ArticlePDF Available

Global biogeochemical impacts of phytoplankton: A trait-based perspective

Journal of Ecology

October 2015
103(6):1384–1396

DOI:10.1111/1365-2745.12438

Authors:

Elena Litchman

Michigan State University

Paula de Tezanos Pinto

Instituto de Botanica Darwinion

Kyle F Edwards

University of Hawaii

Christopher Klausmeier

Michigan State University

Show all 6 authorsHide

Phytoplankton are key players in the global carbon cycle, contributing about half of global primary productivity. Within the phytoplankton, functional groups (characterized by distinct traits) have impacts on other major biogeochemical cycles, such as nitrogen, phosphorus and silica. Changes in phytoplankton community structure, resulting from the unique environmental sensitivities of these groups, may significantly alter elemental cycling from local to global scales. We review key traits that distinguish major phytoplankton functional groups, how they affect biogeochemistry and how the links between community structure and biogeochemical cycles are modelled. Finally, we explore how global environmental change will affect phytoplankton communities, from the traits of individual species to the relative abundance of functional groups, and how that, in turn, may alter biogeochemical cycles. Synthesis . We can increase our mechanistic understanding of the links between the community structure of primary producers and biogeochemistry by focusing on traits determining functional group responses to the environment (response traits) and their biogeochemical functions (effect traits). Identifying trade‐offs including allometric and phylogenetic constraints among traits will help parameterize predictive biogeochemical models, enhancing our ability to anticipate the consequences of global change.

The effects of global change on phytoplankton span nested levels of biological organization influencing the function of phytoplankton, including their contributions to biogeochemical cycles. For example, we can consider the consequences of increased nutrient limitation, at each scale: (a) Collective. Changes in bulk properties of phytoplankton, such as their productivity, can alter energy fluxes and carbon cycling. Increasing stratification can reduce productivity in temperate and tropical oceans. (b) Intergroup. Functional groups within phytoplankton communities can respond differently to shared stressors, altering their relative abundance. Diazotrophs may become more common in nitrogen-limited waters, altering N cycling. (c) Intragroup. Turnover in the identity and abundance of particular species within a group may alter its functioning. Nutrient limitation favours small-celled species, influencing rates of zooplankton predation and carbon export to the deep ocean. (d) Intraspecific. Individual species may change their traits and function as a result of global change, through plasticity or rapid evolution. As with C, this could lead to smaller cell sizes in the case of nutrient limitation.

…

Figures - uploaded by Mridul K. Thomas

Content may be subject to copyright.

Content uploaded by Mridul K. Thomas

Content may be subject to copyright.

SPECIAL FEATURE –ESSAY REVIEW

WHETHER IN LIFE OR IN DEATH: FRESH PERSPECTIVES ON HOW PLANTS AFFECT

BIOGEOCHEMICAL CYCLING

Global biogeochemical impacts of phytoplankton: a

trait-based perspective

Elena Litchman

1,2

*, Paula de Tezanos Pinto

, Kyle F. Edwards

, Christopher A.

Klausmeier

1,5

, Colin T. Kremer

6,7

and Mridul K. Thomas

Kellogg Biological Station, Michigan State University, 3700 E Gull Lake Dr., Hickory Corners, MI 49060, USA;

Department of Integrative Biology, Michigan State University, 288 Farm Lane, East Lansing, MI 48824, USA;

Departamento de Ecolog

ıa, Gen

etica y Evoluci

on, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos

Aires, IEGEBA (CONICET-UBA), C1428EHA, Buenos Aires, Argentina;

Department of Oceanography, University of

Hawaii at Manoa, 1000 Pope Road, Honolulu, HI 96822, USA;

Department of Plant Biology, Michigan State

University, 612 Wilson Road, East Lansing, MI 48824, USA;

Department of Ecology and Evolutionary Biology, Yale

University, P.O. Box 208106, New Haven, CT 06520-8106, USA;

Atmospheric and Oceanic Sciences Program,

Princeton University, 300 Forrestal Road, Sayre Hall, Princeton, NJ 08544, USA; and

Department of Aquatic Ecology,

Eawag: Swiss Federal Institute of Aquatic Science and Technology,

€

Uberlandstrasse 133, 8600 D€

ubendorf,

Switzerland

Summary

1. Phytoplankton are key players in the global carbon cycle, contributing about half of global pri-

mary productivity. Within the phytoplankton, functional groups (characterized by distinct traits) have

impacts on other major biogeochemical cycles, such as nitrogen, phosphorus and silica. Changes in

phytoplankton community structure, resulting from the unique environmental sensitivities of these

groups, may signiﬁcantly alter elemental cycling from local to global scales.

2. We review key traits that distinguish major phytoplankton functional groups, how they affect bio-

geochemistry and how the links between community structure and biogeochemical cycles are modelled.

3. Finally, we explore how global environmental change will affect phytoplankton communities,

from the traits of individual species to the relative abundance of functional groups, and how that, in

turn, may alter biogeochemical cycles.

4. Synthesis. We can increase our mechanistic understanding of the links between the community

structure of primary producers and biogeochemistry by focusing on traits determining functional

group responses to the environment (response traits) and their biogeochemical functions (effect

traits). Identifying trade-offs including allometric and phylogenetic constraints among traits will help

parameterize predictive biogeochemical models, enhancing our ability to anticipate the consequences

of global change.

Key-words: aquatic plant ecology, biogeochemical cycles, cell size, functional groups, global

change, phytoplankton community structure, trade-offs

Introduction

Understanding the links between the structure of primary pro-

ducer communities and biogeochemistry is an important

research frontier bridging community and ecosystem ecology.

Such research is especially important as rapid, human-driven

changes in our environment affect primary producer commu-

nities and, ultimately, global biogeochemical cycles. Phyto-

plankton are major aquatic primary producers, responsible for

about half of global primary productivity each year (Field

et al. 1998). They are key contributors to biogeochemical

cycles, both at present and over the long history of our planet,

*Correspondence author: E-mail: litchman@msu.edu

Journal of Ecology 2015, 103, 1384–1396 doi: 10.1111/1365-2745.12438

and are the subject of extensive experimental, observational

and theoretical attention. Fully appreciating the role of phyto-

plankton in elemental cycling requires characterizing their

diversity that encompasses two domains of life and several

distinct functional groups. These groups differ in how the

environment inﬂuences their growth, in the elements and

their ratios that the groups consume and excrete, and, there-

fore, in their effects on biogeochemistry. We argue that estab-

lishing effective, mechanistic links between community and

ecosystem ecology requires characterizing the traits that deﬁne

these groups and their environmental sensitivities. Ultimately,

this kind of trait-based approach can provide critical insights

into the biogeochemical function of aquatic and terrestrial

communities, both now and in our increasingly human-

impacted world. In this review, we describe major phyto-

plankton groups (including their history, traits and involve-

ment in biogeochemical cycling), how these groups have

been incorporated in biogeochemical models, and discuss

how they may be affected by global change.

What are the advantages of trait-based approaches? Traits

capture aspects of physiology, morphology and life history

that inﬂuence ﬁtness and competitive success. Resource utili-

zation traits connect species abundances and growth with the

chemical compounds that are most limiting for metabolic pro-

cesses, thereby linking ecological processes (species and

community dynamics) with biogeochemical processes through

species performance. For example, the rate at which a cell

can take up phosphorus is an important determinant of its

competitive ability in environments where phosphorus limits

growth. Integrated over the community, these uptake rates

feed back to inﬂuence external phosphorus concentrations. In

other cases, the link between biogeochemistry and ecological

success is less direct, and it may be useful to distinguish

between traits that affect biogeochemical cycles (effect traits)

and the traits that determine how abundances of these groups

would respond to changing conditions (response traits)

(Lavorel & Garnier 2002). Many phytoplankton traits are

both response and effect traits, but many are only one or the

other. For example, the optimal temperature for growth is a

purely response trait (because temperature affects growth but

not vice versa), and carbon export efﬁciency is an effect trait

that is not tightly linked to a particular response trait. Under-

standing how response and effect traits are related will allow

us to understand the links between the community structure

of primary producers and biogeochemistry in a mechanistic

way.

Many groups of primary producers differ both in their

response and effect traits and, consequently, may have differ-

ent impacts on biogeochemical cycling. For example, free liv-

ing or symbiotic nitrogen-ﬁxers can ﬁx atmospheric nitrogen

(N) and do not require other forms of N. Thus, their N

response traits are distinct from those of non-N-ﬁxers. While

non-ﬁxers consume available N, N-ﬁxers can signiﬁcantly

increase N concentration in the environment (soil or water),

thus having contrasting N effect traits as well. The N ﬁxation

response and effect traits are tightly linked, which is true for

other resource acquisition traits.

Different traits are often not independent but connected by

trade-offs, and these trade-offs determine how community

structure and coexistence change under different conditions,

with concomitant changes in biogeochemical processes. N-ﬁx-

ers are often thought to have high phosphorus requirements

compared to other groups and thus exhibit a trade-off between

N and P competitive abilities (Lenton & Klausmeier 2007).

This trade-off can lead to coexistence of N-ﬁxers and non-ﬁx-

ers and can explain shifts in community composition with

changing N: P ratios. Similar trade-offs between N ﬁxation

and competitive abilities for light or iron may also be impor-

tant (Agawin et al. 2007; de Tezanos Pinto & Litchman

2010; Ward et al. 2013). Whether N-ﬁxers occur under par-

ticular conditions has many biogeochemical consequences,

because it determines how much ‘new’nitrogen is being

added to the ocean, because it affects the stoichiometry of

organic matter exported to the deep ocean and because ﬁxed

N can be released into the environment and consumed by

other primary producers and other microbes.

In this review, we focus on biogeochemically relevant traits

and trade-offs to discuss what is known about the feedbacks

between phytoplankton community structure and biogeochem-

istry, with the hope that some of these ideas can also be

applied to terrestrial plants. We begin by reviewing the dis-

tinct biogeochemical signatures and impacts of the major tax-

onomic groups of phytoplankton. Next, we discuss modelling

approaches that link phytoplankton community structure and

key traits to biogeochemical cycles. Finally, we describe

major global change stressors in aquatic environments and

discuss how they can change phytoplankton communities,

selecting for groups and species with certain traits, with con-

sequences for biogeochemistry.

Major phytoplankton groups and their

biogeochemical signatures

Phytoplankton are a paraphyletic group of photoautotrophs

with a complex evolutionary history extending across

2.5–3.5 billion years (Olson & Blankenship 2004; Yoon et al.

2004). Despite this paraphyly, they fall into evolutionarily

distinct functional groups, including one major prokaryotic

group (the cyanobacteria) and a number of eukaryotic groups

(diatoms, green algae, coccolithophorids, dinoﬂagellates and

others).

The evolution of oxygenic photosynthesis in cyanobacteria

was a major advance in metabolic strategy that changed the

fate of our planet forever. This major evolutionary innovation

likely happened only once and changed the atmosphere of our

planet from anoxic to oxygenic after two billion years in the

former state (Knoll 2003). Vast amounts of solar energy were

subsequently channelled into driving geochemical cycles

(Rosing et al. 2006), reducing atmospheric carbon dioxide to

a wide range of organic substances (Holm

en 1992). Signiﬁcant

quantities of molecular oxygen started to accumulate in the

atmosphere about 2.2–2.4 billion years ago (Catling & Zahnle

2002; Holland 2006). The timing of Earth’s oxygenation coin-

cided with the disappearance of large non-mass-dependent

Global biogeochemical impacts of phytoplankton 1385

sulphur isotope (

S and

S) fractionations (Holland 2006)

and of iron banded formations (Isley & Abbott 1999), showing

the impacts of oxygenation on other element cycles.

This process of oxygenation was largely driven by the cy-

anobacteria, which remained the dominant phytoplankton for

well over a billion years and strongly inﬂuence elemental

cycles today. They are the only phytoplanktonic group with

members capable of atmospheric nitrogen ﬁxation (the

absence of N ﬁxation in eukaryotes is poorly understood,

Schopf 1983) that has a major impact on global nitrogen

cycle. Nitrogen ﬁxation is one of the most metabolically

expensive processes in biology (16 ATPs are hydrolysed per

ﬁxed) (Simpson & Burris 1984) and is crucial for Earth’s

nitrogen budget and primary productivity. The evolution of

the ability to ﬁx nitrogen is thought to be extremely ancient

(Staley & Orians 1992), possibly older than oxygenic photo-

synthesis. Several authors infer that the nitrogenase family

had already evolved in the last common ancestor of extant

organisms (Normand et al. 1992; Fani, Gallo & Lio 2000),

though another hypothesis suggests a later origin (see Ray-

mond et al. 2004). Because nitrogen ﬁxation is a process

highly sensitive to the presence of oxygen, the evolution of

oxygenic photosynthesis posed a major physiological burden

on nitrogen ﬁxation in cyanobacteria. This limitation was

overcome by ﬁxing nitrogen during the night (in the absence

of oxygen production), hence separating N ﬁxation and pho-

tosynthesis in time. Another strategy was the separation of

photosynthesis and N ﬁxation in space, by performing N ﬁxa-

tion in anaerobic specialized cells, heterocysts, during day-

light and photosynthesis in vegetative cells. There could still

be other undiscovered strategies.

Marine cyanobacteria are responsible for 25–50% of

natural (i.e. not anthropogenic via the Haber–Bosch process)

global nitrogen ﬁxation, or approximately 4.5–99

mol N year

1

(Mahaffey, Michaels & Capone 2005;

Canﬁeld, Glazer & Falkowski 2010; Zehr 2011). This spatial

extent of nitrogen ﬁxation is limited both by temperature (Sta-

al, Meysman & Stal 2003) and by the availability of iron,

which is needed for the production of the nitrogenase enzyme

(Berman-Frank et al. 2001; Kustka et al. 2003). Conse-

quently, nitrogen ﬁxation occurs most strongly in the tropical

and subtropical oceans, and in regions with high Fe: N ratios

(Monteiro, Dutkiewicz & Follows 2011; Ward et al. 2013).

The process of nitrogen ﬁxation is leaky; between 50 and

90% of ﬁxed nitrogen may be released to surrounding waters

in the form of dissolved organic nitrogen (Glibert & Bronk

1994; Mulholland & Bernhardt 2005), thereby subsidizing

non-ﬁxers. The cyanobacteria possess the highest N: P ratio

of any phytoplankton group, with an average N: P ratio of

around 22:1, relative to 13:1 in eukaryotic taxa (data com-

piled from earlier studies by Deutsch & Weber 2012). This is

considerably higher in the nitrogen-ﬁxing Trichodesmium sp.,

which may be >40:1 due to the high nitrogen requirement for

light-harvesting machinery (Letelier & Karl 1996; data com-

piled from earlier studies by Klausmeier et al. 2004).

Although nitrogen ﬁxation is a unique trait of cyanobacteria,

not all cyanobacteria ﬁx nitrogen. Among those that do not,

Prochlorococcus and Synechococcus play an important role in

the carbon cycle, because they are major components of the

photosynthetic biomass in the oligotrophic oceans (Johnson

et al. 2006).

The eukaryotic taxa originated with the engulfment of a

cyanobacterium by a heterotrophic protozoan, leading to the

formation of a symbiotic relationship between the two,

approximately 1.5 billion years ago (Yoon et al. 2004). The

endosymbiotic cyanobacterium subsequently evolved into the

contemporary chloroplast. Thereafter, this eukaryotic clade

diverged into green and red lineages that differ based on their

pigment composition (Delwiche 1999). Subsequent engulf-

ment of these green and red eukaryotes by heterotrophs led to

secondary and even tertiary endosymbioses: diatoms, for

example, are the product of an endosymbiotic event between

a heterotroph and a red alga (Archibald & Keeling 2002). As

a consequence, contemporary eukaryotic phytoplankton are

composed of two superfamilies. The green superfamily con-

tains the green algae and groups formed by the engulfment of

a green alga by a heterotroph, such as the euglenophytes. The

red superfamily contains the red algae (now largely benthic)

and groups containing red plastids, including diatoms, crypto-

phytes and coccolithophores. While both green and red super-

families primarily use chlorophyll a, the former also

possesses chlorophyll b, while the latter uses chlorophyll c

and several accessory pigments that absorb blue and green

wavelengths (Falkowski et al. 2004; Katz et al. 2004).

The eukaryotic superfamilies exhibit distinct stoichiometries

from each other and the cyanobacteria. Members of the green

superfamily possess higher C: P and N: P ratios than their

counterparts in the red superfamily as well as higher require-

ments for some micronutrients (Fe, Cu and Zn) and lower

requirements for others (Cd, Co and Mn; Quigg et al. 2003).

These differences in stoichiometry are thought to be related to

oceanic environmental conditions when these groups diversi-

ﬁed, and have important implications for biogeographic pat-

terns in ocean N: P ratios, which they both inﬂuence and are

inﬂuenced by Weber & Deutsch (2010). Because functional

groups also differ in their response to other environmental

variables (characterized by their traits), environment-driven

shifts in functional group composition are likely to drive

changes in biogeochemical cycles due to shifts in the average

cell stoichiometry. We discuss a few of the major eukaryotic

groups here, but note that other less studied groups may be of

considerable biogeochemical importance, though we lack data

with which to draw broad conclusions. These include the

non-calcifying haptophytes, chrysophytes, cryptophytes, raphi-

dophytes, rhodophytes and euglenophytes, among others.

Green algae, the group that gave rise to terrestrial plants,

evolved >1 billion years ago and comprised a large propor-

tion of the phytoplankton community till the Mesozoic era,

approximately 250 million years ago. Thereafter, they

declined in abundance and diversity in the oceans, being

apparently out-competed by members of the red superfamily

groups that began to rise in prominence (Falkowski et al.

2004; Katz et al. 2004). This decline has been linked to

changes in the redox state of the oceans, with their high trace

1386 E. Litchman et al.

metal requirements possibly constraining their present marine

distribution and diversity (they remain abundant, diverse and

broadly distributed in freshwaters). Unlike the other groups

described here, they perform no distinctive biogeochemical

functions to our knowledge. They are thought to be excep-

tionally plastic in their biomass N: P ratio, but this is based

largely on experiments with a single freshwater species,

Scenedesmus sp. (Rhee 1978).

Coccolithophorids ﬁrst occur in the fossil record approxi-

mately 220 million years ago and rose to prominence through

the Mesozoic era and then declined in diversity as the dia-

toms diversiﬁed (Bown, Lees & Young 2004; Falkowski

et al. 2004; Katz et al. 2004). They form a major portion of

high latitude phytoplankton communities today, with massive

blooms of one major species, Emiliania huxleyi, occurring

over areas of hundreds of thousands of square kilometres

(Brown & Podest

a 1997). They play a dominant role among

the phytoplankton in the marine calcium cycle through the

formation of calcium carbonate plates (coccoliths), which are

responsible for the formation of calcium carbonate rock for-

mations across the globe. Although coccolith sinking may be

expected to lead to considerable carbon export, the extent of

the export is a function of sea ﬂoor depth and calcium com-

pensation depth. If the ocean ﬂoor is lower than the compen-

sation depth, coccoliths dissolve, returning carbon to the

water. In total, they are thought to be responsible for approxi-

mately 10% of carbon export to the deep ocean (Jin et al.

2006). At the same time, calciﬁcation reduces total alkalinity,

reducing the ability of the surface ocean to take up atmo-

spheric CO

(Passow and Carlson 2012). Coccolithophores

also produce dimethylsulfoniopropionate (DMSP), a precursor

to dimethyl sulphide (DMS), a cloud condensation nucleus.

Though there is a clear mechanistic link between coccolitho-

phorid growth and climate in this case, the evidence for it as

a regulator of climate is weak, though it may still play a

minor role (Charlson et al. 1987; Ayers & Cainey 2007;

Quinn & Bates 2011; Rap et al. 2013).

Diatoms are believed to have originated 160–200 million

years ago and diversiﬁed strongly in the past 60 million

years, at the expense of other groups (Kooistra & Medlin

1996; Falkowski et al. 2004; Sims, Mann & Medlin 2006).

Their recent dominance may be a result of their ability to

store pulses of nutrients in a large vacuole supported by the

silica cell wall and/or protection from predation accorded by

their silica shells (Smetacek 1999; Thingstad et al. 2005;

Litchman, Klausmeier & Yoshiyama 2009). As other func-

tional groups do not use silica, the ability to compete for this

element does not contribute to their success against other

functional groups, but may contribute towards determining

which diatom species predominate. Globally, they play a

dominant role in the carbon cycle, contributing an estimated

20–25% of global primary productivity (Nelson et al. 1995;

Smetacek 1999). Their relatively large cell sizes lead to a

high sinking rate which contributes to major portion of carbon

export to the deep oceans (Nelson et al. 1995). They are also

the primary phytoplankton group associated with the global

silica cycle (with silicoﬂagellates playing a minor role). This

uptake of silica is responsible for the undersaturation of silica

in the surface oceans and leads to the burial of

6.3 910

mol Si year

1

through sinking (Tr

eguer & De La

Rocha 2013). They have low N: P ratios characteristic of the

red superfamily, with an average of approximately 10:1

(Sarthou et al. 2005), and are most abundant in unstable,

nutrient-rich waters with similarly low N:P ratios, typically at

high latitudes and in coastal oceans (Arrigo 2005).

Dinoﬂagellates belong to the red superfamily and became a

major component of the phytoplankton community over

200 million years ago, but have decreased in diversity over

the past 40–60 million years (Katz et al. 2004). They exhibit

exceptionally complex genomes and metabolic capabilities:

members of this motile group exhibit mixotrophy (and even

pure heterotrophy), toxin production, bioluminescence and

tertiary endosymbiosis with complex symbionts such as dia-

toms and prasinophytes (Wisecaver & Hackett 2011). Despite

belonging to the red superfamily, they also contain members

that have appropriated plastids belonging to the green lineage

(Falkowski et al. 2004). They possess among the largest phy-

toplankton cells and are typically poor competitors for nutri-

ents and slow growers under autotrophic conditions

(Litchman et al. 2007); mixotrophy allows them to persist

despite this inefﬁciency, particularly in environments with

high nutrient and organic matter concentrations.

Biogeochemical models of phytoplankton

functional types

Biogeochemical models aim to explain how biogeochemical

cycles operate today and operated in the past, and predict

how they will respond to anthropogenic stressors such as cli-

mate change. Much of the foundational and current work

on marine ecosystem models has used a NPZ (nutrient–

phytoplankton–zooplankton) structure, where phytoplankton

are modelled as a single population that consumes a single

limiting nutrient (typically representing nitrogen) and are con-

sumed by a single zooplankton population (Evans and Par-

slow 1985, Fasham et al. 1990, Franks 2002). NPZ models

are formulated and parameterized by combining laboratory

data on plankton traits, ﬁeld observations of bulk stocks and

rates, and theoretical considerations such as dynamical stabil-

ity. Although they greatly simplify biological complexity,

NPZ models are useful for theoretical exploration and have

successfully represented ecosystem patterns in many situations

(Franks 2002). More complex models have been developed to

address questions such as the role of multiple limiting nutri-

ents and their biogeochemical coupling, because iron, phos-

phorus and silicon may all limit phytoplankton growth in

addition to nitrogen (Flynn 2003, Moore et al. 2013).

It has also become clear that incorporating the functional

diversity of phytoplankton is important for modelling biogeo-

chemical cycles, because community structure varies greatly

in time and space, and different kinds of phytoplankton have

distinct effects on carbon ﬁxation and export, as well as the

cycling of N and Si among other elements (Le Qu

ere et al.

2005, Hood et al. 2006). The most common way biogeo-

Global biogeochemical impacts of phytoplankton 1387

metal requirements possibly constraining their present marine

distribution and diversity (they remain abundant, diverse and

broadly distributed in freshwaters). Unlike the other groups

described here, they perform no distinctive biogeochemical

functions to our knowledge. They are thought to be excep-

tionally plastic in their biomass N: P ratio, but this is based

largely on experiments with a single freshwater species,

Scenedesmus sp. (Rhee 1978).

Coccolithophorids ﬁrst occur in the fossil record approxi-

mately 220 million years ago and rose to prominence through

the Mesozoic era and then declined in diversity as the dia-

toms diversiﬁed (Bown, Lees & Young 2004; Falkowski

et al. 2004; Katz et al. 2004). They form a major portion of

high latitude phytoplankton communities today, with massive

blooms of one major species, Emiliania huxleyi, occurring

over areas of hundreds of thousands of square kilometres

(Brown & Podest

a 1997). They play a dominant role among

the phytoplankton in the marine calcium cycle through the

formation of calcium carbonate plates (coccoliths), which are

responsible for the formation of calcium carbonate rock for-

mations across the globe. Although coccolith sinking may be

expected to lead to considerable carbon export, the extent of

the export is a function of sea ﬂoor depth and calcium com-

pensation depth. If the ocean ﬂoor is lower than the compen-

sation depth, coccoliths dissolve, returning carbon to the

water. In total, they are thought to be responsible for approxi-

mately 10% of carbon export to the deep ocean (Jin et al.

2006). At the same time, calciﬁcation reduces total alkalinity,

reducing the ability of the surface ocean to take up atmo-

spheric CO

(Passow and Carlson 2012). Coccolithophores

also produce dimethylsulfoniopropionate (DMSP), a precursor

to dimethyl sulphide (DMS), a cloud condensation nucleus.

Though there is a clear mechanistic link between coccolitho-

phorid growth and climate in this case, the evidence for it as

a regulator of climate is weak, though it may still play a

minor role (Charlson et al. 1987; Ayers & Cainey 2007;

Quinn & Bates 2011; Rap et al. 2013).

Diatoms are believed to have originated 160–200 million

years ago and diversiﬁed strongly in the past 60 million

years, at the expense of other groups (Kooistra & Medlin

1996; Falkowski et al. 2004; Sims, Mann & Medlin 2006).

Their recent dominance may be a result of their ability to

store pulses of nutrients in a large vacuole supported by the

silica cell wall and/or protection from predation accorded by

their silica shells (Smetacek 1999; Thingstad et al. 2005;

Litchman, Klausmeier & Yoshiyama 2009). As other func-

tional groups do not use silica, the ability to compete for this

element does not contribute to their success against other

functional groups, but may contribute towards determining

which diatom species predominate. Globally, they play a

dominant role in the carbon cycle, contributing an estimated

20–25% of global primary productivity (Nelson et al. 1995;

Smetacek 1999). Their relatively large cell sizes lead to a

high sinking rate which contributes to major portion of carbon

export to the deep oceans (Nelson et al. 1995). They are also

the primary phytoplankton group associated with the global

silica cycle (with silicoﬂagellates playing a minor role). This

uptake of silica is responsible for the undersaturation of silica

in the surface oceans and leads to the burial of

6.3 910

mol Si year

1

through sinking (Tr

eguer & De La

Rocha 2013). They have low N: P ratios characteristic of the

red superfamily, with an average of approximately 10:1

(Sarthou et al. 2005), and are most abundant in unstable,

nutrient-rich waters with similarly low N:P ratios, typically at

high latitudes and in coastal oceans (Arrigo 2005).

Dinoﬂagellates belong to the red superfamily and became a

major component of the phytoplankton community over

200 million years ago, but have decreased in diversity over

the past 40–60 million years (Katz et al. 2004). They exhibit

exceptionally complex genomes and metabolic capabilities:

members of this motile group exhibit mixotrophy (and even

pure heterotrophy), toxin production, bioluminescence and

tertiary endosymbiosis with complex symbionts such as dia-

toms and prasinophytes (Wisecaver & Hackett 2011). Despite

belonging to the red superfamily, they also contain members

that have appropriated plastids belonging to the green lineage

(Falkowski et al. 2004). They possess among the largest phy-

toplankton cells and are typically poor competitors for nutri-

ents and slow growers under autotrophic conditions

(Litchman et al. 2007); mixotrophy allows them to persist

despite this inefﬁciency, particularly in environments with

high nutrient and organic matter concentrations.

Biogeochemical models of phytoplankton

functional types

Biogeochemical models aim to explain how biogeochemical

cycles operate today and operated in the past, and predict

how they will respond to anthropogenic stressors such as cli-

mate change. Much of the foundational and current work

on marine ecosystem models has used a NPZ (nutrient–

phytoplankton–zooplankton) structure, where phytoplankton

are modelled as a single population that consumes a single

limiting nutrient (typically representing nitrogen) and are con-

sumed by a single zooplankton population (Evans and Par-

slow 1985, Fasham et al. 1990, Franks 2002). NPZ models

are formulated and parameterized by combining laboratory

data on plankton traits, ﬁeld observations of bulk stocks and

rates, and theoretical considerations such as dynamical stabil-

ity. Although they greatly simplify biological complexity,

NPZ models are useful for theoretical exploration and have

successfully represented ecosystem patterns in many situations

(Franks 2002). More complex models have been developed to

address questions such as the role of multiple limiting nutri-

ents and their biogeochemical coupling, because iron, phos-

phorus and silicon may all limit phytoplankton growth in

addition to nitrogen (Flynn 2003, Moore et al. 2013).

It has also become clear that incorporating the functional

diversity of phytoplankton is important for modelling biogeo-

chemical cycles, because community structure varies greatly

in time and space, and different kinds of phytoplankton have

distinct effects on carbon ﬁxation and export, as well as the

cycling of N and Si among other elements (Le Qu

ere et al.

2005, Hood et al. 2006). The most common way biogeo-

Global biogeochemical impacts of phytoplankton 1387

nutrient limitation term, and this can cause their regional pat-

terns of growth limitation to be distinct from other phyto-

plankton types, due to variation in the supply of Si relative to

other elements (Moore et al. 2002; Aumont et al. 2003;

Blackford, Allen & Gilbert 2004). They are also thought to

contribute disproportionately to carbon export from the

euphotic zone, due to a number of potential factors: their lar-

ger size (on average), the boom-and-bust phenology of some

species, the ‘ballasting’effect of the silica frustule and

reduced grazing rates (Smetacek 1999; Assmy et al. 2013).

Some of these effects will emerge naturally if the functional

type representing diatoms is given a higher maximum growth

rate or lower grazing rate, as described above. The increase in

export due to larger size or ballasting can be modelled with

an increased sinking rate of non-grazed and/or grazed biomass

(Moore et al. 2002; Litchman et al. 2006; Follows et al.

2007).

Coccolithophores play a distinct biogeochemical role due to

calciﬁcation and ballasting of organic matter and have a dis-

tinct (and likely varied) ecology that is still being unravelled

(Boyd et al. 2010). Their traits have been modelled in a vari-

ety of ways, although a general aim is to reproduce the fact

that they bloom under more stratiﬁed conditions than diatoms,

but are likely poorer nutrient competitors than picophyto-

plankton. Gregg & Casey (2007) give coccolithophores a

half-saturation constant for nitrogen-limited growth that is

higher than that of cyanobacteria but lower than that of dia-

toms, as well as an intermediate maximum growth rate (which

is highest for diatoms in this case) and a relatively high irra-

diance requirement. In this formulation, they essentially have

an intermediate position on a gleaner–opportunist trade-off

axis (Margalef 1978; Kudela 2010; Edwards, Klausmeier &

Litchman 2013). In contrast, the PlankTOM10 model gives

coccolithophores a relatively slow maximum growth rate,

intermediate nutrient afﬁnity traits and a minimum tempera-

ture threshold that is lower than that of the picocyanobacteria

that are the best nutrient competitors (Kwiatkowski et al.

2014, http://lgmacweb.env.uea.ac.uk/green_ocean/model/

code_description/PFT/ﬂuxes.html). An increased efﬁciency of

carbon export for coccolithophores, due to ballasting, can be

implemented with a higher sinking rate (Moore et al. 2002;

Gregg & Casey 2007).

Diazotrophs (nitrogen-ﬁxers) are an important functional

type, due to their unique role in the nitrogen cycle. They

are thought to prosper under extreme nitrogen limitation,

while suffering a number of disadvantages described earlier.

They are often modelled as having a relatively high iron

requirement, which makes them poor competitors when iron

supply is low relative to N and P (Moore et al. 2002; Ward

et al. 2013). They are also typically given a low maximum

growth rate, which means they cannot take advantage of

variable nutrient supply (Tyrell 1999, Moore et al. 2002;

Monteiro, Dutkiewicz & Follows 2011). It may also be

important for diazotrophs to be poor P competitors (Tyrrell

1999; Lenton & Klausmeier 2007; Ward et al. 2013),

although a relatively low P stoichiometry will at least par-

tially offset this (Moore et al. 2002). In combination, these

traits will lead to diazotroph occurrence in stably stratiﬁed

regions with a low supply ratio of N: Fe and N: P (Ward

et al. 2013). Although most ocean biogeochemical models

represent diazotrophs by a single group, some models

include different diazotroph types with distinct traits, such as

ﬁlamentous Trichodesmium, unicellular N-ﬁxers and symbi-

otic N-ﬁxers living inside diatoms (e.g. Monteiro, Follows

& Dutkiewicz 2010).

Responses of phytoplankton communities to

global environmental change and implications

for biogeochemistry

The many dimensions of global change are increasingly affect-

ing marine environments and the phytoplankton communities

that inhabit them. In the following section, we brieﬂy summa-

rize the most important aspects of the ocean that are changing,

including temperature, acidity and nutrient availability. The

traits and functions of phytoplankton groups and species deter-

mine both how they are affected by global change and how

their responses in turn alter global biogeochemistry. In the

remaining text, we discuss how global change stressors inﬂu-

ence phytoplankton across different levels of biological orga-

nization from functional groups to species.

Global change is multifaceted and includes not just CO

emissions and climate change, but also anthropogenic effects

that are independent of these factors. Acidiﬁcation, rising tem-

peratures and changes in the supply of nutrients and light will

pose the greatest challenges for marine ecosystems and phyto-

plankton communities. Several reviews have examined these

effects in greater detail (e.g. Hoegh-Guldberg & Bruno 2010;

Boyd and Hutchins 2012, Doney et al. 2012), so our summary

is brief. World-wide, oceans act as a major sink for both the

rising levels of atmospheric CO

and the increased amount of

heat trapped by this CO

and other greenhouse gasses. Nearly

1/3 of anthropogenic CO

is being absorbed into the oceans,

where it interacts with water, altering seawater carbonate

chemistry and driving ocean acidiﬁcation (Riebesell 2004; Do-

ney et al. 2009; Hoegh-Guldberg & Bruno 2010). Tempera-

tures have risen by ~0.6 °C over the last 100 years within the

surface layers of the ocean (Hoegh-Guldberg & Bruno 2010)

and are predicted to increase by another 1–3°C by the end of

the century (IPCC; Collins et al. 2013). Temperature plays an

important role in physical ocean processes, driving mixing,

vertical stratiﬁcation and currents. In tropical and temperate

regions, warmer temperatures lead to stronger stratiﬁcation

and shallower mixed layers, simultaneously intensifying nutri-

ent limitation and reducing light limitation (as plankton avoid

being mixed to deeper, darker regions) (Beardall, Stojkovic &

Larsen 2009). Regionally, warming may strengthen coastal

upwelling and nutrient supply and alter the depth of mixing in

polar oceans (Hoegh-Guldberg & Bruno 2010). Temperature

and pH also inﬂuence the chemistry of seawater; changes in

both of these variables will alter the solubility and oxidation

state of trace metals such as iron in ways that are not yet fully

understood (Hoffmann et al. 2012). These changes will be sig-

niﬁcant for marine organisms requiring these trace metals

Global biogeochemical impacts of phytoplankton 1389

often as catalysts for critical biochemical processes. The sup-

ply of more abundant (though no less important) nutrients

such as nitrogen and phosphate is also being altered by human

activities. This occurs most dramatically in coastal regions,

through the inﬂux of agricultural run-off and sediments from

rivers, but also includes atmospheric deposition of dust and

combustion by-products. Ultimately, these nutrients affect

microbial communities producing areas of hypoxia, especially

in coastal regions (Diaz & Rosenberg 2008), and altering bio-

geochemical processes. Finally, climate-driven changes in

weather patterns and cloud cover, among other factors, alter

the amount of light (and especially ultraviolet radiation) reach-

ing and penetrating the ocean surface (Beardall, Stojkovic and

Larsen 2009).

Adding to the complexity of global change is the fact that

few environments will experience only one of these changing

stressors: marine organisms may simultaneously face rising

temperatures, changing nutrient levels and acidiﬁcation (Halp-

ern et al. 2008). The combined effects of multiple stressors

can be non-additive and nonlinear, so global change studies

for most environments or species must consider the effects of

a suite of stressors (Crain, Kroeker & Halpern 2008; Boyd

2013; Boyd et al. 2015). Many aspects of global change also

exhibit signiﬁcant regional variability on top of underlying

global trends (Hansen et al. 2006).

As complex as global change is, understanding its effects

on phytoplankton and marine ecosystems presents an even

greater challenge, because these ecological systems, composed

of diverse and interacting species, are exceedingly complex in

their own right. We can organize this complexity by consider-

ing how global change affects phytoplankton at different eco-

logical resolutions, ranging from the collective response of the

phytoplankton as a whole to the ﬁne-scale reactions of indi-

vidual species (Fig. 1). These categories roughly correspond

to the level of detail included in different biogeochemical

models of phytoplankton communities as discussed in ‘bio-

geochemical models of phytoplankton functional types’,

which range from as coarse as one or a few functional

groups, to dozens of species. Determining the appropriate

level of detail to resolve is a critical open research question,

which likely depends on which ecological or biogeochemical

features are of interest, and the degree of precision required.

The following discussion examines each of these categories in

turn, starting with the broadest and most simpliﬁed view.

At the coarsest scale, brieﬂy setting aside the particulars of

species or functional group identity, photoautotrophs such as

phytoplankton are united in their requirement for light, nutri-

ents and CO

to support their growth and productivity. If basic

requirements are either not met or dramatically exceeded (e.g.

photoinhibition or nutrient toxicity), whether through regional

variation or global change, growth and productivity are limited

(Fig. 1a). Modelling studies that include a generic phytoplank-

ton component, with parameters based on these requirements,

improve biogeochemistry models and can reproduce empirical

patterns of nutrient distributions (Kriest, Khatiwala & Oschlies

2010). They may also be able to crudely predict global

Fig. 1. The effects of global change on phytoplankton span nested levels of biological organization inﬂuencing the function of phytoplankton,

including their contributions to biogeochemical cycles. For example, we can consider the consequences of increased nutrient limitation, at each

scale: (a) Collective. Changes in bulk properties of phytoplankton, such as their productivity, can alter energy ﬂuxes and carbon cycling. Increas-

ing stratiﬁcation can reduce productivity in temperate and tropical oceans. (b) Intergroup. Functional groups within phytoplankton communities

can respond differently to shared stressors, altering their relative abundance. Diazotrophs may become more common in nitrogen-limited waters,

altering N cycling. (c) Intragroup. Turnover in the identity and abundance of particular species within a group may alter its functioning. Nutrient

limitation favours small-celled species, inﬂuencing rates of zooplankton predation and carbon export to the deep ocean. (d) Intraspeciﬁc. Individ-

ual species may change their traits and function as a result of global change, through plasticity or rapid evolution. As with C, this could lead to

smaller cell sizes in the case of nutrient limitation.

1390 E. Litchman et al.

nutrient limitation term, and this can cause their regional pat-

terns of growth limitation to be distinct from other phyto-

plankton types, due to variation in the supply of Si relative to

other elements (Moore et al. 2002; Aumont et al. 2003;

Blackford, Allen & Gilbert 2004). They are also thought to

contribute disproportionately to carbon export from the

euphotic zone, due to a number of potential factors: their lar-

ger size (on average), the boom-and-bust phenology of some

species, the ‘ballasting’effect of the silica frustule and

reduced grazing rates (Smetacek 1999; Assmy et al. 2013).

Some of these effects will emerge naturally if the functional

type representing diatoms is given a higher maximum growth

rate or lower grazing rate, as described above. The increase in

export due to larger size or ballasting can be modelled with

an increased sinking rate of non-grazed and/or grazed biomass

(Moore et al. 2002; Litchman et al. 2006; Follows et al.

2007).

Coccolithophores play a distinct biogeochemical role due to

calciﬁcation and ballasting of organic matter and have a dis-

tinct (and likely varied) ecology that is still being unravelled

(Boyd et al. 2010). Their traits have been modelled in a vari-

ety of ways, although a general aim is to reproduce the fact

that they bloom under more stratiﬁed conditions than diatoms,

but are likely poorer nutrient competitors than picophyto-

plankton. Gregg & Casey (2007) give coccolithophores a

half-saturation constant for nitrogen-limited growth that is

higher than that of cyanobacteria but lower than that of dia-

toms, as well as an intermediate maximum growth rate (which

is highest for diatoms in this case) and a relatively high irra-

diance requirement. In this formulation, they essentially have

an intermediate position on a gleaner–opportunist trade-off

axis (Margalef 1978; Kudela 2010; Edwards, Klausmeier &

Litchman 2013). In contrast, the PlankTOM10 model gives

coccolithophores a relatively slow maximum growth rate,

intermediate nutrient afﬁnity traits and a minimum tempera-

ture threshold that is lower than that of the picocyanobacteria

that are the best nutrient competitors (Kwiatkowski et al.

2014, http://lgmacweb.env.uea.ac.uk/green_ocean/model/

code_description/PFT/ﬂuxes.html). An increased efﬁciency of

carbon export for coccolithophores, due to ballasting, can be

implemented with a higher sinking rate (Moore et al. 2002;

Gregg & Casey 2007).

Diazotrophs (nitrogen-ﬁxers) are an important functional

type, due to their unique role in the nitrogen cycle. They

are thought to prosper under extreme nitrogen limitation,

while suffering a number of disadvantages described earlier.

They are often modelled as having a relatively high iron

requirement, which makes them poor competitors when iron

supply is low relative to N and P (Moore et al. 2002; Ward

et al. 2013). They are also typically given a low maximum

growth rate, which means they cannot take advantage of

variable nutrient supply (Tyrell 1999, Moore et al. 2002;

Monteiro, Dutkiewicz & Follows 2011). It may also be

important for diazotrophs to be poor P competitors (Tyrrell

1999; Lenton & Klausmeier 2007; Ward et al. 2013),

although a relatively low P stoichiometry will at least par-

tially offset this (Moore et al. 2002). In combination, these

traits will lead to diazotroph occurrence in stably stratiﬁed

regions with a low supply ratio of N: Fe and N: P (Ward

et al. 2013). Although most ocean biogeochemical models

represent diazotrophs by a single group, some models

include different diazotroph types with distinct traits, such as

ﬁlamentous Trichodesmium, unicellular N-ﬁxers and symbi-

otic N-ﬁxers living inside diatoms (e.g. Monteiro, Follows

& Dutkiewicz 2010).

Responses of phytoplankton communities to

global environmental change and implications

for biogeochemistry

The many dimensions of global change are increasingly affect-

ing marine environments and the phytoplankton communities

that inhabit them. In the following section, we brieﬂy summa-

rize the most important aspects of the ocean that are changing,

including temperature, acidity and nutrient availability. The

traits and functions of phytoplankton groups and species deter-

mine both how they are affected by global change and how

their responses in turn alter global biogeochemistry. In the

remaining text, we discuss how global change stressors inﬂu-

ence phytoplankton across different levels of biological orga-

nization from functional groups to species.

Global change is multifaceted and includes not just CO

emissions and climate change, but also anthropogenic effects

that are independent of these factors. Acidiﬁcation, rising tem-

peratures and changes in the supply of nutrients and light will

pose the greatest challenges for marine ecosystems and phyto-

plankton communities. Several reviews have examined these

effects in greater detail (e.g. Hoegh-Guldberg & Bruno 2010;

Boyd and Hutchins 2012, Doney et al. 2012), so our summary

is brief. World-wide, oceans act as a major sink for both the

rising levels of atmospheric CO

and the increased amount of

heat trapped by this CO

and other greenhouse gasses. Nearly

1/3 of anthropogenic CO

is being absorbed into the oceans,

where it interacts with water, altering seawater carbonate

chemistry and driving ocean acidiﬁcation (Riebesell 2004; Do-

ney et al. 2009; Hoegh-Guldberg & Bruno 2010). Tempera-

tures have risen by ~0.6 °C over the last 100 years within the

surface layers of the ocean (Hoegh-Guldberg & Bruno 2010)

and are predicted to increase by another 1–3°C by the end of

the century (IPCC; Collins et al. 2013). Temperature plays an

important role in physical ocean processes, driving mixing,

vertical stratiﬁcation and currents. In tropical and temperate

regions, warmer temperatures lead to stronger stratiﬁcation

and shallower mixed layers, simultaneously intensifying nutri-

ent limitation and reducing light limitation (as plankton avoid

being mixed to deeper, darker regions) (Beardall, Stojkovic &

Larsen 2009). Regionally, warming may strengthen coastal

upwelling and nutrient supply and alter the depth of mixing in

polar oceans (Hoegh-Guldberg & Bruno 2010). Temperature

and pH also inﬂuence the chemistry of seawater; changes in

both of these variables will alter the solubility and oxidation

state of trace metals such as iron in ways that are not yet fully

understood (Hoffmann et al. 2012). These changes will be sig-

niﬁcant for marine organisms requiring these trace metals

Global biogeochemical impacts of phytoplankton 1389

community composition alters the functional contributions of

a group, global change can affect biogeochemistry at the in-

tragroup level. Such dynamics occur in response to climate

change within an intermediate complexity Earth system

model, as the ranges of individual species expand and con-

tract based on their trait differences (including growth rates,

temperature optima, size and half-saturation constants), lead-

ing to substantial turnover in the composition of local com-

munities (Dutkiewicz, Scott & Follows 2013). Several

examples exist of between-species trait variation with the

potential to produce intragroup level biogeochemical effects,

limited seemingly only by the number of traits and environ-

mental factors studied so far. Biogeochemical rates are funda-

mentally linked to biological rates, such as growth rate.

Growth rates are strongly linked to temperature, yet individual

species often differ dramatically in their temperature depen-

dence (Thomas et al. 2012; Boyd et al. 2013). In turn, these

differences lead to predicted range shifts as oceans warm

(Thomas et al. 2012). Looking beyond temperature, different

coccolithophorid species and strains engage in calciﬁcation to

varying degrees, with important consequences for productivity

and carbon cycling. In nitrogen-ﬁxing cyanobacteria,

responses to CO

are diverse: elevated CO

stimulated growth

and N

ﬁxation in Trichodesmium species and Crocosphaera

watsonii, important open ocean diazotrophs (Behrenfeld et al.

2006; Beardall, Stojkovic and Larsen 2009, Hutchins et al.

2013), but inhibited these processes in coastal Nodularia

spumigena (Czerny, Barcelos e Ramos & Riebesell 2009;

Boyce, Lewis & Worm 2010). In this case, turnover in cyano-

bacteria species driven by rising CO

might actually support

consistent levels of function (N

ﬁxation) in the face of global

change. Overall, the effects of species turnover within func-

tional group will depend on which species are favoured by

the combinations of global change stressors experienced by

communities regionally and how these species contribute to

ecosystem function.

Finally, within a functional group, individual species need

not have ﬁxed traits (Fig. 1d). Even within primarily asexual

phytoplankton, species consist of multiple strains with differ-

ent traits, representing standing genetic and phenotypic varia-

tion. In addition to between-species differences in the

sensitivity of N ﬁxation rates to pCO

, Hutchins et al. (2013)

demonstrated signiﬁcant strain-level variation. Given this vari-

ation, increasing CO

concentrations may select for particular

strains, inﬂuencing future biogeochemistry in complex ways.

The traits of species also respond to environmental changes

through physiological, behavioural and evolutionary mecha-

nisms. For example, the lipid content of several species of

Antarctic diatom (in the genus Navicula) declines with

increasing temperature, making them a poorer resource for

zooplankton (Teoh, Phang & Chu 2012). Many phytoplankton

are capable of nutrient storage and exhibit variable stoichiom-

etries through a variety of physiological mechanisms.

Changes in the stoichiometry of individual species can be of

similar magnitude to effects driven by turnover between spe-

cies (Behrenfeld et al. 2006; Finkel et al. 2009). This can

inﬂuence their value as a food resource as mentioned

previously. In addition to physiological changes, evolutionary

responses are likely within the phytoplankton due to their

small size, huge populations and rapid generation times.

These responses might allow species to adapt to conditions

imposed by global change, suggesting that (over longer

time-scales) communities might recover from short-term per-

turbations. Understanding the effects of evolutionary

responses to global change is a complex process, requiring

detailed, speciﬁc knowledge of how environments will

change, how quickly phytoplankton can adapt and what

genetic constraints or trade-offs limit them. Experiments

designed to investigate these questions are becoming more

common, but much work remains to be done.

Ultimately, understanding how the effects of global change

will propagate through phytoplankton communities to inﬂu-

ence biogeochemical cycling requires understanding the indi-

vidual and combined effects of stressors on phytoplankton

from species to functional groups. The key to obtaining this

knowledge rests on elucidating the links between environ-

ment, species traits and trade-offs, and their performance and

function.

Conclusions

Understanding the feedbacks between phytoplankton commu-

nity structure and biogeochemistry is a burgeoning area of

research. Distinct functional groups have been shaping the

biogeochemistry of our planet, from the period of oxygena-

tion ushered in by the cyanobacteria to contemporary patterns

of N ﬁxation and carbon cycling in the oceans. Understanding

the physiology and ecology of these groups is continuing. We

are still learning about new members of the phytoplankton

(Cuvelier et al. 2010), the environmental sensitivities and

traits of even the well-studied groups (Boyd et al. 2010),

approaches for efﬁciently modelling diverse plankton commu-

nities and multiple element cycles (Stock, Dunne & John

2014), and methods for conceptually managing complex glo-

bal change (Boyd et al. 2015). These are crucial problems in

an era of global change. Quantifying the traits and trade-offs

that affect functional group composition will help connect

community responses to biogeochemical causes and effects.

A challenging empirical aspect of this approach is the num-

ber of important traits that must be gathered, especially given

trait variation between functional groups, species and popula-

tions. Which traits do we need to measure in order to predict

functional group dynamics? Predicting the outcome of compe-

tition between two populations requires knowledge of traits

relating to competition for resources (macronutrients,

micronutrients and light) and avoidance of predators, all of

which differ between groups (Litchman and Klausmeier 2008;

Boyd et al. 2010; Edwards et al. 2012, 2015). However, this

task is made easier because of several fundamental con-

straints, including both allometric and phylogenetic relation-

ships. Many traits scale strongly with cell size and

temperature, including cellular stoichiometry (Yvon-Durocher

et al. 2015). Due to the physical constraints that size and

shape place on cells’ability to obtain nutrients, cell size cor-

1392 E. Litchman et al.