equivalent
spherical diameter, ESD) as labelled below the graph and by thermal norms from
cold adapted to warm adapted from left to right in between each vertical
dotted line. Differences are relative to those shown in Fig. 6b.
Links to diversity along the Atlantic Meridional Transect
Using the results of this study, we can hypothesize as to why richness of
coexisting nano- and micro-eukaryotes along the AMT (Figs. 1a, 15a) has the
observed patterns. We consider the modelled diversity within the three
dimensions along the transect (Fig. 15b, c, d). All three dimensions have high
diversity along the northern edge of the Southern Ocean (labelled A in Fig. 15),
suggesting that all controls (supply rate of limiting nutrient, imbalance in
the supply of different nutrients, top-down control and transport) are at play
in setting the maximum richness seen here in both the model and observations
(Fig. 1a, d). Thermal and functional richness decrease southward, leading to
the drop in total richness observed poleward. Absolute nutrient supplies are
still high enough to maintain size diversity, but the N : Si supply ratios are
no longer conducive to maintaining coccolithophores (Fig. 13e, f, g), and their diversity decreases as is observed (Fig. 1b, e). In this southernmost region
there is also no longer the mixing of different water masses between
the subtropical and Southern Ocean to promote large thermal norm diversity.
However, diatom diversity (due here to size classes) increases (Fig. 1b, e),
driven by the large gradient in silicic acid supply rate (Fig. 9d).
Modelled nano- and micro-eukaryote normalized richness
along the Atlantic transect. Annual mean richness normalized to the maximum in
a transect similar to AMT for (a) all diatoms, coccolithophores and
dinoflagellates – this panel is the same as Fig. 1b; (b) size classes; (c) biogeochemical functional groups; and (d) thermal norms. The normalization factor is
given on the bottom right of each panel. Note that pico-phytoplankton and
diazotrophs are not included in this analysis as they were not part of the
observations. Dashed lines and text (A, B, C, D) are used to locate regions
discussed in the text.
All three dimensions have an even sharper decrease equatorward of the
Southern Ocean boundary, leading to much lower total diversity observed in
the South Atlantic subtropical gyre (labelled B in Fig. 15). Here the lower
absolute nutrient supply likely leads to a reduction in size classes, silicic
acid supply rates drop dramatically (Fig. 9d) and functional diversity
decreases. The lack of mixing of water masses reduces the thermal norm
diversity. Nearer the Equator (labelled C), both size and functional
diversity are high, leading to the observed increase in total diversity.
Here an increased supply of nutrients (Fig. 9) from equatorial upwelling,
including slightly higher Si supply rates, is probably important for
allowing additional size classes and diatoms to exist. In the region of the
Canary upwelling region (labelled D), there is an increase in diversity in
the model and observations. Here increased size class and thermal norm
diversity are possibly responsible, a result of the nutrient-rich upwelled
water mixing with surrounding water masses as it is transported offshore
(see Clayton et al., 2013). The model underestimates this increase since the
model's coarse resolution does not capture the mesoscale filaments
associated with these upwelling features found in the real ocean.
Limitations of this study
This study must be understood within the context of the limitations of the
model. Models are by definition simplified constructs that attempt to
capture the essence of a real system. The model here has a more complex
ecosystem than many other marine models but is still limited in terms of the
parameterization choices. For instance, the size-dependent grazing assumes a
10-to-1 preference as suggested by observations and used in many other
studies (Fenchel, 1987; Kiorboe, 2008; Ward et al., 2012; Baird et al., 2004).
However, there are many examples of grazing that break these preference
rules (Jeong et al., 2010; Weisse et al., 2016; Sommer et al., 2018). The model
assumes fixed elemental ratios in the plankton. This too is an
oversimplification, and variable ability to store nutrients and modify
cellular quotas is an important trait that likely allows for levels of
coexistence (Edwards et al., 2011). This level of stoichiometric complexity
is not incorporated here. However, the model carries almost 750 unique
tracers to account for all the phytoplankton, variable Chl a, and the inorganic and organic pools. To include variable stoichiometry would add
over 2000 more tracers, which is computationally unfeasible for this study.
Each functional group has a different absorption spectrum, though these are
modified with size (see Text S1.3 and Fig. S1); we
recognize that this has a large implication for the pico-phytoplankton whose
accessory pigments are quite different. Using a version of this model, but
with differing absorption spectra for the pico-phytoplankton, Hickman et al. (2010) showed that such a difference was responsible for some niche
separation, especially vertically. The results of this study should be
interpreted in light of these and other simplifications.
The model considers only three axes of phytoplankton traits. We anticipate
that additional axes such as morphology (e.g. shape, spines), motility (e.g.
flagella), chains and colony formation, nutrient storage abilities, and
symbiosis will each have their own controlling mechanisms. Such traits might
allow the model to capture more species and, particularly, more large-sized
phytoplankton types. Previous studies have suggested other controllers of
phytoplankton distributions when considering other traits, for instance the
importance of trade-offs between nutrient acquisition and storage (e.g.
Edwards et al., 2011) or the effect of symbioses (e.g. Follett et al., 2018;
Tréguer et al., 2017). Here, we have specifically designed the model to only
consider the three dimensions for simplicity. Including additional trait
dimensions will likely lead to alterations to the patterns of diversity and
will be important for follow-on studies, especially as our knowledge of the
trade-offs of each trait dimension becomes clearer. For instance, the fact
that the model underestimates diversity in the subtropical gyres suggests
that additional dimensions are likely important in these regions.
Our results are also dependent on the resolution of different axes of trait
space. Likely in the real ocean there is a similar (though more complex)
coarse resolution of functional groups but much higher (potentially
continuous) resolution of size classes and thermal norms. Total diversity
may therefore be influenced more by these two axes than established in this
study. Our model only captures a tiny (probably orders of magnitude less)
amount of the diversity found in the real ocean. Including more resolution
along these axes and including additional trait axes would allow for further
diversity but is beyond the scope of our study. This study should be viewed
as only a step in the understanding of controls of diversity and provides new
evidence to explain the “paradox of the plankton” (Hutchinson, 1961).
However, that we can capture the major patterns of the AMT (Figs. 1,
S7) suggests that we have included some of the most
important mechanisms.
Given computational constraints with this complexity of ecosystem model, we
have used a coarse-resolution physical model that does not capture explicit
meso- or subscale features. Previous studies (e.g. Clayton et al., 2013;
Lévy et al., 2014) have shown the importance of such features in modulating
diversity. Meso- and sub-mesoscale features give rise to temporal increases
in nutrient supplies (see e.g. Clayton et al., 2017), and, according to our
results, this suggests temporal increases in size classes during such
events. Sub- and mesoscale mixing in frontal regions will also enhance
species richness in hot spots (Clayton et al., 2013) but also result in a general increase in species richness (Lévy et al., 2014).
Discussion
We have used ecological theories and a numerical model to examine the
controls on phytoplankton diversity along a number of trait dimensions. We
find that each dimension has a different set of controls. Observed total
diversity is an integrated function of the richness along each trait
dimension and is thus controlled by many different mechanisms. By focusing
on the mechanisms, we can understand the patterns of diversity at the
fundamental level. Such insight provides us with a perspective to predict
changes that might occur in diversity in, for instance, a warming world.
Our results suggest that observed patterns of total diversity (or for
any grouping of phytoplankton types, such as for nano- and micro-eukaryotes
along the AMT) are a result of multiple controllers: supply rate of limiting
resource; imbalance in the supply of different resources relative to
competitors' demands; top-down control, particularly in terms of
size-dependent grazing; and transport processes. The importance of both
resource supply and resource imbalance (or resource supply ratio) has
previously been demonstrated by Cardinale et al. (2009) for lake habitats and
more recently for other natural phytoplankton assemblages (Lewandowska et
al., 2016).
In this study we have synthesized the previously known theory and a numerical
model. The results explain why previous model results have had sometimes
contradictory results. In ecosystem models that only considered two
dimensions of diversity (functional groups and thermal norms, Barton et al.,
2010; Clayton et al., 2013), different patterns where obtained relative to a
model that only considered size (Ward et al., 2014). For instance, the
hot spots of diversity in western boundary currents were not apparent in the
study of Ward et al. (2014) since thermal norm diversity was not included in
that study. Similarly, the lack of high diversity along the edge of the
Southern Ocean in Barton et al. (2010), seen in this study and in the AMT
observations (Fig. 1), was due to the lack of a size trait dimension in that
study. This stresses that diversity in models needs to be understood in
terms of the traits that are included. This obviously bring up the questions
raised in Sect. 6: what additional patterns will be apparent in models
that include additional, or other, trait dimensions? This is an exciting avenue for
future study.
The drivers we found in this study (supply rate of limiting resource,
imbalance in the supply of different resources relative to the better
competitor's demands, size-dependent grazing and transport processes) have
little to do with factors such as temperature or latitude that have been
investigated by correlations to diversity patterns (see e.g. Hillebrand and
Azovsky, 2001; Hillebrand, 2004; Irigoien et al., 2004; Smith et al., 2007;
Rodriguez-Ramos et al., 2015; Powell and Glazier, 2017). However, there may be
some occasions when there are correlations between such factors as
temperature and, for instance, nutrient supply rates, thereby somewhat
confounding correlation and causation. Though observational studies have
hypothesized a multi-factorial control on diversity in the ocean (e.g.
Rodriguez-Ramos et al., 2015), they were unable to find significant
correlations with any combination of factors such as latitude, temperature
or biomass, or even nutrient concentrations. Correlating with factors such
as temperature, latitude is a logical first step for trying to understand
observed patterns of diversity, as these are often the only additional data
that are available from a field study, and for instance “latitude” could
potentially stand in for a range of biotic and abiotic processes. Our study,
however, suggests that, to some degree, these factors are unlikely to help
disentangle controllers of diversity. For instance, in our study, it is the
mixing of different temperature water masses, potentially hinted at by local
temperature variances, rather than temperature itself, that is important at
least for one dimension of diversity. A previous study focusing on the
interactions between microbes (and hence community structure) showed few
statistical links to nutrient concentrations (Lima-Mendez et al., 2015).
However, nutrient supply rates (a harder variable to measure) have been shown
to strongly influence the taxonomic and size structure of marine
phytoplankton communities (see e.g. Mouriño-Carballido et al., 2016).
Diversity controls inferred by correlations with environmental factors or
from niche modelling (e.g. Righetti et al., 2019, who make use of statistical
inferences on species biogeography) likely miss important drivers. For
instance, biotic interactions (competition and grazing) and impacts of
transport (two mechanisms we have shown to be important) cannot easily be
captured using such statistical techniques.
Biomass and productivity are dictated by the supply rate of the limiting
nutrient, and therefore our study found an increase in size diversity with
increased productivity and biomass. This compares well to the observations
of Marañón et al. (2015) and Acevedo-Trejos et al. (2018), who found an increase in size classes with higher productivity. However, we caution that
it is nutrient supply rate (not productivity) that is the controlling
mechanism. Obviously, nutrient supply rate (a bottom up process) cannot
alone lead to high size diversity. Top-down processes are essential for the
build-up of size classes with higher nutrient supply (see also Poulin and
Franks, 2010). Considering only correlations with productivity would lead
one to miss this important biotic interaction as a control on diversity. In
our model, the top-down control was size-specific grazing, but similar patterns
could be achieved with kill-the-winner type parameterizations (Vallina et al.,
2014a) or by imposing species-specific grazers or viruses.
Though transport of phytoplankton most strongly controls the thermal norm
diversity, we did find that it modulates the extent of the regions for all
traits. For instance, diatoms die out in the central subtropical gyres when
transport is turned off in EXP-3, and the largest size classes become less
competitive without transport (Figs. 10d, S16). Our
explanations of the different controls on the diversity along different
trait axes should be understood as focusing on the most important
components. The real system has multiple controlling mechanisms working
together.
The discussion of marine phytoplankton diversity must also be considered in
light of the limited but also different types of observational datasets
(see review by Johnson and Martiny, 2015). Different techniques tend to
capture just some aspects of diversity; for instance, different axes are
distinguished when instruments measure just size (e.g. by flow cytometer,
laser diffractometer), pigments (e.g. through high-performance liquid chromatography), or morphologic and taxon differences
(e.g. microscopy). Only recently have studies incorporated diversity from a
genomic perspective (e.g. de Vargas et al., 2015). Genomic diversity tends to
capture a much higher diversity than other methods, with a long tail of rare
species not captured by other measurements (Busseni, 2018). Thus,
“diversity” depends on the definition and/or on the measurement used.
Observational datasets are, however, sparse and only capture a small
temporal and spatial pattern of biodiversity. Only recently has it become commonly understood that diversity studies require consistent datasets
(e.g. Rodriguez-Ramos et al., 2015; Sal et al., 2013) and that
sampling biases can skew results (Cermeño et al., 2013).
Conclusions
In this study we have disentangled some of the multiple controls on marine
phytoplankton species richness (or types), a metric of diversity. We have
shown through theory and a model that diversity within different dimensions
of phytoplankton traits is controlled by disparate drivers. The number of
coexisting size classes of phytoplankton is largely controlled by the
magnitude of the limiting resource supply rate and the strength of the
size-specific top-down processes; functional group coexistence is partly
controlled by the imbalance in the supply rate of different resources
relative to competing species' demands; the number of phytoplankton types
with different thermal optima that can coexist is strongly controlled by
the amount of mixing of different water masses. Transport in general expands
the range of phytoplankton habitats and leads to higher local diversity.
That each controller affects a different dimension of diversity is important
to explain why diversity patterns in models that include only one or two of
the traits will have different results to one that includes all three.
Likely including other traits (e.g. morphology, symbioses) controlled by
different (as yet not understood) mechanisms will lead to additional
components to the patterns of diversity.
This study offers an explanation to why there have often been conflicting
results in observational studies that have attempted to link diversity to
factors such as temperature or productivity. Even when they do show
correlations with diversity, it can sometimes be only because such factors
are also correlated with some of the actual drivers (such as nutrient supply
rates), and results will also be specific to the dimensions of diversity
measured. Models such as this one, though still only capturing a tiny amount
of the full diversity of the ocean, can be a good tool to address both
consistency and sampling biases, as well as for providing insight into
controlling mechanisms as we have done here. By understanding the
mechanistic controls on diversity, we are in a better position to understand
how diversity might have been different in the past, how it changes with
interannual variability and how it might alter in a future ocean.
In this appendix we detail the theoretical framework we mention in the main
text. This framework starts from the simplest system of equations to define
the marine ecosystem. We note that the computer model discussed in the main
text has much more complex equations, including temperature and light
constraints on growth, multiple limiting nutrients, multiple variants of one
of those resources (NH4, NO2 and NO3) with differing
affinities, additional loss terms (e.g. sinking), and more complicated
parameterization of grazing and a food web including carnivorous
zooplankton.
Theory. We consider a system of phytoplankton biomass (B) sustained
by nutrients (R):
A1dRdt=-μmaxRR+kRB+SR,A2dBdt=μmaxRR+kRB-MB,
where μmax is maximum growth rate, kR is the half-saturation constant
for growth, SR is the supply of resource R and M is the phytoplankton loss
term (we will consider different assumptions of M below).
Steady state
Here we synthesize the theoretical
underpinning that we have previously presented (Dutkiewicz et al., 2009;
Ward et al., 2013, 2014; Lévy et al., 2014; Follows et al., 2018).
Those studies have in turn been informed from the seminal work of Tilman (1982) and Armstrong (1994).
We assume a steady state and solve the biomass equation (Eq. A2):
R∗=kRMμmax-M.
This is the concentration that the phytoplankton will draw the resource down
to in a steady state. In a system with J phytoplankton, the one with the lowest
Rj* will draw the nutrients down to this concentration and all others
will be excluded.
Grazing allows coexistence
If we now consider a
system of J phytoplankton (Bj) and K zooplankton (Zk), where each
phytoplankton has a specific grazer, we can write the loss rate now as
M=m+gkjZk. Here gkj is the per biomass grazing rate of
zooplankton k on phytoplankton j, and m is a linear loss rate (resolving cell
death and other losses). In this case
Rj∗=kRj(m+gkjZk)μmaxj-(m+gkjZk).
Note that this is not an explicit solution as Zk is itself a complex
function of the parameters. However, this equation can provide us with the
insight that with higher grazing Rj* increases.
For a situation where Rj* increases with size (in the absence of grazing),
the smallest phytoplankton (j=1) will outcompete others in the absence of
grazing. However as grazing pressure increases, Rj=1j=2* of this
smallest type will increase. When this Rj=1* becomes large enough it
reaches the Rj=2* of the second smallest phytoplankton (assume for now
that this plankton is not grazed) and the two phytoplankton will be able to
coexist. This situation occurs when there is higher resource supply
(SR) allowing for a larger biomass of both phytoplankton and
zooplankton. With even higher nutrient supply, similar grazing control of
the j=2 phytoplankton (Rj=2*) will allow a third phytoplankton–zooplankton
pair to coexist with the others in the system. This system however does
require a separate grazer per phytoplankton, or a strong kill-the-winner
parameterization. This theory explains the coexistence of several size
classes in the ecosystem model (Figs. 8b, S10). For more
details, see Ward et al. (2014) and Follows et al. (2018).
Multiple limiting resources allow coexistence
If
we now consider a system of two phytoplankton (Bj, where j is 1 or 2)
limited by different resources (Ri where i is A or C), we suggest that this system
can allow for coexistence. To explore when the two types can coexist we
expand Eqs. (A1) and (A2) (where the biomass is in units of element A) such that
A5dRAdt=-μmax1RARA+kRA1B1-μmax2RCRC+kRC2B2+SRA,A6dRCdt=-μmax1RARA+kRA1ΥAC1B1-μmax2RCRC+kRC2ΥAC2B2+SRC,A7dB1dt=μmax1RARA+kRA1B1-M1B1,A8dB2dt=μmax2RCRC+kRC2B2-M2B2,
where ΥAC1 represents the stoichiometric ratio requirements of B1
for element A and C. SRA andSRC are the supply rate of nutrient A
and C respectively. If one of the phytoplankton (B1) has a much higher
growth rate than the other (B2) it will be a better competitor for both
resources (A and C). We find, solving the above equations in a steady state, that
coexistence is possible if
SRASRC>ΥAC1.
There must be excess supply of the resource limiting the slower-growing
phytoplankton relative the stoichiometric demands of the faster-growing
phytoplankton for coexistence to occur.
For the case of an Fe-limited diazotroph (which can fix its own nitrogen)
and a faster-growing DIN-limited non-diazotroph, coexistence occurs when
SFeSN>ΥNFe1, where ΥNFe1 represents the
stoichiometric demands of the non-diazotroph. We can write a similar
inequality for any other nutrient limiting the diazotrophs (e.g. P) and
find that diazotrophs survive where both SFe and SP are supplied in
excess of the non-diazotroph requirements (Fig. 13b, c). See Ward et al. (2013)
and Follows et al. (2018) for more details.
Similarly, the equations in a steady state suggest that for slower-growing
non-diatoms to coexist with the fast-growing diatoms, the diatoms must be
silicic acid limited. In a situation where the non-diatoms are DIN limited,
then coexistence occurs if SNSSi>ΥSiN1, where
ΥSiN1 represents the stoichiometric demands of the diatom. Again,
similar inequalities are applicable if other nutrients limit the
non-diatoms (e.g. P, Fe), and we find that non-diatoms can exist where DIN,
Fe and P are supplied in excess of the diatom requirements (Fig. 13e, f, g).
Physical transport can allow coexistence
As
discussed in Lévy et al. (2014), physical transport can also modify R*. Here were
recognize that Eq. (A2) should be expanded for a moving ocean to
dBdt=μmaxRR+kRB-MB+TB+VB,
where T represents the per unit biomass advection of plankton,
T=-1B∇uB, where u is the local three-dimensional velocity vector; V represents per unit biomass vertical
mixing, V=1B∂∂z(K∂B∂z), where K is the vertical
mixing coefficient; and z indicates the vertical dimension. With these
additions,
R∗=kR(M-T-V)μmax-(M-T-V).
Thus T and V provide additional means for phytoplankton to have a similar R*. If a
phytoplankton type is less competitive at a location, it can still have a
similar R* to a locally better adapted type if there is a steady influx of it
from an upstream location. We clearly see this effect in the (generally)
expanded biogeography of phytoplankton with advection relative to the
experiment without advection (Figs. 14, 16).
Non-steady state
In a previous study (Dutkiewicz et al., 2009)
we found that this steady-state theory was applicable in a model in the
subtropics and in the summer months in some of the high-latitude regions. We
contend that when looking at annual coexistence this theoretical
understanding still provides insight even in non-steady-state regions such
as the highly seasonal high latitudes (as was done in Ward et al., 2014).
However, we do acknowledge that the processes are more complex in these
regions. Such regions generally have a succession of dominance of different
types. As long as there is a long enough period of favourable conditions for
each type, the phytoplankton can coexist, though with seasonally varying
biomass. We explain the succession by considering Eq. (2) in a non-steady
state:
1BdBdt=μmaxRR+kR-M,
such that the biomass normalized tendency term is dictated by the net growth
rate (μmaxRR+kR-M). At any moment (or with a short lag)
the phytoplankton with the largest net growth rate can dominate
(temporally).
Spring bloom
As suggested in Dutkiewicz et al. (2009), the fastest-growing phytoplankton will dominate at the beginning of
the spring bloom when the nutrients are plentiful RR+kR∼1
and grazing is small, such that Eq. (12) reduces to
1BdBdt=μmax.
That is, the phytoplankton with the largest μmax will dominate. In
the model here, this is the smallest diatoms.
Grazing allows coexistence
If we now consider two
phytoplankton (B1, B2) both limited by the same nutrient, R, and each
having its own specific grazer (Z1, Z2), then M becomes
M=m+gkjZk. If we assume μmax1>μmax2, then B1 will
dominate when there is no grazer control. However, when Z1 is large
enough, and Z2 is small or negligible, it is possible for
μmax1RR+kR1-m-g11Z1<μmax2RR+kR2-m.
In this case B2 can grow in and potentially dominate the system
temporarily. Similarly, as grazing control limits B2, a third species
with slower growth but also lower grazing might be able to follow on the
succession. This is shown in the model for a location in the North Atlantic
with a succession of diatoms of increasing size in the spring bloom period
(see Fig. 12).
Multiple limiting resources allow coexistence
We
can also consider Eqs. (A7) and (A8) (two phytoplankton types limited by
different nutrient) in a non-steady-state case. If B1 is the faster-growing species, it may still be outcompeted (at least temporarily) by the slower-growing species if
μmax1RARA+kRA1<μmax2RCRC+kRC2.
That is, B2 can succeed B1 if the nutrient limitation of B1
becomes severe enough that its net growth drops lower than its competitor,
which is limited (less) by a different nutrient. An example is a strongly
silicic acid limited diatom later in the seasonal progression succeeded by a
nitrate-limited coccolithophore, as in the model example (Fig. 12). Provided
each type has sufficiently long favourable conditions each year, it will
continue to coexist at any location though at lower abundances for part of
the year.
Physical transport allows coexistence
We can use
the biomass normalized tendency formulation to consider the circumstances
where physical transport has an impact (see Eq. A10):
1BdBdt=μmaxRR+kR-M+T+V.
Temporarily a phytoplankton type might have the fastest tendency if T or V are
particularly strong (i.e. there is a strong supply of that type to the
location through advection or mixing). Such circumstances may occur in
highly energetic regions where there is a constant advected supply of
different types (e.g. a fast-moving western boundary current). A highly
varying set of environmental conditions will also help in this situation.
For instance if μmax is assumed to have a temperature-mediated
component (as in the numerical model, Text S1.1, Eq. S1.4),
then many different types would temporarily have the best environment.
However, these beneficial conditions may not occur often enough or long
enough to maintain coexistence without the constant supply of a new
population. This is the situation in the hot spots of diversity seen in the
default experiment but which disappear in the experiment with no advection
(Fig. 11). See more discussion in Clayton et al. (2013). We note that the
hot spots do not appear in either the size class or functional group
richness, suggesting that the temporal best environment can be provided
by varying temperatures, but no such temporary optimal situation occurs in
these circumstances for the other dimensional controls
Code availability
The global physical/circulation model (MITgcm) is
available at: http://mitgcm.org (last access: 4 February 2020). The ecosystem component is available as a git repository from: https://gud.mit.edu/git/gud (last access: 4 February 2020). Version and modifications used for this study are
available at: 10.7910/DVN/EOTT9H (Dutkiewicz and Jahn, 2019).
Data availability
Model output used in this study is available at:
10.7910/DVN/JUQCFG (Dutkiewicz, 2019).
The supplement related to this article is available online at: https://doi.org/10.5194/bg-17-609-2020-supplement.
Author contributions
SD conceived the experimental design, conducted the
biogeochemical/ecosystem/optical model simulations, and performed most of
the analysis. OJ was responsible for the numerical code, with input from
AEH on the phytoplankton absorption (Fig. S1). PC provided
the Atlantic Meridional Transect observational data (Figs. 1, 5, S6, S7). BAW and MJF provided input on the theoretical
interpretations. BAW provided the schematic in Fig. 2. DT provided input on the
grazing parameterizations and the data for Fig. S2. SD led the
writing with input from all authors.
Competing interests
The authors declare that they have no conflict of interest.
Acknowledgements
Stephanie Dutkiewicz, Michael J. Follows and Oliver Jahn received funding from NASA (grant no. NNX16AR47G and 80NSSC17K0561).
This work was also supported by the Simons Collaboration on Computational
Biogeochemical Modeling of Marine Ecosystems (CBIOMES) (grant no. 549931).
Financial support
This research has been supported by NASA (grant nos. NNX16AR47G and 80NSSC17K0561) and the Simons Foundation (grant no. 549931).
Review statement
This paper was edited by Jack Middelburg and reviewed by Adam C. Martiny and one anonymous referee.
ReferencesAcevedo-Trejos, E., Brandt, G., Bruggeman, J., and Merico, A.: Mechanisms
shaping size structure and functional diversity og phytoplankton communities
in the ocean, Sci. Rep.-UK, 5, 8918, 10.1038/srep08918, 2015.Acevedo-Trejos, E., Marañón, E., and Merico, A: Phytoplankton size
diversity and ecosystem function relationships across oceanic regions, P.
R. Soc. B., 285, 20180621, 10.1098/rspb.2018.0621, 2018.Agawin, N. S. R. and Agustí, S.: Abundance, frequency of dividing
cells and growth rates of Synechococcus sp. (cyanobacteria) in the stratified Northwest
Mediterranean Sea, J. Plankton Res., 19, 1599–1615, 1997.
Agawin, N. S. R., Rabouille, S., Veldhuis, M. J. W., Servatius, L., Hol, S., van
Overzee, H. M. J., and Huisman, J.: Competition and facilitation between
unicellular nitrogen-fixing cyanobacteria and non-nitrogen-fixing
phytoplankton species, Limnol. Oceanogr., 52, 2233–2248, 2007.
Allen, A. P., Gillooly, J. F., and Brown, J. H.: Recasting the species–energy
hypothesis: the different roles of kinetic and potential energy in
regulating biodiversity, in: Scaling biodiversity, edited by: Storch, D., Marquet, P. A., and Brown, J. H.,
283–299, Cambridge University Press, Cambridge, 2007.
Armstrong, R. A.: Grazing limitation and nutrient limitation in marine
ecosystems: steady state solutions of an ecosystem model with multiple food
chains, Limnol. Oceanogr., 39, 597–608, 1994.
Baird, M. E., One, R. R., Suthers, I. M., and Middleton, J. H.: A plankton
population model with biomechanics descriptions of biological processes in
an idealized 2D ocean basin, J. Marine Syst., 50, 199–222, 2004.Barton, A. D., Dutkiewicz, S., Flierl, G., Bragg, J., and Follows, M. J.:
Patterns of Diversity in Marine Phytoplankton, Science, 327, 5972,
10.1126/science.1184961, 2010.Barton, A. D., Finkel, Z. V., Ward, B. A., Johns, D. G., and Follows, M. J.: On the
roles of cell size and trophic strategy in North Atlantic diatom and
dinoflagellate communities, Limnol. Oceanogr., 58, 254–266,
10.4319/lo.2013.58.1.0254, 2013.
Barton, A. D., Irwin, A. J., Finkel, Z. V., and Stock, C.: Anthropogenic climate
change drives shift and shuffle in North Atlantic phytoplankton communities,
P. Natl. Acad. Sci. USA, 113, 2964–2969, 2016.Beardall, J., Allen, D., Bragg, J., Finkel, Z. V., Flynn, K. J., Quigg,
A., Rees, T. A. V., Richardson, A., and Raven, J. A.: Allometry and
stoichiometry of unicellular, colonial and multicellular phytoplankton, New
Phytol., Tansley Review, 181, 295–309,
10.1111/j.1469-8137.2008.02660.x, 2008.
Bec, B., Collos, Y., Vaquer, A., Mouillot, D., and Sourchu, P.: Growth rate
peaks at intermediate cell size in marine photosynthetic picoeukaryotes,
Limnol. Oceanogr., 53, 863–867, 2008.Bergman, B., Sandh, G., Lin, S., Larsson, J., and Carpenter,
E. J.: Trichodesmium – a widespread marine cyanobacterium with unusual nitrogen fixation
properties, FEMS Microbiol. Rev., 37, 286–302, 2013.
Bissenger, J. E., Montagnes, D. J. S., Harples, J., and Atkinson, D.:
Predicting marine phytoplankton maximum growth rates from temperature:
Improving on the Eppley curve using quantile regression, Limnol. Oceangr., 53,
487–493, 2008.Boatman, T. G., Lawson, T., and Geider, R. J.: A key marine diazotroph in a
changing ocean: The interacting effects of temperature, CO2 and light on the
growth of Trichodesmium erythraeum IMS101, PLOS ONE, 12, e0168796, 10.1371/journal.pone.0168796,
2017.Boyd, P., Rynearson, T. A., Armstrong, E. A., Fu, F., Hayashi, K.,
Hu, Z., Hutchins, D. A., Kudela, R. M., Litchman, E.,
Mulholland, M. R., Passow, U., Strzepek, R. F., Whittaker, K. A., Yu, E.,
and Thomas M. K.: Marine phytoplankton temperature versus growth response from
polar to tropical waters – outcome of a scientific community-wide study,
PlosOne, 8, e63091, 10.1371/journal.pone.0063091, 2013.Breitbarth, E., Wohler, J., Klas, J., LaRoche, J., and Peekan, I.: Nitrogen
fixation and growth rates of Trichodesmium IMS101 as a function of light intensity, Mar.
Ecol.-Prog. Ser., 359, 25–36, 2008.
Buitenhuis, E. T., Pangerc, T., Franklin, D. J., Le Quere, C., and Malin, G.:
Growth rates of six coccolithophorid strains as a function of temperature,
Limnol. Oceanogr., 53, 1181–1185, 2008.Buitenhuis, E. T., Li, W. K. W., Vaulot, D., Lomas, M. W., Landry, M. R., Partensky, F., Karl, D. M., Ulloa, O., Campbell, L., Jacquet, S., Lantoine, F., Chavez, F., Macias, D., Gosselin, M., and McManus, G. B.: Picophytoplankton biomass distribution in the global ocean, Earth Syst. Sci. Data, 4, 37–46, 10.5194/essd-4-37-2012, 2012.Buitenhuis, E. T., Vogt, M., Moriarty, R., Bednaršek, N., Doney, S. C., Leblanc, K., Le Quéré, C., Luo, Y.-W., O'Brien, C., O'Brien, T., Peloquin, J., Schiebel, R., and Swan, C.: MAREDAT: towards a world atlas of MARine Ecosystem DATa, Earth Syst. Sci. Data, 5, 227–239, 10.5194/essd-5-227-2013, 2013.
Busseni, G.: Exploring diatom functional and taxonomic diversity on a global
scale through an integrative approach, Doctoral Dissertation, The Open
University and Statione Zoologica Anton Dohrn Napoli, Naples, Italy, 2018.Cardinale, B. J., Hillebrand, H., Harpole, W. S., Gross, K., and Ptacnik, R.:
Separating the influence of resource “availability” from resource
“imbalance” on productivity diversity relationships, Ecol. Lett., 12, 475–487, 10.1111/j.1461-0248.2009.01317.x, 2009.Cardinale, B. J., Matulich, K. L., Hooper, D. U., Byrnes, J. E., Duffy, E.,
Gamfeldt, L., Balvanera, P., O'Connor, M. I., and Gonzalez, A.: The
functional role of producer diversity in ecosystems, Am. J. Bot., 98,
572–592, 10.3732/ajb.1000364, 2011.Cermeño, P., Dutkiewicz, S., Falkowski, P. G., Follows, M. J., Harris,
R. P., and Schofield, O.: The role of oceanic nutricline depth in regulating
Earth's carbon cycle, P. Natl. Acad. Sci. USA, 105, 20344–20349,
10.1073/pnas.0811302106, 2008.Cermeño, P., Rodriguez-Ramos, T., Dornelas, M., Figueiras, F.,
Marañón, E., Teixeira, I. G., and Vallina, S. M.: Species richness in
marine phytoplankton communities is not correlated to ecosystem
productivity, Mar. Ecol.-Prog. Ser., 488, 1–9, 10.3354/meps10443, 2013.Cermeño, P., Chouciño, P., Fernández-Castro, B., Figueiras,
F. G., Marañón, E., Marrasé, C., Mouriño-Carballido, B.,
Pérez-Lorenzo, M., Rodríguez-Ramos, T., Teixeira, I. G., and Vallina,
S. M.: Marine Primary Productivity Is Driven by a Selection Effect, Front.
Mar. Sci., 3, 173, 10.3389/fmars.2016.00173, 2016.
Chaudhary, C., Saeedi, H., and Costello, M. J.: Bimodality of latitudinal
gradients in marine species richness, Trends Ecol. Evol., 31,
670–676, 2016.Christaki, U., Jacquet, S., Dolan, J. R., Vaulot, D., and Rassoulzadegan, F.:
Differential grazing and growth on Prochlorococcus and Synechococcus by two contrasting ciliates, Limnol.
Oceanogr., 44, 52–61, 1999.
Chust, G., Irigoien, X., Chave, J., and Harris, R. P.: Latitudinal
phytoplankton distribution and the neutral theory of biodiversity, Global
Ecol. Biogeogr., 22, 531–543, 2013.
Clayton, S., Dutkiewicz, S., Jahn, O., and Follows, M. J.: Dispersal, eddies,
and the diversity of marine phytoplankton, Limnol. Oceanogr.: Fluids and
Environments, 3, 182–197, 2013.Clayton, S., Dutkiewicz, S., Jahn, O., Hill, C., Heimbach, P., and Follows, M. J.: Biogeochemical versus ecological consequences of modeled ocean physics, Biogeosciences, 14, 2877–2889, 10.5194/bg-14-2877-2017, 2017.de Vargas, C., Audic, S., Henry, N., Decelle, J., Mahé, F., Logares, R.,
Lara, E., Berney, C., Le Bescot, N., Probert, I., Carmichael, M., Poulain,
J., Romac, S., Colin, S., Aury, J.-M., Bittner, L., Chaffron, S., Dunthorn,
M., Engelen, S., Flegontova, O., Guidi, L., Horák, A., Jaillon, O.,
Lima-Mendez, G., Lukeš, J., Malviya, S., Morard, R., Mulot, M., Scalco,
E., Siano, R., Vincent, F., Zingone, A., Dimier, C., Picheral, M., Searson,
S., Kandels-Lewis, S., Tara Oceans Coordinators, Acinas, S. G. Bork, P.,
Bowler, C., Gorsky, G., Grimsley, N., Hingamp, P., udicone, D., Not, F.,
Ogata, H., Pesant, S., Raes, J., Sieracki, M. E., Speich, S., Stemmann,L.,
Sunagawa, S., Weissenbach, J., Wincker, P., and Karsenti, E.: Eukaryotic
plankton diversity in the sunlit
ocean, Science, 348, 1261605, 10.1126/Science.1261605, 2015.Dutkiewicz, S.: Diversity Model Output, Harvard Dataverse, V1, 10.7910/DVN/JUQCFG, 2019.Dutkiewicz, S. and Jahn, O.: Diversity Model Code, Harvard Dataverse, V1, 10.7910/DVN/EOTT9H, 2019.Dutkiewicz, S., Follows, M. J., and Bragg, J. G.: Modeling the coupling of
ocean ecology and biogeochemistry, Global Biogeochem. Cy., 23, GB4017,
10.1029/2008GB003405, 2009.Dutkiewicz, S., Ward, B. A., Monteiro, F., and Follows, M. J.: Interconnection
between nitrogen fixers and iron in the Pacific Ocean: Theory and numerical
model, Global Biogeochem. Cy., 26, GB1012, 10.1029/2011GB004039,
2012.Dutkiewicz, S., Ward, B. A., Scott, J. R., and Follows, M. J.: Understanding predicted shifts in diazotroph biogeography using resource competition theory, Biogeosciences, 11, 5445–5461, 10.5194/bg-11-5445-2014, 2014.Dutkiewicz, S., Hickman, A. E., Jahn, O., Gregg, W. W., Mouw, C. B., and Follows, M. J.: Capturing optically important constituents and properties in a marine biogeochemical and ecosystem model, Biogeosciences, 12, 4447–4481, 10.5194/bg-12-4447-2015, 2015a.Dutkiewicz, S., Morris, J., Follows, M. J., Scott, J., Levitan, O.,
Dyhrman, S., and Berman-Frank, I.: Impact of ocean acidification on the
structure of future phytoplankton communities, Nature Climate Change, 5, 1002–1006,
10.1038/nclimate2722, 2015b.
Edwards, K. F., Klausmeier, C. A., and Litchman, E.: Evidence for a three-way
trade-off between nitrogen and phosphorus competitive abilities and cell
size in phytoplankton, Ecology, 92, 2085–2095, 2011.
Edwards, K. F., Thomas, M. K., Klausmeier, C. A., and Litchman, E.: Allometric
scaling and taxonomic variation in nutrient utilization traits and maximum
growth rate of phytoplankton, Limnol. Oceanogr., 57, 554–566, 2012.
Eppley, R. W.: Temperature and phytoplankton growth in the sea, Fish. Bull.,
70, 1063–1085, 1972.Falkowski, P. G., Barber, R. T., and Smetacek, V.: Biogeochemical controls and
feedback on ocean primary production, Science, 281, 200–206,
10.1126/science.281.5374.200, 1998.Fenchel, T. M.: Ecology of Protozoa, Springer, Berlin Heidelberg, p. 187, 1987.Finkel, Z. and Irwin, A. J.: Modeling size-dependent photosynthesis: light
absorption and the allometric rule, J. Theor. Biol., 204,
361–369, 10.1006/jtbi.2000.2020, 2000.
Finkel, Z. V., Beardall, J., Flynn, K. J., Quigg, A., Rees, T. A. V., and
Raven, J. A.: Phytoplankton in a changing world: cell size and elemental
stoichiometry, J. Plankton Res., 32, 119–137, 2010.
Follows, M. J., Dutkiewicz, S., Ward, B. A., and Follett, C. N.: Theoretical
interpretation of subtropical plankton biogeography, in: Microbial Ecology of the Oceans, edited by: Gasol, J. and Kirshman, D., 3rd Edn.,
John Wiley, Hoboken, NJ, p. 467, 2018.Follett, C. L., Dutkiewicz, S., Karl, D. M., Inomura, K., and Follows, M. J.:
Why is there a summertime bloom of diatom-diazotroph associations in the
North Pacific subtropical gyre?, ISME J., 12, 1543–1557, 10.1038/s41396-017-0012-x,
2018.
Fuhrman, J. A.: Microbial community structure and its functional implications,
Nature, 459, 193–199, 2009.Garcia, N. S. and Hutchins, D. A.: Light-limited growth rates modulate nitrate inhibition of dinitrogen fixation in the marine unicellular cyanobacterium Crocosphaera watsonii, PLOS ONE, 9, e114465, 10.1371/journal.pone.0114465, 2014.
Geider, R. J., MacIntyre, H. L., and Kana, T. M.: A dynamic regulatory model
of photoacclimation to light, nutrient and temperature, Limnol. Oceanogr.,
43, 679–694, 1998.
Grover, J. P.: Dynamics of competition among microalgae in variable
environments: experimental test of alternative models, Oikos, 62, 231–243,
1991a.
Grover, J. P.: Resource competition in a variable environment: phytoplankton
growing according to the variable-internal-stores model, Am. Nat., 138,
811–835, 1991b.
Guidi, L., Stemmann, L., Jackson, G. A., Ibanez, F., Claustre, H., Legendre, L.,
Picheral, M., and Gorsky, G.: Effects of phytoplankton community on
production, size and export of large aggregates: a world-ocean analysis,
Limnol. Oceanogr., 54, 1951–1963, 2009.Hansen, B. B., Bjornsen, B. W., and Hansen, P. J.: Zooplankton grazing and
growth: Scaling within the 2–2000 mm body size range, Limnol. Oceanogr.,
42, 687–704, 10.4319/lo.1997.42.4.0687, 1997.Hickman, A. E., Dutkiewicz, S., Williams, R. G., and Follows, M. J.: Modelling
the effects of chromatic adaptation on phytoplankton community structure in
the oligotrophic ocean, Mar. Ecol.-Prog. Ser., 406, 1–17,
10.3354/meps08588, 2010.
Hillebrand, H.: On the generality of the latitudinal diversity gradient,
Am. Nat., 163, 192–211, 2004.
Hillebrand, H. and Azovsky, A. I.: Body size determines the strength of the
latitudinal diversity gradient, Ecography, 24, 251–256, 2001.
Holling, C. S.: Some characteristics of simple types of predation and
parasitism, Can. Entomol., 91, 385–398, 1959.
Huisman, J. and Weissing, F. J.: Light-limited growth and competition for
light in well mixed aquatic environments: An elementary model, Ecology, 75,
507–520, 1994.Hutchins, D. A., Fu, F. X., Zhang, Y., Warner, M. E., Feng, Y., Portune, K.,
Bernhardt, P. W., and Mulholland, M. R.: CO2 control of Trichodesmium N-2 fixation,
photosynthesis, growth rates, and elemental ratios: Implications for past,
present, and future ocean biogeochemistry, Limnol. Oceanogr., 52,
1293–1304, 10.4319/lo.2007.52.4.1293, 2007.
Hutchinson, G. E.: The paradox of the plankton, Am. Nat., 95, 137–145, 1961.
Irigoien, X., Huisman, J., and Harris, R. P.: Global biodiversity patterns of
marine phytoplankton and zooplankton, Nature, 429, 863–867, 2004.Jeong, H. J., Yoo, Y. D., Kim, J. S., Seon, K. A., Kang, N. S., and Kim, T. H.:
Growth, Feeding and Ecological Roles of the Mixotrophic and Heterotrophic
Dinoflagellates in Marine Planktonic Food Webs, Ocean Sci. J., 45, 65–91,
10.1007/s12601-010-0007-2, 2010.Johnson, R., Strutton, P. G., Wright, S. W., McMinn, A., and Meiners, K. M.:
Three improved Satellite Chlorophyll algorithms for the Southern Ocean, J.
Geophys. Res.-Oceans, 118, 3694–3703, 10.1002/jgrc.20270, 2013.Johnson, Z. I. and Martiny, A. C.: Techniques for quantifying phytoplankton
biodiversity, Annu. Rev. Mar. Sci., 7, 299–324, 10.1146/annurev-marine-010814-015902, 2015.Johnson, Z. I., Zinser, E. R., Coe, A., McNulty, N. P., Woodward, E. M. S.,
and Chisholm, S. W.: Niche partitioning among Prochlorococcus ecotypes along ocean-scale
environmental gradients, Science, 311, 1937–1740, 2006.Kiorboe, T.: Turbulence, phytoplankton cell size, and the structure of
pelagic food webs, Adv. Mar. Biol., 29, 1–72,
10.1016/S0065-2881(08)60129-7, 1993.
Kiorboe, T.: A mechanistic approach to plankton ecology, Princeton University Press, Princeton, 224 pp., 2008.
Kooijman, S. A. L. M.: Dynamic Energy and Mass Budget in Biological Systems, Cambridge Univ. Press, Cambridge, UK, 2000.Kranz, S. A., Levitan, O., Richter, K.-U., Prasil, O., Berman-Frank, I., and
Rost, B.: Combined effects of CO2 and light on the N2-fixing cyanobacterium
Trichodesmium IMS101: Physiological responses, Plant Physiol., 154, 334–345, 10.1104/pp.110.159145, 2010.Kremer, C. T. and Klausmeier, C. A.: Species packing in eco-evolutionary
models of seasonally fluctuating environments, Ecol. Lett., 20, 1158–1168,
10.1111/ele.12813, 2017.
Lampert, W.: Zooplankton research: the contribution of limnology to general
ecological paradigms, Aquat. Ecol., 31, 19–27, 1997.Leblanc, K., Arístegui, J., Armand, L., Assmy, P., Beker, B., Bode, A., Breton, E., Cornet, V., Gibson, J., Gosselin, M.-P., Kopczynska, E., Marshall, H., Peloquin, J., Piontkovski, S., Poulton, A. J., Quéguiner, B., Schiebel, R., Shipe, R., Stefels, J., van Leeuwe, M. A., Varela, M., Widdicombe, C., and Yallop, M.: A global diatom database – abundance, biovolume and biomass in the world ocean, Earth Syst. Sci. Data, 4, 149–165, 10.5194/essd-4-149-2012, 2012.Lévy, M., Jahn, O., Dutkiewicz, S., and Follows, M. J.: Phytoplankton
diversity and community structure affected by oceanic dispersal and
mesoscale turbulence, Limnol. Oceanogr.: Fluids and Environments, 4, 67–84, 10.1215/21573689-2768549,
2014.Lévy, M., Jahn, O., Dutkiewicz, S., Follows, M. J., and d'Ovidio, F.: The
dynamical landscape of marine phytoplankton diversity, Journal of Royal
Society Interface, 12, 20150481, 10.1098/rsif.2015.0481, 2015.Lewandowska, A. M. Biermann, A., Borer, E. T., Cebrian-Piqueras, M. A.,
Declerck, S. A. J., De Meester, L., Van Donk, E., Gamfeldt, L., Gruner, D. S.,
Hagenah, N., Harpole, W. S., Kirkman, K. P., Klausmeier, C. A., Kleyer, M.,
Knops, J. M. H., Lemmens, P., Lind, E. M., Litchman, E., Mantilla-Contreras,
J., Martens, K., Meier, S., Minden, V., Moore, J. L., Olde Venterink, H.,
Seabloom, E. W., Sommer, U., Striebel, M., Trenkamp, A., Trinogga, J., Urabe,
J., Vyverman, W., Van de Waal, D. B., Widdicombe, C. E., and Hillebrand, H.:
The influence of balanced and imbalanced resource supply on biodiversity –
functioning relationship across ecosystems, Philos. T. R. Soc. B, 371,
20150283, 10.1098/rstb.2015.0283, 2016.Lima-Mendez, G., Faust, K., Henry, N. C., Decelle, J., Colin, S., Carcillo,
F., Chaffron, S., Ignacio-Espinosa, J. C., Roux, S., Vincent, F. J., Bittner,
L., Darzi, Y., Wang, J., Audic, S., Berline, L., Bontempi, G., Cabello,
A. M., Coppola, L., Cornejo-Castillo, F. M., d'Ovidio, F. D., Meester, L. D.,
Ferrera, I., Garet-Delmas, M., Guidi, L., Lara, E., Pesant, S., Royo-Llonch,
M., Salazar, G., Sánchez, P., Sebastián, M. G., Souffreau, C.,
Dimier, C., Picheral, M., Searson, S., Kandels-Lewis, S., Gorsky, G., Not,
F., Ogata, H., Speich, S., Stemmann, L., Weissenbach, J., Wincker, P.,
Acinas, S. G., Sunagawa, S., Bork, P., Sullivan, M. B., Karsenti, E., Bowler,
C., Vargas, C. D., and Raes, J.: Determinants of community structure in the
global plankton interactome, Science, 22, 1262073, 10.1126/science.1262073, 2015.Litchman, E., de Tezanos Pinto, P., Klausmeier, C. A., Thomas, M. K., and
Yoshiyama, K.: Linking traits to species diversity and comminuty structure
in phytoplankton, Hydrobiologia, 653, 15–28, 10.1007/s10750-010-0341-5,
2010.Luo, Y.-W., Doney, S. C., Anderson, L. A., Benavides, M., Berman-Frank, I., Bode, A., Bonnet, S., Boström, K. H., Böttjer, D., Capone, D. G., Carpenter, E. J., Chen, Y. L., Church, M. J., Dore, J. E., Falcón, L. I., Fernández, A., Foster, R. A., Furuya, K., Gómez, F., Gundersen, K., Hynes, A. M., Karl, D. M., Kitajima, S., Langlois, R. J., LaRoche, J., Letelier, R. M., Marañón, E., McGillicuddy Jr., D. J., Moisander, P. H., Moore, C. M., Mouriño-Carballido, B., Mulholland, M. R., Needoba, J. A., Orcutt, K. M., Poulton, A. J., Rahav, E., Raimbault, P., Rees, A. P., Riemann, L., Shiozaki, T., Subramaniam, A., Tyrrell, T., Turk-Kubo, K. A., Varela, M., Villareal, T. A., Webb, E. A., White, A. E., Wu, J., and Zehr, J. P.: Database of diazotrophs in global ocean: abundance, biomass and nitrogen fixation rates, Earth Syst. Sci. Data, 4, 47–73, 10.5194/essd-4-47-2012, 2012.
Marañón, E., Cermeño, P., López-Sandoval, D. C.,
Rodríguez-Ramos, T., Sobrino, C., Huete-Ortega, M., Blanco, J. M., and
Rodríguez, J.: Unimodal size scaling of phytoplankton growth and the
size dependence of nutrient uptake and use, Ecol. Lett., 16, 371–379,
2013.Marañón, E., Cermeño, P., Latasa, M., and Tadonléké, R. D.:
Resource supply alone explains the variability of marine phytoplankton size
structure, Limnol. Oceanogr., 60, 1848–1854, 10.1002/lno.10138, 2015.
Margalef, R.: Life-forms of phytoplankton as survival alternatives in an
unstable environment, Oceanol. Acta, 1, 493–509, 1978.
Marshall, J., Adcroft, A., Hill, C. N., Perelman, L., and Heisey, C.: A
finite-volume, incompressible Navier–Stokes model for studies of the ocean
on parallel computers, J. Geophys. Res., 102, 5753–5766, 1997.Matthiessen B. and Hillebrand, H.: Dispersal frequency affect local biomass
production by controlling local diversity, Ecol. Lett., 9, 652–662,
10.1111/j.1461-0248.2006.00916.x, 2006.
McCann, K.: The diversity-stability debate, Nature, 405, 228–233, 2000.
Miller, T. E. H., Burns, J. H., Munguia, P., Walters, E. L., Kneitel, J. M.,
Richards, P. M., Mouquet, N., and Buckley, H. L.: A critical review of twenty
years use of the resource-ratio theory, Am. Nat., 165, 439–448, 2005.Monteiro, F. M., Bach, L. T., Brownlee, C., Brown, P., Rickaby, R. E. M.,
Tyrrell, T., Beaufort, L., Dutkiewicz, S., Gibbs, S., Gutowska, M. A., Lee,
R., Poulton, A. J., Riebesell, U., Young, J., and Ridgwell, A.: Why marine
phytoplankton calcify?, Sci. Adv., 2, e1501822, 10.1126/sciadv.1501822, 2016.Moore, L. R., Rocap, G., and Chisholm, S.: Physiology and molecular
phylogeny of coexisting Prochorococcus ecotypes, Nature, 393, 464–446, 1998.
Moore, T. S., Campell, J. W., and Dowel, M. D.: A class-based approach to
characterizing and mapping the uncertainty of the MODIS ocean chlorophyll
product, Remote Sens. Environ., 113, 2424–2430, 2009.Morel, A., Ahn, Y.-H., Partensky, F., Vaulot, D., and Claustre, H.:
Prochlorococcus and Synechococcus: A comparative study of their optical properties in relation to their
size and pigmentation, J. Mar. Res., 51, 617–649, 1993.Mouriño-Carballido, B., Hojas, E., Cermeño, P., Chouciño, P.,
Fernández-Castro, B., Latasa, M., Marañón, E., Morán,
X. A. G., and Vida, M.: Nutrient supply controls picoplankton community structure
during three contrasting seasons in the northwestern Mediterranean Sea, Mar.
Ecol.-Prog. Ser., 543, 1–19, 10.3354/meps11558, 2016.Mousing, E. A., Richardson, K., Bendtsen, J., Cetinic, I., and Perry, M. J.:
Evidence of small-scale spatil structuring of phytoplankton alpha- and
beta-diversity in the open ocean, J. Ecol., 104, 1682–1695,
10.1111/1365-2745.12634, 2016.O'Brien, C. J., Peloquin, J. A., Vogt, M., Heinle, M., Gruber, N., Ajani, P., Andruleit, H., Arístegui, J., Beaufort, L., Estrada, M., Karentz, D., Kopczyńska, E., Lee, R., Poulton, A. J., Pritchard, T., and Widdicombe, C.: Global marine plankton functional type biomass distributions: coccolithophores, Earth Syst. Sci. Data, 5, 259–276, 10.5194/essd-5-259-2013, 2013.Pančić, M., Rodriguez Torres, R., Almeda, R., and Kiørboe,
T.: Silicified cell walls as a defensive trait in diatoms, P.
R. Soc. B, 286, 20190184, 10.1098/rspb.2019.0184, 2019.
Poulin, F. J. and Franks, P. J. S.: Size-structured planktonic ecosystems:
constraints, controls and assembly instructions, J. Plankton
Res., 32, 1121–1130, 2010.Powell, M. G. and Glazier, D. S.: Asymmetric geographic range expansion
explains the latitudinal diversity gradients of four major taxa of marine
plankton, Paleobiology, 42, 196–208, 10.1017/pab.2016.38, 2017.Prowe, A. E. F., Pahlow, M., Dutkiewicz, S., Follows, M. J., and Oschlies,
A.: Top-down control of marine phytoplankton diversity in a global ecosystem
model, Prog. Oceanogr., 101, 1–13,
10.1016/j.pocean.2011.11.016, 2012.
Ptacnik, R., Solimini, A. G., Andersen, T., Tamminsen, T., Brettum, P.,
Lepisto, L., Willen, E., and Rekolainen, S.: Diversity predicts stability and
resource use efficiency in natural phytoplankton communities, P. Natl.
Acad. Sci. USA, 105, 5134–5138, 2008.Raven, J. A.: Why are there no picoplanktonic O2 evolvers with volumes less
than 10-19 m3?, J. Plankton Res., 16, 565–580, 1994.Reich, P. B., Tilman, D., Isbell, F., Mueller, K., Hobbie, S. E., Flynn,
D. F. B., and Eisenhauer, N.: Impacts of biodiversity loss excalates through
time as redundancy fades, Science, 336, 589–592,
10.1126/science.1217909, 2012.Righetti, D., Vogt, M., Gruber, N., Psomas, A., and Zimmermann, N. E.: Global
patterns of phytoplankton diversity driven by temperature and environmental
variability, Sci. Adv., 5, eaau6253, 10.1126/sciadv.aau6253, 2019.
Rodriguez-Ramos, M., Marañón, E., and Cermeno, P.: Marine nano- and
microphytoplankton diversity: redrawing global patterns from
sampling-standardized data, Global Ecol. Biogeogr., 24, 527–538, 2015.
Sal, S., López-Urrutia, Á., Irigoien, X., Harbour, D. S., and Harris,
R. P.: Marine microplankton diversity database: ecological archives E094-149,
Ecology, 94, p. 1658, 2013.
Sarthou, G., Timmermans, K. R., Blain, S., and Treguer, P.: Growth physiology
and fate of diatoms in the ocean: A review, J. Sea Res., 53, 25–42, 2005.Schade, J. D., Espeleta, J. F., Klausmeier, C. A., McGroddy, M. E.,
Thomas, S. A., and Zhang, L.: A conceptual framework for ecosystem
stoichiometry: Balancingresource supply and demand, Oikos, 109, 40–51,
10.1111/j.0030-1299.2005.14050.x, 2005.Schartau, M., Landry, M. R., and Armstrong, R. A.: Density estimation of
plankton size spectra: a reanalysis of IronEx II data, J. Plankton
Res., 32, 1167, 10.1093/plankt/fbq072, 2010.Shi, D., Kranz, S. A., Kim, J.-M., and Morel, F. M. M.: Ocean acidification
slows nitrogen fixation and growth in the dominant diazotroph Trichodesmium
under low-iron conditions, P. Natl. Acad. Sci. USA,
10.1073/pnas.1216012109, 2012.Smith, V. H.: Microbial diversity-productivity relationships in aquatic
ecosystems, FEMS Microbial Ecol., 62, 181–186,
10.1111/j.1574-6941.2007.00381.x, 2007.Snow, J. T., Schlosser, C., Woodward, E. M. S., Mills, M. M., Achterberg, E. P.,
Bibby, T. S., and Moore, C. M.: Environmental controls on the biogeography of
diazotrophy and Trichodesmium in the Atlantic Ocean, Global Biogeochem. Cy., 29, 865–884,
10.1002/2015GB005090, 2015.
Sommer, U.: Nitrate and silicate competition among Antarctic phytoplankton,
Mar. Biol., 91, 345–351, 1986.Sommer, U., Charalampous, E., Genitsaris, S., and Moustaka-Gouni, M.:
Benefits, costs and taxonomic distribution of marine phytoplankton body
size, J. Plankton Res., 39, 494–508, 10.1093/plankt/fbw071, 2017.Sommer, U., Charalampous, E., Cotti, M., and Moustaka-Gouni, M.: Big fish
eat small fish: Implications for food chain length?, Community Ecol.,
19, 107–115, 10.1556/168.2018.19.2.2, 2018.
Stramski, D., Bricaud, A., and Morel, A.: Modeling the inherent optical
properties of the ocean based on the detailed composition of the planktonic
community, Appl. Optics, 40, 2929–2945, 2001.
Tang, E. P. Y.: The allometry of algal growth rates, J. Plankton Res., 17,
1325–1335, 1995.
Taniguchi, D. A. A., Landry, M. R., Franks, P. J. S., and Selph, K. E.:
Size-specific growth and grazing rates for picophytoplankton in coastal and
oceanic regions of the eastern Pacific, Mar. Ecol.-Prog. Ser., 509,
87–101, 2014.Terseleer, N., Bruggeman, J., Lancelot, C., and Gypens, N.: Trait-based
respresentation of diatom functional diversity in a plankton functional type
model of the eutrophied southern North Sea, Limnol. Oceanogr., 59, 1958–1972,
10.4319/lo.2014.59.6.1958, 2014.
Tilman, D.: Resource competition between planktonic algae: An experimental
and theoretical approach, Ecology, 58, 338–348, 1977.
Tilman, D.: Resource Competition and Community Structure, Pop. Biol.,
Vol. 17, 296 pp., Princeton Univ. Press, Princeton, N.J., 1982.Thomas, M. K., Kremer, C. T., Klausmeier, C. A., and Litchman, E.: A global
pattern of thermal adaptation in marine phytoplankton, Science, 338, 1085–1088, 10.1126/science.1224836, 2012.Tréguer, P., Bowler, C., Moriceau, B., Dutkiewicz, S., Gehlen, M.,
Leblanc, K., Aumont, O., Bittner, L., Dugdale, R., Finkel, Z., Iudicone, D.,
Jahn, O., Guidi, L., Lasbleiz, M., Levy, M., and Pondaven, P.: Influence of
diatoms on the ocean biological pump, Nat. Geosci., 11, 27–37,
10.1038/s41561-017-0028-x, 2017.
Vallina, S., Cermeno, P., Dutkiewicz, S., Loreau, M., and Montoya, J. M.:
Phytoplankton functional diversity increases ecosystem productivity and
stability, Ecol. Model., 361, 184–196, 2017.
Vallina, S. M., Ward, B. A., Dutkiewicz, S., and Follows, M. J.: Maximal
foraging with active prey-switching: a new “kill the winner” functional
response and its effect on global species richness and biogeography, Prog.
Oceanogr., 120, 93–109, 2014a.Vallina, S. M., Follows, M. J., Dutkiewicz, S., Montoya, J., Cermeno, P., and
Loreau, M.: Global relationship between phytoplankton diversity and
productivity in the ocean, Nat. Commun., 5, 4299, 10.1038/ncomms5299,
2014b.Verdy, A., Follows, M., and Flierl, G.: Optimal phytoplankton cell size in an
allometeric model, Mar. Ecol.-Prog. Ser., 379, 1–12, 10.3354/meps07909,
2009.Ward, B. A.: Temperature-Correlated Changes in Phytoplankton Community
Structure Are Restricted to Polar Waters, PLOS ONE, 10, e0135581,
10.1371/journal.pone.0135581, 2015.
Ward, B. A., Dutkiewicz, S., Jahn, O., and Follows, M. J.: A size structured
food-web model for the global ocean, Limnol. Oceanogr., 57,
1877–1891, 2012.
Ward, B. A., Dutkiewicz, S., Moore, C. M., and Follows, M. J.: Iron, phosphorus
and nitrogen supply ratios define the biogeography of nitrogen fixation,
Limnol. Oceanogr., 58, 2059–2075, 2013.Ward, B. A., Dutkiewicz, S., and Follows, M. J.: Modelling spatial and
temporal patterns in size-structured marine plankton communities: top-down
and bottom-up controls, J. Plankton Res., 36, 31–47,
10.1093/plankt/fbt097, 2014.
Ward, B. A., Marañón, E., Sauterey, B., Rault, J., and Claessen, C.: The
size-dependence of phytoplankton growth rates: a trade-off between nutrient
uptake and metabolism, Am. Nat., 189, 170–177, 2017.Webb, E. A., Ehrenreich, I. M., Brown, S. L., Valois, F. W., and Waterbury,
J. B.: Pheotypic and genotypic characterization of multiple strains of the
diazotrophic cyanobacterium, Crocosphaera watsonii, isolated from the open ocean, Environ.
Microbiol., 11, 338–348, 2009.Weisse, T., Anderson, R., Arndt, H., Calbet, A., Hansen, P. J., and Montagnes,
D. J. S.: Functional ecology of aquatic phagotrophic protists – Concepts,
limitations, and perspectives, Eur. J. Protistol., 55, 50–74,
10.1016/j.ejop.2016.03.003, 2016.Wilson, J. B., Spijkerman, E., and Huisman, J.: Is there really insufficient
support for Tilman's R* concept? A comment on Miller et al., Am. Nat., 169,
700–706, 2007.
Wilson, S. T., Aylward, F. O., Ribalet, F., Benedetto, B., Case, J. R.,
Connell, P. E., Eppley, J. M., Ferran, S., Fitzsimmons, J. N., Hayes, C. T.,
Romano, A. E., Turk-Kubo, K. A., Vislova, A., Armbrust, E. V., Carno, D. A.,
Church, M. J., Zehr, J. P., Karl, D. M., and DeLong, E. F.: Coordinated
regulation of growth, activity and transcription in natural populations of
the unicellular nitrogen-fixing cyanobacterium Crocosphaera, Nat. Microbiol., 2, 17118,
10.1038/nmicrobiol.2017.118, 2017.
Worm, B., Hillebrand, H., and Sommer, U.: Consumer versus resource control
of species diversity and ecosystem functioning, Nature, 417, 848–851, 2002.
Wunsch C. and Heimbach, P.: Practical global ocean state estimation, Physica
D, 230, 197–208, 2007.