Download - Daniel G. Boyce

Ecology Letters, (2015)
doi: 10.1111/ele.12434
IDEA AND
PERSPECTIVE
Daniel G. Boyce,1,2* Kenneth T.
Frank2 and William C. Leggett1
From mice to elephants: overturning the ‘one size fits all’
paradigm in marine plankton food chains
Abstract
It is widely believed that consumer control is a weak regulator of marine phytoplankton communities. It remains unclear, however, why this should be the case when marine consumers routinely
regulate their prey at higher trophic levels. One possibility is that the weak consumer control of
phytoplankton communities results from the inability of field researchers to effectively account for
consumer–prey trophic relationships operating at the scale of the plankton. We explored this issue
by reviewing studies of trophic control in marine plankton. Experimental studies indicate that size
is a critical determinant of feeding relationships among plankton. In sharp contrast, of the 51 field
studies reviewed, 78% did not distinguish among the sizes or species of phytoplankton and their
consumers, but instead assumed a general bulk phytoplankton–zooplankton trophic connection.
Such an approach neglects the possibility that several trophic connections may separate the smallest phytoplankton (0.2 lm) from the larger zooplankton (~ 1000 lm), a remarkable size differential exceeding that between a mouse (~10 cm) and an elephant (~2500 cm). The size-based
approach we propose integrates theory, experiments and field observations and has the potential
to greatly enhance our understanding of the causes and consequences of recently documented
restructuring of plankton communities.
Keywords
Consumer control, marine plankton, resource control, size-based predation, trophic control.
Ecology Letters (2015)
INTRODUCTION
It is a common belief that over most of the ocean, annual trajectories of marine phytoplankton biomass are primarily regulated by resource availability rather than by consumers
(Margalef 1978; Aebischer et al. 1990; Di Lorenzo et al. 2013).
This view is largely based on statistical analyses of annual time
series that generally indicate strong climate-driven resource
regulation of phytoplankton (Behrenfeld et al. 2006), and weak
grazer regulation (Micheli 1999; Shurin et al. 2002). This perspective shapes our view of how marine ecosystems will
respond to ongoing climate change, and of the factors driving
observed changes in phytoplankton standing stock (Boyce
et al. 2010, 2014) and phytoplankton community restructuring
(Polovina et al. 2008; Mor
an et al. 2010; Polovina & Woodworth 2012) over decadal to multi-decadal scales. For example,
resource-driven models suggest that phytoplankton concentrations at high latitudes will increase with warming, due to
reduced upper ocean mixing (Mora et al. 2013). However, this
contrasts sharply with experimental evidence that indicate phytoplankton concentrations at these latitudes will decline, due to
increased consumer control by grazers as temperature increases
(O’Connor et al. 2009; Lewandowska et al. 2014).
Notwithstanding the current widely accepted bottom-up
view of phytoplankton dynamics, mechanistic explanations of
1
Department of Biology, Queen’s University, Kingston, ON, Canada K7L 3N6
2
Ocean Sciences Division, Bedford Institute of Oceanography, PO Box 1006,
why consumer control should be the dominant regulator of
heterotrophs but not of phytoplankton is lacking. It has been
hypothesised that ecosystem traits such as the degree of prey
switching, or the levels of zooplankton diversity or omnivory
may buffer against strong consumer regulation of phytoplankton (Sommer 2008). Another viable possibility is that the
widely accepted hypothesis of weak consumer control of phytoplankton communities is an artefact resulting from the
widespread inability of field-based studies to adequately
resolve predator–prey trophic connections within marine
microbial communities.
Here, we use published studies to document the extent to
which trophic connectedness has been unaccounted for in field
studies of plankton trophodynamics. We then revisit the
importance of size-based predation in trophic ecology and
review recent advances that enable the estimation of phytoplankton size and community dynamics at macroecological
scales. Finally, we employ longstanding and well-established
experimentally derived size-based predation relationships to
advance an alternate approach to the estimation of trophic
connectedness, trophic control and food chain structure in
marine plankton. Our objective is to facilitate a more accurate
resolution of the trophic controls that regulate plankton communities at large scales, and thereby to improve our understanding of the causes and consequences of recently observed
*Correspondence: E-mail: dgb1@queensu.ca
Dartmouth,NS, Canada B2Y 4A2
© 2015 John Wiley & Sons Ltd/CNRS
2 D. G. Boyce, K. T. Frank and W. C. Leggett
large-scale changes in marine plankton communities (Polovina
et al. 2008; Boyce et al. 2010, 2014; Mor
an et al. 2010;
Polovina & Woodworth 2012).
QUANTIFICATION OF TROPHIC CONTROL
The relative importance of resource availability vs. consumer
control in structuring ecosystems has been difficult to resolve
because these types of trophic control are not directly observable in situ and must instead be estimated. In short-term or
small-scale studies, trophic control can be quantified using gut
content or biochemical tracer analyses, electronic tagging or
dilution experiments. However, these methods are impractical
when attempting to assess trophic control for the complete
spectrum of species, over large spatial areas or over multiple
years. Moreover, small-scale experimental approaches may
not accurately reproduce the complexity and realism of natural systems. Consequently, observable patterns such as the
interaction strength between spatial or temporal gradients in
the abundance of predators and their prey are commonly and
extensively used as state indicators of trophic control in natural systems (Fig. 1). While many such state indicators of trophic control have been used, perhaps the most common is the
correlation coefficient (rTC), which has been used extensively
to quantify the type and strength of trophic control in marine
(i.e. Frank et al. 2006), terrestrial (i.e. Post et al. 1999) and
freshwater (i.e. Carpenter & Kitchell 1993) studies and is commonly featured in contemporary ecological textbooks (Moloney et al. 2011). The validity of this and other state indicator
approaches relies on two critical assumptions:
(1) Trophic connectedness: There is an implicit assumption
that the predator and prey species are trophically coupled
in space and time.
(2) Causality: The interaction between predator and prey
arises due to a causative trophic effect, rather than in
response to an independent factor (sensu the Moran
effect; Moran 1953).
These assumptions are interdependent: species must be trophically connected to have a causative trophic effect. Application of the method to non-trophically coupled species could
therefore obscure the detection of consumer control, leading
the researcher to overestimate the prevalence of resource control (Fig. 1). For this reason, accurate estimation of the
degree of trophic connectedness is a critical prerequisite to the
establishment of causality and to quantifying trophic control.
How to establish this connectedness is the primary focus of
our review.
The trophic connectedness of species and topology of food
webs has been successfully estimated using the stochastic theoretical cascade model (Cohen & Newman 1985). Within the cascade model, the structure of the food web is determined by the
trophic position of the component species: species of higher trophic position may consume only those occupying a lower position. The theoretical cascades model has been adapted by
empiricists, and it is now common and widespread to use species trophic positions to estimate food web structure and trophic connectivity. In turn, the trophic position of an individual
species requires detailed knowledge of the diet preferences of
© 2015 John Wiley & Sons Ltd/CNRS
Idea and Perspective
the consumer and the trophic position of the prey. Accordingly,
we use ‘trophic positioning model’ to refer to the use of estimated trophic levels to determine food chain structure and feeding connectivity. The trophic position, or fractional trophic
level, of any consumer species can be estimated as
TLi ¼ 1 þ Rj TLj DCij ;
ð1Þ
where TLj is the fractional trophic level of the prey j, and DCij
is the fraction of species j in the diet of i (Pauly & Palomares
2005). While larger-bodied species are easily identified, are
often extensively harvested, and their diets determined through
dissection, detailed diet information for small-bodied species
such as microscopic plankton are difficult to obtain. Moreover, planktonic species may exhibit frequent prey-switching
behaviour (Stibor et al. 2004; Sommer & Sommer 2006), and
can be mixotrophic (Zubkov & Tarran 2008). Therefore, trophic levels for plankton species may be less reliably estimated,
and are likely to vary in space and time (Sommer et al. 2002;
Sommer & Sommer 2006). Indeed, the trophic positioning
model may be more data-intensive and difficult to apply in
aquatic systems, because ontogenetic changes in diet are much
more common in these than in terrestrial systems. Thus, many
marine species are likely to possess a different fractional trophic level at each ontogenetic stage, thus reducing the usefulness of a species-level trophic estimate.
Notwithstanding these complexities, researchers have, and
continue, to commonly employ the trophic positioning model
when studying trophic control in marine ecosystems and have
thereby made the implicit assumption that all phytoplankton
(trophic level 1) are identical in their trophic suitability, and
that all grazing zooplankton (trophic level 2) are identical in
their prey preferences (i.e. Smith 1982; Aebischer et al. 1990;
Daskalov et al. 2007).
It is now well-understood that plankton food chains are
complex, with several trophic transfers potentially separating
the smallest phytoplankton cells from the largest zooplankton
(Pomeroy 1974; Azam et al. 1983; Stibor et al. 2004). For
example, small phytoplankton cells (0.2 lm) cannot be efficiently grazed by larger zooplankton (~ 1000 lm) of the size
that are typically retained by commonly used zooplankton
sampling platforms. Instead, phytoplankton cells in this size
range may be grazed by smaller heterotrophic nanoflagellates
or ciliates within a microbial food chain (Fig. 2a), which, due
to their small size, are rarely retained by sampling instruments
and almost never considered in measures of total zooplankton
biomass.
This convenient, but in many ways misleading conceptual
model leads to a misrepresentation of food chain structure
and to ‘top-heavy’ food chains, in which the trophic relationships between higher trophic level consumer species are much
better represented and understood than is the case for plankton, which are often simplistically and unrealistically represented as bulk biomass pools of phytoplankton and
zooplankton (Fig. 2b). This approach neglects the complex
predator and prey dynamics of plankton food webs and has
at best obscured, and at worst biased the understanding of
the trophic controls operating on plankton communities at
macroecological scales. We therefore argue that an alternate,
size-based approach to the assessment of trophic connectivity
Idea and Perspective
Size-based structuring of plankton ecosystems 3
Trophic level 3
(i.e. small fish)
Trophic level 2
(i.e. zooplankton)
Trophic level 1
(Primary producers)
(a) Trophic connectedness. Trophic linkages
between species within the foodweb are
developed
Consumer control
Resource control
(b) Measurements of abundance. Temporal or
spatial measurements of abundance of different
components of foodweb are assessed
Relative abundance
3
2
1
0
−1
−2
−3
1900
1940
1900
1980
1940
1980
Year
3
Trophic control suggested by state indicators
Consumer control
Resource control
Trophic level 1
r = 0.93
*
1
0
−1
−2
r = −0.94
r = 0.93
2
1
0
−1
−2
−3
3
* Application of method to non-trophically
coupled foodweb components (top right and
bottom left panels in c) will bias the
detection of consumer control but not of
resource control
r = 0.86
2
−3
3
Trophic level 2
Interaction strength between abundance
measurements of trophically coupled foodweb
components indicates the type and
strength of trophic control. Positive
interactions indicate resource control
and negative interactions denote
consumer control
Trophic level 3
(c) State indicators of trophic control.
2
r = 0.86
*
r = −0.92
1
0
−1
−2
−3
−3 −2 −1
0
1
Trophic level 3
2
3 −3 −2 −1
0
1
Trophic level 2
2
3 −3 −2 −1
0
1
2
3
Trophic level 1
Figure 1 Statistical state indicators of trophic control. Simulated time-series of abundance of trophically connected food web components (species or
functional groups) are used to infer the type and strength of trophic control at macroecological scales for a three trophic level food web. (a) Trophic
connections between species within the food chain are developed. (b) Simulated time-series of relative biomass for different trophic levels are plotted over
the last century (1900–2010). (c) The interaction between abundance time-series of trophically coupled food chain components provides an index of trophic
control. In this case, the interaction metric is the correlation coefficient and is shown on each plot. *Application of the method to food web components
that are not trophically coupled does not prevent the detection of resource control or weak trophic control, but may lead to the incorrect conclusion that
consumer controlled systems are actually resource control (bottom left panel in c). For all plots, colours depict the food web components. Green denotes
trophic level 1, yellow denotes trophic level 2 and red denotes trophic level 3. Columns denote simulations of consumer (left) and resource control (right).
and control in the phytoplankton would be a substantial
improvement over the current trophic positioning model.
IMPORTANCE OF SIZE
Organism size is an critical determinant of many ecological
processes including metabolism, energy use, mobility, produc-
tion and mortality (Woodward et al. 2005). In the trophic
context, organism size sets first-order constraints on feeding
behaviour, and thus regulates food web structure and the
pathways and efficiencies by which primary production flows
through ecosystems (Ryther 1969; Sheldon et al. 1972; Cohen
et al. 1993; Sommer et al. 2002; Barnes et al. 2010). For vertebrate consumers, feeding relationships are well-described by
© 2015 John Wiley & Sons Ltd/CNRS
4 D. G. Boyce, K. T. Frank and W. C. Leggett
Idea and Perspective
Fish, mammals
and seabirds
(a) 10000
Predator length (mm)
(a)
Trophic position
Mesozooplankton
Ciliates
Habitat
Freshwater
Pond
Stream
1000
100
10
Nanoflagellates
r2 = 0.72
1
Trophic transfer
2
5
10
Energy flow
100
(b)
Nano- Micro-phytoplankton
101
102
103
104
105
100
200
500
1000
(b)
Fish, mammals
and seabirds
Habitat
10000
106
Organism size (μm)
Trophic position
50
Prey length (mm)
Predator length (mm)
Pico-
20
1000
100
Estuarine
Transition zone
Shelf
Seasonal ice zone
Open ocean
Nearshore
Demersal
Pelagic
10
1
r2 = 0.83
0.2
0.5
1.0
2.0
5.0
10.0 20.0
50.0 100.0 200.0
500.0
Prey length (mm)
Heterotroph
Zooplankton
Autotroph
Cannibal
Phytoplankton
Figure 2 Food web structure. (a) Qualitative model depicting the flux of
primary production through the microbial marine food chain. The
number of trophic transfers, and hence the efficiency of energy transfer,
between primary producers and fishes depends on the size structure of the
community. Arrows depict the energy flow and blue circles show trophic
transfers through the food chain. Colours depict the different pathways
by which primary production flows up the food chain. Figure was
adapted from Azam et al. (1983). (b) Visual representation of the
estimated trophic structure obtained from a published food web model of
the Icelandic marine ecosystem (Buchary 2001). Circles represent different
species or species groups, semi-transparent lines represent feeding
relationships between components. Top consumers are represented as
dark circles and are at the top of the diagram, light blue circles are lower
trophic level consumers, cannibalistic species are orange and
phytoplankton are green.
predator–prey body size relationships that explain a large
fraction of the variability in feeding interactions within freshwater (Fig. 3a), and marine (Fig. 3b) ecosystems (Reuman &
Cohen 2004; Brose et al. 2006; Barnes et al. 2010). These sizebased relationships appear to be broadly general, existing
both within, and across different habitats and ecosystems, and
for species spanning multiple taxonomies and body sizes. Such
© 2015 John Wiley & Sons Ltd/CNRS
Figure 3 The size of vertebrate consumers and their prey. (a) The average
size (mm) of vertebrate predators and their prey from freshwater
ecosystems (data from Brose et al. 2006). (b) The average size of
predators and their prey from more than 27 marine ecosystems (Brose
et al. 2006; Barnes et al. 2010). Colours depict the habitat from which the
measurements were collected. Lines and shaded areas depict the model II
OLS linear regression line and 95% confidence limits about the mean
respectively. The dashed line in (a) is a generalised additive model
regression fitted to the data. None of the relationships involve
phytoplankton.
predator–prey size relationships are believed to arise primarily
due to gape-size constraints (Bremigan & Stein 1994), for
which body size is a more easily measured and reliable proxy.
Body size is also incorporated into some ecosystem food web
models. For instance size has been used to parameterise predator attack rates within a newly developed general end-to-end
ecosystem model (Harfoot et al. 2014), and appears to be closely aligned to niche space within widely used niche models.
Indeed, niche models using organism size alone to form the
niche axis are capable of correctly predicting 57–68% of the
trophic connections in the food web (Petchey et al. 2008; Williams et al. 2010).
The importance of size-based predation has also been demonstrated experimentally in plankton food chains. Consistent
with the conceptual model of the microbial food chain
(Fig. 2a), an early freshwater experiment by Burns (Burns
1968) found a strong linear relationship between the carapace
length of a filter-feeding zooplankton (Cladocera) and the
Idea and Perspective
length of spherical plastic beads (1–80 lm diameter) that
served to simulate their prey. Using a collection of experimental and field observations derived from aquatic foodwebs,
Warren (1987) found that organism length could explain 90%
of the variability in predator–prey feeding relationships
among invertebrates (P < 0.0001; n = 28). In a series of mesocosm experiments, Stibor et al. (2004) demonstrated contrasting food web configurations and food chain length that were
highly dependent on the initial size structure of the phytoplankton community. An ecosystem supported mainly of
large-sized phytoplankton (105 lm3) resulted in a three-level
food chain (large phytoplankton–copepods–predators) and
strong predator effects on phytoplankton. Alternatively, an
ecosystem supported primarily of smaller phytoplankton
(52 lm3) led to a four-level food chain (small phytoplankton–
ciliates–copepods–predators) and weak predator effects. The
pooled results from these experiments yielded a neutral
response of the phytoplankton community to predator manipulation. These findings further emphasise the importance of
accounting for size structure in trophic ecology.
Despite the demonstrated importance of size in structuring
ecological food chains, our literature review revealed that the
potential influence of size-based predation is rarely incorporated into the design or analysis of field-based studies of trophic control in the marine plankton (see Supporting
information for details of literature search). Only ~ 20% (10
of 51) of field-based and non-experimental studies involving
plankton trophic dynamics we reviewed evaluated the roles of
organism size or species composition in determining trophic
connectedness. This proportion declined to 12.5% (3 of 24)
when only studies conducted at macroecological scales (spatial
scales greater than ‘station’ and temporal scales greater than
5 years) were considered. Overall, the proportion of studies
that evaluated organism size or species composition was greatest for short term, small-scale studies, and least for macroecological studies. Of the studies evaluated, zooplankton
abundance was commonly obtained using plankton nets or
the continuous plankton recorder (CPR), from which the minimum zooplankton size can be easily inferred from the mesh
size of the instrument. In contrast, phytoplankton abundance
was most frequently assessed using measurements of total
chlorophyll-a concentration (mg m3), ocean colour (‘greenness’) or primary production (mg C m2 D) – methods that
subsume the entire phytoplankton assemblage and lump species spanning several orders of magnitude in size (~ 1–
1000 lm). Incredibly, this size differential is approximately
proportional to that between a mouse and an elephant (~ 10–
2500 cm or ~ 0.05–15 000 lbs).
Zooplankton grazing on phytoplankton was assumed in all
51 studies we reviewed. However, the reality is that by
neglecting to consider the potential effects of size-based predation (Fig. 2a), it is possible that no direct trophic connectedness existed between zooplankton and phytoplankton in
~ 80% of the studies reviewed. Equally relevant, in those
studies in which plankton size was reported, the average minimum size of the zooplankton retained by the sampling instrumentation was 331 lm. This contrasts dramatically with
evidence that microzooplankton between 20 and 200 lm are
the dominant grazers in the world oceans, consuming between
Size-based structuring of plankton ecosystems 5
60 and 75% of daily primary production (Calbet & Landry
2004; Landry & Calbet 2004). This difference between the
minimum size of zooplankton retained by sampling instrumentation (331 lm) and the size of the dominant grazers (20–
200 lm), suggests that there may be at least one trophic transfer between the sampled zooplankton assemblage and phytoplankton. Hence, the assumption of direct trophic
connectedness between zooplankton and phytoplankton in the
majority of the 51 published field studies we reviewed is inappropriate.
As an alternative to assuming that all phytoplankton are
grazed by all zooplankton, as many studies currently do, we
propose that estimates of plankton size may be used to more
accurately resolve trophic connectedness in plankton food
chains. The idea of size-based food chains is not new. Size is
a directly observable and measureable trait, and is directly
related, through theory and observation, to the trophic
dynamics of all organisms. The size-based approach has
proved successful in predicting trophic connectedness in ecosystem food web models (Petchey et al. 2008; Williams et al.
2010), and several studies have found that body size is interchangeable with the trophic position of an organism (Elton
1927; Trebilco et al. 2013).
ESTIMATING PLANKTON SIZE AT
MACROECOLOGICAL SCALES
The quantification of phytoplankton cell size or species composition has historically been a difficult and painstaking process requiring microscopic enumeration, flow cytometry, sizefractionated filtration, various particle size distribution techniques or other time consuming methods (Table 1 contains
full summary of available methods and Table S1 contains references). This may explain why size-based predation information is rarely incorporated into plankton trophic studies.
Obtaining continuous time-series of phytoplankton size measurements over regional spatial scales using such methods has
been virtually impossible given limitations linked to vessel and
personnel cost and availability and the expense of sampling
remote ocean areas. The few decadal (> 10 year) time-series
of phytoplankton size or community composition that do
exist, originate from fixed station locations, such as the
Hawaiian Ocean Time-Series (HOTS) and Bermuda Atlantic
Time-Series (BATS). The one exception is the continuous
plankton recorder (CPR) time-series which has been sampling
the northeast Atlantic Ocean more or less continuously since
~ 1940. However, the CPR uses a 270 lm mesh that may not
accurately sample the entire zooplankton or phytoplankton
size spectrum (Dippner & Krause 2013). For example, picophytoplankton (< 2 lm diameter) comprise more than 90%
of the phytoplankton biomass in some oligotrophic regions of
the north Atlantic (Tarran et al. 2000) and elsewhere (Li et al.
1983), and an even greater proportion of primary production
(Stockner & Anita 1986).
Recently, however, remote sensing methodologies that
employ spectral-, abundance-, or ecological-based methods,
particle size distributions and biogeochemical models (see
Table 1 for complete list of these methods and Table S1 for
references) have dramatically increased the ability to estimate
© 2015 John Wiley & Sons Ltd/CNRS
6 D. G. Boyce, K. T. Frank and W. C. Leggett
Idea and Perspective
Table 1 Methods of resolving phytoplankton cell size or community composition
Method
Microscopic
Flow cytometry
Particle size distribution
Sub-method
Coulter counter
Scattering properties
Video
Holographic imagery
Acoustics
Size-fractionated filtration
Pigment (HPLC)
Molecular
Particle size distribution
Spectral-based
Abundance-based
Ecological based
Biogeochemical model
Description
PSC
PFT
Sampling
Enumerated under microscope
Enumerated by laser
Measure displacement volume of particles
Optical scatterometers use inversion algorithms to
estimate particle size from scattering signatures
Underwater video microscope
3D particle distributionstored in 2D
Sonar
Seawater filtered through differently sized filters
Size derived from pigment analyses
Determined via genetic variations
Uses power law distribution of particle size spectrum
Distinguish PFTs based on optical signatures
Based on relationships between trophic state and
phytoplankton community composition
Empirical relationship to relate environment to phytoplankton
Forced by nutrients and physical conditions
Yes
Yes
Yes
Yes
Yes
No
No
No
In
In
In
In
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
No
No
No
No
Yes
Yes
Yes
Yes
Yes
In situ
In situ
In situ
In situ
Remote
In situ
Remote
Remote
Remote
No
No
Yes
Yes
Remote sensing
Remote sensing
situ
situ
situ
situ
sensing
sensing
sensing
sensing
PSC, phytoplankton size class; PFT, phytoplankton functional type. Sampling denotes if the method has been adapted to use remotely sensed observations
or primarily uses in situ methods. Refer to Table S1 for references to these methods.
phytoplankton size and species composition at macroecological scales. These new techniques now make it possible to
approximate the phytoplankton community composition and/
or size structure at daily to weekly time scales and at a near
global resolution.
These new methods differ in their complexity, data requirements, accuracy and output (Fig. 4). The output products
generated are either phytoplankton size classes (PSCs), the
estimated spherical diameter of the phytoplankton cells, or
phytoplankton functional types (PFTs). Although PFTs are
primarily defined based on biogeochemical function rather
than size, per se, many routinely fall into distinct size categories (Roy et al. 2013). The differences between the algorithms
employed to achieve these outputs are significant and have led
to international workshops and collaborations designed to
compare and validate their outputs against in situ observations (http://pft.ees.hokudai.ac.jp/satellite/index.shtml), and a
report has recently been published by the International
Ocean-Colour Coordinating Group (IOCCG) on resolving
PFT’s and PSCs’s using space-borne technologies (Sathyendranath 2014).
HOW EXPERIMENTS, MODELS AND OBSERVATIONS
CAN BE USED TO BETTER UNDERSTAND PLANKTON
TROPHIC DYNAMICS: A PRELIMINARY FRAMEWORK
Here, we provide an example of how the results of experiments conducted at smaller scales can be used in combination
with estimates of phytoplankton sizes derived using newly
developed remote-sensing-based techniques, together with field
observations, to improve the understanding of plankton trophic dynamics.
We compiled measurements on the size of plankton predators and their prey originating from 63 published laboratory
studies (Table S2), and previously assembled by three peerreviewed studies (Hansen et al. 1994; Fuchs & Franks 2010;
Wirtz 2012). These extracted measurements encompass all
© 2015 John Wiley & Sons Ltd/CNRS
major marine and freshwater plankton taxa and include cell
diameters ranging between 0.6 and 2 9 105 lm. This accumulated experimental evidence indicates that a relationship consistent with the well-established predator–prey body size
relationships in freshwater and marine vertebrate consumers
(Fig. 3) also exists for marine plankton (Fig. 5a). The
observed log–log linear relationship relating the size of plankton predators and their prey is based on long-standing ecological theory, and can be formally expressed as
ESDPRD / ESDPRY ;
ð2Þ
where ESDPRD and ESDPRY are the equivalent spherical
diameters of the plankton predator and prey respectively.
Extending this representation, we fitted ordinary least-squares
log–log linear models to the extracted experimental measurements to derive the following predictive equations
ESDPRD ¼ 18:5ESDPRY 0:985
ESDPRY ¼ 0:313ESDPRD 0:744
ð3Þ
ð4Þ
The above equations explain 73% of the variance in the
relationships on a log–log scale and can be used with confidence to estimate trophic connectivity within plankton food
chains. Given the diameter of a plankton predator or prey,
the equations can be used to estimate the prediction intervals
within which a given proportion of future observations will
fall (Fig. 5a). These intervals can be then interpreted as the
probable predator–prey size limits; points falling outside these
intervals are statistically unlikely to be trophically coupled.
The above equations were estimated on a log–log scale, but
are presented on the response scale for clarity.
This approach can be used to better represent and parameterise trophic studies involving plankton. For example
eqns 3 and 4 can be used in conjunction with remote sensing
observations to predict the probable grazer size range, and
when applied sequentially may be used to predict the average
microbial food chain length at any location at any time (Menden-Deuer & Lessard 2000; Barnes et al. 2011; Polovina &
Idea and Perspective
Size-based structuring of plankton ecosystems 7
Nanophytoplankton (2 to 20 μm)
50
40
30
20
10
70
60
50
40
30
20
10
0
Microphytoplankton (20 to >200 μm)
0
80
60
Size classes
40
20
Percent of total chl (% T chl)
Picophytoplankton (0.5 to 2 μm)
Phytoplankton cell diameter
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
−0.1
0.8
0.25
0.6
Functional types
0.4
0.2
0.00
1.0
0.8
Synechococcus + prochlorococcus (0.2 to 2 μm)
0.6
0.0
Diatoms (10 to 200 μm)
Frequency of occurence (%)
1.0
0.50
Log10 Diameter (μm)
Spherical diameter
Remotely sensed ocean
surface observations
0.4
0.2
0.0
Coccolithophores (5 to 100 μm)
Figure 4 Climatalogical global spatial gradients in phytoplankton cell size or functional types. Schematic depicts how available sea surface parameters
(green box) may be used to estimate phytoplankton size class from biogeochemical models (Brewin et al. 2010; red arrows), functional groups using
spectral-based methods (Alvain et al. 2005, 2008; blue arrows) or spherical cell diameter using empirical methods (Barnes et al. 2011; Polovina &
Woodworth 2012; green arrow). The methods used to derive these estimate are presented in Table 1 and Table S1; further details of the data used to
produce this figure are in the supporting information.
Woodworth 2012). Following the approach developed by
Barnes et al. (2011), and adapted by Polovina & Woodworth
(2012), the median phytoplankton cell diameter (MD50) can be
derived as
MD50 ¼ 2:14MB50 0:35
ð5Þ
where MB50 is the median phytoplankton cell weight, which
was estimated by Barnes et al. (2011) as
log10 ðMB50 Þ ¼ 1:34 0:043SST þ 0:929ðlog10 ðChlÞÞ
ð6Þ
where SST is the sea surface temperature (°C) and Chl is the
chlorophyll-a concentration (mg m3). Predictive eqns 5 and 6
are based on ecological theory, observations and experiments
that show that increasing temperature leads to reductions in
phytoplankton cell size (Daufresne et al. 2009; Mor
an et al.
2010). The method is based on a space-for-time (ergodic)
assumption (Rosenzweig 1998) derived from 361 water samples
collected from 12 ocean regions ranging from tropical to temperate, and including upwelling and oligotrophic regions
(Barnes et al. 2011). The predictive model (eqn 6) explains 50%
of the variance in phytoplankton cell size (Barnes et al. 2011).
To demonstrate the utility of this approach, we used eqns 5
and 6 to predict the median phytoplankton cell size globally
during March of 2005 using remotely sensed monthly average
SST and Chl measurements (Fig. 5b; see Supporting information for description of data). We used the above methods
(eqns 5 and 6) to estimate phytoplankton cell size rather than
one of the alternate approaches shown in Fig. 4 or listed in
Table 1, because the method is capable of resolving the continuous cell size spectrum, rather than discrete size classes or functional types. Estimated median phytoplankton cell diameters
across the global ocean during this time period ranged between
0.4 and 17 lm (Fig. 5b), which is consistent with the lower
(0.2 lm; Agawin et al. 2000) and upper (> 200 lm) bounds of
empirically derived estimates. This empirical size range is
expected to be wider, since it includes sizes assessed across
many times and locations, whereas the range estimated here
(Fig. 4) includes sizes assessed during a single month and year.
The emergent spatial patterns of phytoplankton cell size
(Fig. 5b) are primarily related to factors influencing ocean
mixing, temperature and nutrient concentration, consistent
with the results of independent studies (Margalef 1978; Atkin© 2015 John Wiley & Sons Ltd/CNRS
8 D. G. Boyce, K. T. Frank and W. C. Leggett
(a)
Zooplankton diameter (μm)
1x107
Idea and Perspective
r2 = 0.73
1x105
1x103
Prediction interval
50%
60%
70%
80%
90%
95%
99%
1x101
1x10–1
1
10
100
1000
10000
Prey diameter (μm)
(b)
Median phytoplankton diameter (μm)
0.8
Mean 14.8
Min
0.9
Max 246
1.3
2.0
3.0
5.0
8.0
13.0
23.2
1.4
386
36.5
2.2
604
57.5
3.5
947
90.5
5.5
1485
142.5
8.7
2331
224.4
13.7
3662
Optimal grazer diameter range (μm)
(c)
Zooplankton diameter (μm)
1x107
1x105
1x103
1x101
Predator feeding type
Active, (r2 = 0.9 )
Mixed, (r2 = 0.83 )
Passive, (r2 = 0.15 )
1x10–1
1
10
100
1000
10000
Prey diameter (μm)
Figure 5 Size-based predation in microbial ecosystems. (a) The estimated
spherical diameter of marine plankton prey and their consumers derived
from experiments (Table S1). Line depicts the linear regression fit. Shaded
areas represent the prediction intervals about the mean for different levels
of confidence; for the 95% interval, 95% of future values are predicted to
fall within the interval. (b) Estimated median phytoplankton cell diameter
and probable grazer size range estimated globally during March, 2005.
Colours represent the size of phytoplankton cells and of grazers; white
areas contain no data. (c) Same as in (a), but with colours depicting the
primary feeding mode of the zooplankton consumer. Linear regression
parameters were estimated within consumer feeding types.
© 2015 John Wiley & Sons Ltd/CNRS
son 1994; Mor
an et al. 2010). Locations such as the oligotrophic gyres, where temperatures are warmer, ocean mixing
is low and nutrients are limiting tend to favour the smallest
phytoplankton cells. Locations where temperatures are cooler,
mixing is sufficient to deliver nutrients to surface waters and
land-based runoff is high tend to support large-celled phytoplankton assemblages (Fig. 5b). We then used these estimates
of phytoplankton cell diameter to predict the average optimal
size range for grazers from eqn 3 during the same time interval (Fig. 5b). We chose a 95% prediction interval, which
assumes that there is a 95% probability that future observations will fall within the predicted size interval.
This approach could be used to modify the design of sampling programmes aimed at resolving trophic dynamics in
plankton and to determine whether trophic connectedness is
probable. For instance, phytoplankton in the south Pacific gyre
are among the smallest in the ocean (0.2 lm Agawin et al.
2000; Fig. 5b), and the predicted probable grazer size range of
these picophytoplankton is between 0.2 and 65 lm, which is far
below the retained size of most zooplankton sampling platforms (331 lm in our literature review). This knowledge could
be used to estimate the number of trophic transfers separating
picophytoplankton from a zooplankton assemblage of known
size, or to increase the minimum size resolution of the zooplankton sampling instrument. Alternately, the probable prey
size range for a zooplankton assemblage sampled with the
CPR, which uses a 270 lm mesh, is predicted to fall between
1.8 and 228 lm. This knowledge could be used to estimate the
average size or community composition of the contemporaneous phytoplankton assemblage and to determine whether trophic connectedness is probable. Finally, the method could also
be used to estimate the number of trophic transfers and, following the ten per cent law (Lindeman 1942), the transfer efficiency
of primary production up the microbial food chain. For example it has been hypothesised that the small nekton-to-primary
production ratio in open ocean ecosystems (< 0.5%), relative to
upwelling ecosystems (~ 1%; Iverson 1990), arises due to the
small size of primary producers there (Sommer et al. 2002). The
size-based model (Figs 2a and 5a), would suggest that the small
size of primary producers in open ocean ecosystems leads to a
longer food chain with more trophic transfers, and a reduced
fraction of primary production available to higher trophic levels, and may explain the observed difference between the energetic transfer efficiencies in open ocean vs. upwelling
ecosystems. The approach could be further expanded to estimate the total productivity of ecosystems or the capacity of different ecosystems to support fisheries.
Our earlier analysis (Fig. 5a) indicated that phytoplankton
size explains 73% of the variability in the size of their consumer on a log–log scale. Other factors such as the nutritional
content (Cebrian & Cebrain 1999), organism shape (Warren
& Lawton 1987), or the defensive adaptations of phytoplankton may well explain much of the residual variability. Predator feeding modes can also influence size-based feeding
relationships. For instance, small freshwater piranhas may
occasionally consume much larger organisms by group feeding, and large baleen whales can consume zooplankton by filter-feeding. However, such extreme examples are uncommon
(Barnes et al. 2010).
Idea and Perspective
Following the approach used by Wirtz (2012), wherein predator feeding strategy and morphology are considered, we
incorporated the feeding mode of the consumer (active, mixed
or passive) into the simple predictive framework presented
here. This produced different regression parameters and levels
of uncertainty (Fig. 5c). The resulting predator–prey relationship was most variable for passive consumers that filter their
prey from the water using sieve- or comb-like apparatus. Alternately, active consumers that select food individually and are
more constrained by gape-size exhibited a very strong predator–prey size relationship (r2 = 0.9). Prey motility (Tiselius &
Jonsson 1990) and chemical quality (DeMott 1988) may also
influence the prey selection of active zooplankton consumers.
Given the limited experimental data available and the high
explanatory power of the derived predator–prey relationship
(Fig. 5a), we did not attempt to discriminate between different
zooplankton feeding strategies when constructing empirical
predator–prey size relationships (eqns 3 and 4). Incorporation
of these behavioural traits, while requiring an enhanced
knowledge of the zooplankton community, could, ultimately
lead to narrower prediction intervals of the predator–prey size
relationships and improved accuracy and precision of the predictions (Fig. 5c). Given that experimental evidence suggests a
low degree of coexistence between active and passive zooplankton consumers in marine ecosystems (Becker et al. 2004;
Sommer & Sommer 2006), improved knowledge of feeding
strategies may be a productive avenue of future research.
CONCLUSIONS AND IMPLICATIONS
Our analyses suggest that the trophic positioning model, currently widely employed in studies of trophic control in marine
systems, may be inappropriate for resolving trophic connectedness at the level of the plankton, for which trophic position
is difficult to estimate. As an alternative, we propose a traitbased approach based on organism size as a more mechanistically relevant and practical method of resolving trophic connectedness among plankton. Such size-based approaches are
widely used to parameterise predator attack rates in general
ecosystem models (Harfoot et al. 2014), and to define niche
space within niche models. Organism size has also been shown
to be an accurate indicator of trophic position (Elton 1927;
Trebilco et al. 2013). In fact, organism size was used in classical ecological pyramids to denote a species’ position in the
food chain (Elton 1927), and was only later replaced by the
trophic level (Lindeman 1942; Odum & Brewer 1959). Recent
experiments have also convincingly demonstrated that the size
structure of phytoplankton communities can have strong
effects on food chain length and on the detection of trophic
effects (Stibor et al. 2004; Sommer & Sommer 2006).
The advantages of the approach we propose are considerable.
The accurate trophic positioning of plankton in the trophic
positioning model requires identification of the organism and
detailed knowledge of the organisms diet, both of which can be
difficult, costly and time consuming to obtain. Moreover, given
the data-intensive nature of the trophic positioning model, spatio-temporal changes in food chain structure cannot be readily
accounted for. In contrast, plankton size can be directly measured (zooplankton) or estimated globally at high temporal res-
Size-based structuring of plankton ecosystems 9
olution using remote sensing techniques (phytoplankton).
Given knowledge of the size of the plankton predator or prey,
experimentally based empirical relationships can then be used
to determine trophic linkages among plankton of different sizes.
The theory of size-based predation predicts that microbial food
web structure is spatially and temporally dynamic. To further
explore this, eqns 5 and 6 were used to estimate phytoplankton
cell diameter from remotely sensed monthly averaged surface
SST and Chl from 1998 to 2008 (see Supporting information
for details). We then fit statistical models to these observations
to assess the space and time scales of phytoplankton cell size
variability, and hence microbial food chain structure across the
global ocean over this time period (Fig. 4, 5b and 6). Reductions in phytoplankton cell size over the past decade (1998–
2008) were apparent in the Arctic Ocean and in the major oligotrophic gyres of all oceans (Fig. 6a). These Arctic Ocean trends
may be driven by ice melt, and associated freshening of the
ocean there (Li et al. 2009), whereas similar trends in the gyres
have been attributed to ocean warming, reduced mixing and
resulting nutrient limitation (Polovina et al. 2008; Polovina &
Woodworth 2012). In contrast, increases in cell sizes are evident
at most coastal locations, in the Southern Ocean and in low-latitude tropical waters. It remains unclear what drives increases
in size at these locations, but the trend in coastal locations is
likely influenced by land-based eutrophication (Jickells 1998),
which favours large-celled species. In addition to this large spatial (Figs 4 and 5b) and decadal (Fig. 6a) variability, phytoplankton cell diameter also varies at intra (Fig. 6b) and
interannual (Fig. 6c) timescales, and in response to decadal and
multi-decadal climate fluctuations such as El Ni~
no (Fig. 6d).
These dynamic changes in phytoplankton cell size may cause
food chain length and structure to vary on seasonal, yearly, decadal and basin-scales. The existence of non-steady-state dynamics in phytoplankton cell size, food chain structure and the
transfer of energy, are currently overlooked in the design of
most field studies, but could be readily incorporated within the
size-based approach detailed above.
The size-based approach we propose is appealingly simple,
adaptable to temporal and spatial variability, and could lead to
an improved understanding of plankton dynamics and of how
plankton communities will respond to ongoing climate change.
It is important to also highlight that the size-based approach we
propose should serve to compliment and improve existing
approaches, such as trophodynamic models, which often use
the trophic positioning model to determine food chain structure. For example organism size could be used to estimate trophic connectivity in situations in which the trophic level cannot
be adequately resolved. Food chain connections estimated from
organism size could also be used to validate connections based
on trophic levels. In this manner, the size-based approach may
serve to increase the robustness of existing approaches.
There is mounting evidence that marine phytoplankton
communities are undergoing widespread alterations in distribution, abundance and community composition. Observational
studies
suggest
that
marine
phytoplankton
concentrations have declined globally over the past century
(Boyce et al. 2010, 2014), that the average cell size of phytoplankton communities is shrinking (Li et al. 2009; Mor
an
et al. 2010), and that the least productive regions of the ocean
© 2015 John Wiley & Sons Ltd/CNRS
10 D. G. Boyce, K. T. Frank and W. C. Leggett
Idea and Perspective
(a)
0.010
0.000
−0.005
Phytoplankton cell diameter rate
of change log10[μm yr–1]
0.005
14
12
10
8
6
4
2
4
6
8
10
12
Inter-annual
10
(c)
8
6
4
2
1998
2000
2002
Month
2004
2006
2008
Phytoplankton cell diameter [σ]
(b)
Phytoplankton cell diameter [μm]
Intra-annual
Multi-decadal
2
(d) 2
r = −0.66
1
1
0
0
−1
−1
−2
−2
1997
Year
2001
2005
Inverse of El Nino index [σ]
Phytoplankton cell diameter [μm]
−0.010
2009
Year
Figure 6 Spatio-temporal trends in phytoplankton cell diameter. (a) Estimated time trends in the median phytoplankton cell diameter from 1998 to 2008.
Red and yellow are increasing trends, blue is declining, white indicates areas with limited data availability. (b and c) Examples of estimated changes in
phytoplankton cell diameter at (b) intraannual and (c) interannual timescales within individual 1° 9 1° cells for different ocean locations. Lines are from
the mean and 95% confidence interval about the mean estimated from generalised additive models. (d) Co-variation between phytoplankton cell size
(black) and the inverse of the El ni~
no southern oscillation index (red) at an example location in the eastern central Pacific Ocean. Both variables were Zstandardised and are in units of standard deviations from the mean (r).
are expanding (Polovina et al. 2008). These trends are predicted to continue over the next century (Polovina et al. 2011;
Mora et al. 2013), with probable effects on climate, marine
ecosystem structure, geochemical cycling and fisheries. For
example, because large particles sink more rapidly than small
ones, a reduction in average phytoplankton cell size will likely
lead to reduced export production to the deep ocean
(Rodrıguez et al. 2001), the majority of which originates from
phytoplankton (Ruhl et al. 2008). As export production
declines, the sequestration of CO2 to the deep ocean and the
oceanic drawdown of CO2 from the atmosphere may also be
reduced, with effects on the global carbon cycle and climate
(Falkowski 2012). Reduced export production will also have
strong impacts on deep-sea ecosystems, which are almost
entirely sustained by this downward flux of organic matter.
These changes in phytoplankton abundance and community
composition have been largely attributed to physically based
processes associated with ocean warming, such as upper ocean
stratification and nutrient limitation, yet it is unclear how biological control may have driven, or responded to such
changes. Such knowledge is sorely needed, as the majority of
general ocean circulation models used to forecast future ocean
conditions do not currently incorporate potentially important
biological processes such as the influence of environmental
© 2015 John Wiley & Sons Ltd/CNRS
forcing on predator–prey interactions and trophic control. An
enhanced capacity to incorporate such biological information,
made possible in part by an incorporation of size-based
approaches, could well improve the realism and forecasting
accuracy of these models. Experimental studies suggest that
continued warming may lead to strengthened consumer control of phytoplankton (O’Connor et al. 2009), but this effect
is also believed to be context-specific (Lewandowska et al.
2014). It is likewise unclear what effect altered phytoplankton
concentration, size structure and community composition will
have on consumers and on the transfer of energy through the
food chain. Resolving these important questions will require a
greater understanding of the structure of microbial food webs
and of how environmental change will modify this structure.
ACKNOWLEDGEMENTS
We thank B. Worm, B. Li, S. Craig and T. Eddy, for providing valuable comments on this manuscript.
AUTHORSHIP
All authors contributed to the design of the study, data
interpretation and manuscript editing; D.G.B. assembled the
Idea and Perspective
data, performed the statistical analyses and drafted the manuscript.
REFERENCES
Aebischer, N.J., Coulson, J.C. & Colebrook, J.M. (1990). Parallel longterm trends across four marine trophic levels and weather. Nature, 347,
753–755.
Agawin, N.S.R., Duarte, C.M. & Agusti, S. (2000). Nutrient and
temperature control of the contribution of picoplankton to
phytoplankton biomass and production. Limnol. Oceanogr., 45, 591–600.
Alvain, S., Moulin, C., Dandonneau, Y., Breon, F.M. & Breon, F.M.
(2005). Remote sensing of phytoplankton groups in case 1 waters from
global SeaWiFS imagery. Deep Sea Res. Part I Oceanogr. Res. Pap.,
52, 1989–2004.
Alvain, S., Moulin, C., Dandonneau, Y. & Loisel, H. (2008). Seasonal
distribution and succession of dominant phytoplankton groups in the
global ocean: a satellite view. Global Biogeochem. Cycles, 22, 1–15.
Atkinson, D. (1994). Temperature and organism size: a biological law for
ectotherms. Adv. Ecol. Res., 25(25), 1–58.
Azam, F., Fenchel, T., Field, J.G., Gray, J.S., Meyerreil, L.A. &
Thingstad, F. (1983). The ecological role of water-column microbes in
the sea. Marine Ecology-Progress Series, 10, 257–263.
Barnes, C., Maxwell, D., Reuman, D.C. & Jennings, S. (2010). Global
patterns in predator-prey size relationships reveal size dependency of
trophic transfer efficiency. Ecology, 91, 222–232.
Barnes, C., Irigoien, X., De Oliveira, J.A.A., Maxwell, D. & Jennings, S.
(2011). Predicting marine phytoplankton community size structure from
empirical relationships with remotely sensed variables. J. Plankton Res.,
33, 13–24.
Baum, J.K. & Worm, B. (2009). Cascading top-down effects of changing
oceanic predator abundances. J. Anim. Ecol., 78, 699–714.
Becker, C., Feuchtmayr, H., Brepohl, D., Santer, B. & Boersma, M.
(2004). Differential impacts of copepods and cladocerans on lake
seston, and resulting effects on zooplankton growth. Hydrobiologia,
526, 197–207.
Behrenfeld, M.J., O’Malley, R.T., Siegel, D.A., McClain, C.R.,
Sarmiento, J.L., Feldman, G.C. et al. (2006). Climate-driven trends in
contemporary ocean productivity. Nature, 444, 752–755.
Boyce, D., Lewis, M. & Worm, B. (2010). Global phytoplankton decline
over the past century. Nature, 466, 591–596.
Boyce, D.G., Dowd, M., Lewis, M.R. & Worm, B. (2014). Estimating
global chlorophyll changes over the past century. Prog. Oceanogr., 122,
163–173.
Bremigan, M.T. & Stein, R.A. (1994). Gape-dependent larval foraging
and zooplankton size: implications for fish recruitment across systems.
Can. J. Fish Aquat. Sci., 51, 913–922.
Brewin, R.J.W., Sathyendranath, S., Hirata, T., Lavender, S.J., Barciela,
R.M. & Hardman-Mountford, N.J. (2010). A three-component model
of phytoplankton size class for the Atlantic Ocean. Ecol. Model., 221,
1472–1483.
Brose, U., Jonsson, T., Berlow, E.L., Warren, P., Banasek-Richter, C.,
Bersier, L.F. et al. (2006). Consumer-resource body-size relationships in
natural food webs. Ecology, 87, 2411–2417.
Buchary, E.A. (2001). Preliminary reconstruction of the Icelandic marine
ecosystem in 1950 and some predictions with time series data: models
and analyses. Fish. Cent. Res. Rep., 9(296), 306.
Burns, C.W.W. (1968). Relationship between body size of filter-feeding
cladocera and the maximum size of particle ingested. Limnol.
Oceanogr., 13, 675–678.
Calbet, A. & Landry, M.R. (2004). Phytoplankton growth,
microzooplankton grazing, and carbon cycling in marine systems.
Limnol. Oceanogr., 49, 51–57.
Carpenter, S.R. & Kitchell, J.F. (1993). The Trophic Cascade in Lakes.
Cambridge University Press, Cambridge.
Cebrian, J. & Cebrain, J. (1999). Patterns in the fate of production in
plant communities. Am. Nat., 154, 449–468.
Size-based structuring of plankton ecosystems 11
Cohen, J.E. & Newman, C.M. (1985). A stochastic theory of community
food webs: I. models and aggregated data. Proc. R. Soc. Lond., B, Biol.
Sci., 224, 421–448.
Cohen, J.E., Pimm, S.L., Yodzis, P. & Saldana, J. (1993). Boyd sizes of
animal predators and their animal prey in food webs. J. Anim. Ecol.,
62, 67–78.
Daskalov, G.M., Grishin, A.N., Rodionov, S. & Mihneva, V. (2007).
Trophic cascades triggered by overfishing reveal possible mechanisms of
ecosystem regime shifts. Proc. Natl. Acad. Sci. USA, 104, 10518–10523.
Daufresne, M., Lengfellner, K. & Sommer, U. (2009). Global warming
benefits the small in aquatic ecosystems. Proceedings of the National
Academy of Sciences of the United States of America, 106(31), 12788–
12793. doi:10.1073/pnas.0902080106
DeMott, W.R. (1988). Discrimination between algae and artificial
particles by fresh-water and marine copepods. Limnol. Oceanogr., 33,
397–408.
Di Lorenzo, E., Mountain, D., Batchelder, H.P., Bond, N. & Hoffmann,
E.E. (2013). Advances in marine ecosystem dynamics from US
GLOBEC: the horizontal-advection bottom-up forcing paradigm.
Oceanography, 26, 22–33.
Dippner, J.W. & Krause, M. (2013). Continuous plankton recorder
underestimates zooplankton abundance. J. Mar. Syst., 111–112, 263–268.
Elton, C.S. (1927). Animal Ecology. MacMillan, New York, NY.
Falkowski, P. (2012). The power of plankton. Nature, 483, S17–S20.
Frank, K.T., Petrie, B., Shackell, N.L. & Choi, J.S. (2006). Reconciling
differences in trophic control in mid-latitude marine ecosystems. Ecol.
Lett., 9, 1096–1105.
Fuchs, H. & Franks, P. (2010). Plankton community properties
determined by nutrients and size-selective feeding. Mar. Ecol. Prog.
Ser., 413, 1–15.
Hansen, B., Bjornsen, P.K. & Hansen, P.J. (1994). The size ratio between
planktonic predators and their prey. Limnol. Oceanogr., 39, 395–403.
Harfoot, M.B.J., Newbold, T., Tittensor, D.P., Emmott, S., Hutton, J.,
Lyutsarev, V. et al. (2014). Emergent global patterns of ecosystem
structure and function from a mechanistic general ecosystem model.
PLoS Biol., 12, e1001841.
Iverson, R.L. (1990). Control of marine fish production. Limnol.
Oceanogr., 35, 1593–1604.
Jickells, T.D. (1998). Nutrient biogeochemistry of the coastal zone.
Science, 281, 217–222.
Landry, M. & Calbet, A. (2004). Microzooplankton production in the
oceans. ICES J. Mar. Sci., 61, 501–507.
Lewandowska, A.M., Boyce, D.D.G.D., Hofmann, M., Matthiessen, B.,
Sommer, U. & Worm, B. (2014). Effects of sea surface warming on
marine plankton. Ecol. Lett., 17, 614–623.
Li, W.K.W., Rao, D.V.S., Harrison, W.G., Smith, J.C., Cullen, J.J.,
Irwin, B. et al. (1983). Autotrophic picoplankton in the tropical ocean.
Science, 219, 292–295.
Li, W.K.W., McLaughlin, F.A., Lovejoy, C. & Carmack, E.C. (2009).
Smallest algae thrive as the Arctic Ocean freshens. Science, 326, 539.
Lindeman, R.L. (1942). The trophic-dynamic aspect of ecology. Ecology,
23, 399–418.
Margalef, R. (1978). Life-forms of phytoplankton as survival alternatives
in an unstable environment. Oceanol. Acta, 1, 493–509.
Menden-Deuer, S. & Lessard, E.J. (2000). Carbon to volume relationships
for dinoflagellates, diatoms, and other protest plankton. Limnol.
Oceanogr., 45, 569–579.
Micheli, F. (1999). Eutrophication, fisheries, and consumer-resource
dynamics in marine pelagic ecosystems. Science, 285, 1396–1398.
Moloney, C.L., Jarre, A., Kimura, S., Mackas, D.L., Maury, O.,
Murphy, E.J. et al. (2011). Dynamics of marine ecosystems: ecological
processes. In: Marine Ecosystems and Global Change (eds Barange, M.,
Field, J.G., Harris, R.P., Hofmann, E.E. & Perry, I.R.). Oxford
University Press, Oxford, pp. 440–441.
Mora, C., Wei, C.-L., Rollo, A., Amaro, T., Baco, A.R., Billett, D. et al.
(2013). Biotic and human vulnerability to projected changes in ocean
biogeochemistry over the 21st century. PLoS Biol., 11, 1–14.
© 2015 John Wiley & Sons Ltd/CNRS
12 D. G. Boyce, K. T. Frank and W. C. Leggett
Moran, P.A.P. (1953). The statistical analysis of the Canadian lynx cycle,
II. Synchronization and meteorology. Aust. J. Zool., 1, 291–298.
Mor
an, X.A.G., L
opez-Urrutia, A.,
Calvo-Dıaz, A., Li, W.K.W.,
Moran, X.A.G., Lopez-Urrutia, A. et al. (2010). Increasing importance
of small phytoplankton in a warmer ocean. Global Change Biol., 16,
1137–1144.
O’Connor, M.I., Piehler, M.F., Leech, D.M., Anton, A. & Bruno, J.F.
(2009). Warming and resource availability shift food web structure and
metabolism. PLoS Biol., 7, 1–6.
Odum, E.P. & Brewer, R. (1959). Fundamentals of ecology. (G. W.
Barrett, Ed.) (2nd ed.). Philadelphia: Saunders. 546.
Pauly, D. & Palomares, M. (2005). Fishing down marine food web: it is far
more pervasive than we thought. Bulletin of Marine Science., 76, 197–211.
Petchey, O.L., Beckerman, A.P., Riede, J.O. & Warren, P.H. (2008). Size,
foraging, and food web structure. Proc. Natl. Acad. Sci. USA, 105,
4191–4196.
Polovina, J.J. & Woodworth, P.A. (2012). Declines in phytoplankton cell
size in the subtropical oceans estimated from satellite remotely-sensed
temperature and chlorophyll, 1998–2007. Deep Sea Res. Part II Top.
Stud. Oceanogr., 77–80, 82–88.
Polovina, J.J., Howell, E.A. & Abecassis, M. (2008). Ocean’s least
productive waters are expanding. Geophys. Res. Lett., 35, L03618.
Polovina, J.J., Dunne, J.P., Woodworth, P.A. & Howell, E.A. (2011).
Projected expansion of the subtropical biome and contraction of the
temperate and equatorial upwelling biomes in the North Pacific under
global warming. ICES J. Mar. Sci., 68, 986–995.
Pomeroy, L.R. (1974). Oceans food web, a changing paradigm.
Bioscience, 24, 499–504.
Post, E., Peterson, R.O., Stenseth, N.C. & McLaren, B.E. (1999).
Ecosystem consequences of wolf behavioural response to climate.
Nature, 401, 905–907.
Reuman, D.C. & Cohen, J.E. (2004). Trophic links’ length and slope in
the Tuesday Lake food web with species’ body mass and numerical
abundance. J. Anim. Ecol., 73, 852–866.
Rodrıguez, J., Tintore, J., Allen, J.T., Blanco, J.M., Gomis, D., Reul, A.
et al. (2001). Mesoscale vertical motion and the size structure of
phytoplankton in the ocean. Nature, 410, 360–363.
Rosenzweig, M.L. (1998). Preston’s ergodic conjecture: the accumulation
of species in space and time. In: Biodiversity dynamics: turnover of
populations, taxa, and communities (eds McKinney, M.L. & Drake,
J.A.). Columbia University Press, New York, pp. 311–348.
Roy, S., Sathyendranath, S., Bouman, H. & Platt, T. (2013). The global
distribution of phytoplankton size spectrum and size classes from their
light-absorption spectra derived from satellite data. Remote Sens.
Environ., 139, 185–197.
Ruhl, H.A., Ellena, J.A. & Smith, K.L. (2008). Connections between
climate, food limitation, and carbon cycling in abyssal sediment
communities. Proc. Natl. Acad. Sci. USA, 105, 17006–17011.
Ryther, J.H. (1969). Photosynthesis and fish production in the sea.
Science, 166, 72–76.
IOCCG (2014). Phytoplankton Functional Types from Space.
Sathyendranath, S. (ed.), Reports of the International Ocean-Colour
Coordinating Group, No. 15, IOCCG, Dartmouth, Canada.
Sheldon, R.W., Sutcliff, W.H. & Prakash, A. (1972). Size distribution of
particles in the ocean. Limnol. Oceanogr., 17, 327–340.
Shurin, J.B., Borer, E.T., Seabloom, E.W., Anderson, K., Blanchette,
C.A., Broitman, B. et al. (2002). A cross-ecosystem comparison of the
strength of trophic cascades. Ecol. Lett., 5, 785–791.
© 2015 John Wiley & Sons Ltd/CNRS
Idea and Perspective
Smith, P.E. (1982). Primary production and the anchovy population in
the Southern California Bight: comparison of time series. Limnol.
Oceanogr., 27, 1–17.
Sommer, U. (2008). Trophic cascades in marine and freshwater plankton.
Int. Rev. Hydrobiol., 93, 506–516.
Sommer, U. & Sommer, F. (2006). Cladocerans versus copepods: the
cause of contrasting top-down controls on freshwater and marine
phytoplankton. Oecologia, 147, 183–194.
Sommer, U., Stibor, H., Katechakis, A., Sommer, F. & Hansen, T.
(2002). Pelagic food web configurations at different levels of nutrient
richness and their implications for the ratio fish production: primary
production. Hydrobiologia, 484, 11–20.
Stibor, H., Vadstein, O., Diehl, S., Gelzleichter, A., Hansen, T., Hantzsche,
F. et al. (2004). Copepods act as a switch between alternative trophic
cascades in marine pelagic food webs. Ecol. Lett., 7, 321–328.
Stockner, J.G. & Anita, N.J. (1986). Alagl picoplankton from marine and
fresh-water ecosystems - a multidiscliplinary perspective. Can. J. Fish
Aquat. Sci., 43, 2472–2503.
Tarran, G.A., Zubkov, M.V., Sleigh, M.A., Burkhill, P.H., Yallop, M. &
Burkill, P.H. (2000). Microbial community structure and standing
stocks in the NE Atlantic in June and July of 1996. Deep-Sea Research
II, 48, 963–985.
Tiselius, P. & Jonsson, P.R. (1990). Foraging behavior of 6 calanoid
copepods- observations and hydrodynamics analysis. Marine EcologyProgress Series, 66, 23–33.
Trebilco, R., Baum, J.K., Salomon, A.K. & Dulvy, N.K. (2013).
Ecosystem ecology: size-based constraints on the pyramids of life.
Trends Ecol. Evol., 28, 423–431.
Warren, P.H. & Lawton, J.H. (1987). Invertebrate predator-prey body
size relationships: an explanation for upper triangular food webs and
patterns in food web structure? Oecologia, 74, 231–235.
Williams, R.J., Anandanadesan, A. & Purves, D. (2010). The probabilistic
niche model reveals the niche structure and role of body size in a
complex food web. PLoS ONE, 5, e12092.
Wirtz, K. (2012). Who is eating whom? Morphology and feeding type
determine the size relation between planktonic predators and their ideal
prey. Mar. Ecol. Prog. Ser., 445, 1–12.
Woodward, G., Ebenman, B., Emmerson, M., Montoya, J.M., Olesen,
J.M., Valido, A. et al. (2005). Body size in ecological networks. Trends
Ecol. Evol., 20, 402–409.
Zubkov, M.V. & Tarran, G.A. (2008). High bacterivory by the smallest
phytoplankton in the North Atlantic Ocean. Nature, 455, 224–226.
SUPPORTING INFORMATION
Additional Supporting Information may be downloaded via
the online version of this article at Wiley Online Library
(www.ecologyletters.com).
Editor, Tim Wootton
Manuscript received 26 January 2015
First decision made 2 March 2015
Manuscript accepted 9 March 2015