Content uploaded by Simon Jennings
Author content
All content in this area was uploaded by Simon Jennings on Dec 06, 2017
Content may be subject to copyright.
Part III
Practical Applications
259
261
Metabolic Ecology: A Scaling Approach, First Edition. Edited by Richard M. Sibly, James H. Brown, Astrid Kodric-Brown.
© 2012 John Wiley & Sons, Ltd. Published 2012 by John Wiley & Sons, Ltd.
Chapter 21
Marine Ecology
and Fisheries
Simon Jennings , Ken H. Andersen ,
and Julia L. Blanchard
SUMMARY
1 Variations in metabolic rate with body size and
temperature drive patterns of energy acquisition
and use. These have consequences for the life histo-
ries of marine plants and animals and the structure
and function of populations, communities, and
ecosystems.
2 Small primary producers and size - based preda-
tion characterize marine ecosystems. Since the
transfer of energy to predators at each step in the
food chain is ineffi cient, less energy is available for
larger individuals at higher trophic levels.
3 Changes in energy supply and demand with size
account for observed trends in abundance with size.
These are often described with “ size spectra, ” log -
log relationships between the total abundance of
individuals and their body size. Slopes of size spectra
are remarkably consistent among ecosystems that
differ in productivity and temperature.
4 Models of the processes structuring size spectra
rely on the principles of the metabolic theory of
ecology (MTE). Species identities are often ignored
in these models, because many of the key processes
depend more on body size than on taxonomic
identity.
5 The limited parameter demands of size - based
models simplify application to diverse systems at
many scales. Size - based models have been used to
investigate the effects of food chain coupling and
the role of prey size selection, as well as to predict
fi sh production, community size composition, and
fi shing impacts.
6 Accounting for species identity in a size - based
framework provides additional insight into life -
history trade - offs and food web structure. The
approach also supports analysis and advice on
management issues that are species focused.
7 Applications of size - and species - based models
include the analysis and prediction of life histories
in a community context and the assessment of
trade - offs between fi shing and conservation.
8 Metabolic ecology underpins the assessment and
management of fi shing impacts on populations,
communities, and ecosystems. The drivers for
developing these methods remain strong because
they are providing new insights into marine ecology
and because fi shery managers increasingly demand
advice that places fi shery management in a com-
munity and ecosystem context.
262 Practical applications
the fi nal sections we describe how the understanding
of life histories and community interactions has been
used for measuring fi shing impacts and developing
assessment and management tools. The chapter pro-
vides examples of the ways in which fundamental
advances in the understanding of ecological processes
can be used to solve applied problems.
21.2 LIFE HISTORIES
To achieve population persistence, adults that die must
be replaced. That replacement can be achieved by
many alternate life histories is amply demonstrated by
the variety of life forms that persist in the sea. These
extant life forms must all achieve 1 : 1 replacement
despite differences in size, longevity, age at maturity,
and reproductive output – a consequence of life - history
trade - offs (see Sibly, Chapter 5 ). If there were no trade -
offs every animal would start reproducing at birth,
suffer no mortality, and produce large numbers of
young at frequent intervals as it got infi nitely older: the
“ Darwinian demon ” of Law (1979) .
Body size and temperature have a profound infl u-
ence on observed life histories through their effects on
metabolic rate. With metabolic rate driving the “ pace
of life, ” other aspects of the life history show compen-
satory adjustment. These compensatory adjustments
serve to maximize lifetime reproductive output and
are remarkably consistent within and among taxa.
Indeed, relationships among life - history parameters
such as reproductive output, size and age at maturity,
maximum size and age, natural mortality, and growth
rate are often used to predict parameters for taxa where
details of the life history are not known (Beverton and
Holt 1959 ; Pauly 1980 ; Gislason et al. 2010 ). Both
growth and natural mortality have long been used as
life - history parameters, although their values and
associated trade - of fs are more easily understood in a
community context (section 21.6 ).
Temperature effects are manifest across the latitudi-
nal ranges of marine species. With increasing tempera-
ture increasing metabolic rate and the “ pace of life, ”
aspects of the life history such as age at maturity and
annual reproductive output respond to maintain life-
time reproductive output (Beverton 1992 ). Thus
species ’ populations inhabiting warmer waters have
faster growth, early maturity at smaller size, and
higher annual reproductive output.
21.1 INTRODUCTION
Marine ecosystems cover around two - thirds of the
Earth ’ s surface and account for nearly half of global
primary production. The body mass of taxa found in
these ecosystems spans 20 orders of magnitude, from
bacteria to whales, and mean sea temperatures range
from less than 0 ° C at the poles to almost 30 ° C in the
tropics. The marine environment provides food for
people, with fi sheries yielding around 100 million
tonnes annually. Metabolic ecology helps us to under-
stand structure and processes in marine ecosystems
and to assess and manage the effects of fi shing.
Metabolic rate varies with body mass and tempera-
ture (see Brown and Sibly, Chapter 2 ). Variations in
metabolic rate drive patterns of energy acquisition and
use, with consequences for the life histories of marine
plants and animals and the structure and function of
their ecosystems. Thus smaller species have higher
metabolic rates per unit body mass, faster growth,
higher maximum population growth rates, greater
annual reproductive output, higher natural mortality,
and shorter lifespan. Conversely, larger species have
lower metabolic rates, slower growth, lower maximum
population growth rates, lower annual reproductive
output, lower natural mortality, and greater longevity.
These differences in life history af fect species ’ responses
to their physical and biological environment and, for
those animals exploited by humans, to fi shing (Fenchel
1974 ; Banse and Mosher 1980 ; Charnov 1993 ; Brown
et al. 2004 ; see also Sibly, Chapter 5 ).
We begin this chapter by seeking to understand how
metabolism infl uences the life histories of marine
animals and their responses to temperature. Our focus
is unashamedly fi shy, since metabolic rates (Winberg
1956 ; Clarke and Johnston 1999 ) and life histories
(Beverton and Holt 1959 ; Charnov 1993 ) have been
so comprehensively studied in this group of signifi cant
ecological, economic, and social importance. The prin-
cipal link between population processes and the meta-
bolic processes described in other chapters of this book
are related to the role of metabolism in setting the
“ pace of life. ” The “ pace of life ” determines other
aspects of the life history and responses to mortality,
both from other predators in the sea and from humans
fi shing. In the core of the chapter we consider the life
histories of species in the context of communities and
ecosystems, and how species interactions lead to the
characteristic structuring of marine communities. In
Marine ecology and fi sheries 263
21.3 FOOD WEBS
Metabolism creates an energy demand and consumers
acquire the energy to meet this demand by feeding on
primary producers and/or other consumers. This
establishes a food web (Petchey and Dunne, Chapter 8 ;
Hechinger, Lafferty, and Kuris, Chapter 19 ), where
most energy passes from smaller to larger individuals
to provide the resources needed to metabolize, grow,
and reproduce, and ultimately to achieve the 1 : 1
replacement needed for persistence. The transfer of
energy through the food web is ineffi cient, so the aggre-
gate production of predators is always less than the
production of their prey.
Trophic levels are used to measure the number of
steps in a food chain that lead to a given consumer,
with primary producers typically assigned a trophic
level of 1. When multiple feeding pathways support a
consumer, as is usually the case, the trophic level is
likely to be fractional. For this reason there is a trophic
continuum in the sea, with trophic level rising almost
continuously as a function of body mass (Fig. 21.1 ).
The relationship tends to be strongest in pelagic
systems and weakest in coastal systems, where phyto-
plankton production accounts for a smaller proportion
of total primary production and larger primary pro-
ducers such as macroalgae, sea grasses, or corals may
be grazed directly by large herbivores (invertebrates,
fi sh, reptiles, and mammals).
Globally, phytoplankton account for approximately
90% of total marine primary production (Duarte and
Cebri á n 1996 ). Consequently most grazers are also
small, and energy is transported from the phytoplank-
ton to the macrofauna through predation by larger
organisms on smaller ones. Several predation events
may occur before energy reaches the largest fi sh, which
is why marine food chains often extend to 4 or 5 trophic
levels (Vander Zanden and Fetzer 2007 ). With meta-
bolic rate setting the pace of life for individuals, the
smaller individuals and species in a food web turn over
faster and consume more than the larger ones, so rates
of energy fl ux per unit mass are higher. These differen-
tial rates affect community interactions and the
balance of species abundances (Fig. 21.1 .).
Marine food webs have been conceptualized in three
principal ways, refl ecting interactions among species,
interactions among size - classes, and interactions
among species and size - classes. Abstractions based
purely on body mass of individuals are rarely used in
While 1 : 1 replacement needs to be achieved on
average, mortality rates will fl uctuate with changes in
abiotic conditions and predator abundance. Further,
for fi shed populations, there is additional mortality
that must be tolerated for the population to persist.
Population responses to increased mortality rates are
governed by maximum population growth rate and the
strength of compensation. Both are linked to body size.
Maximum population growth rates are faster in
smaller species with early maturity and higher annual
reproductive output (Myers et al. 1999 ; Denney et al.
2002 ) (see also Sibly, Chapter 5 ). For a given species,
they also increase with temperature. Compensation is
the capacity for increased population growth as mor-
tality increases. Compensation increases with body size
(Goodwin et al. 2006 ). The balance between compen-
sation and population growth rates explains why popu-
lations of larger species often yield large catches as
they are fi shed more heavily but, once they become
depleted, their low maximum population growth rates
prevent recovery. Small species, conversely, can be sig-
nifi cantly depleted as they are fi shed more heavily but
bounce back more quickly once fi shing effor t is reduced.
We have been referring to body size or mass as con-
venient correlates of metabolic rate and life history, but
this is a generalization that does not apply in a consist-
ent way across all animal groups. The generalization is
often considered acceptable for many pelagic species,
where differences in the amounts of metabolically
active tissue and structural tissue are quite limited, but
for some benthic species, and especially for deep - sea
species, a large proportion of body mass is metaboli-
cally inert or has low metabolic activity (Drazen and
Seibel 2007 ). This includes structural materials such
as chitin and shell, and also lipids that provide buoy-
ancy in deep - sea species and other fi shes that lack
swim bladders. Here, other measures such as energy
content of living tissue may be more appropriate.
Further, it should be recognized that body sizes of indi-
viduals within a fi sh species can vary by many orders
of magnitude, from egg sizes around 1 mg in most
species to maximum body sizes of several hundred kg
in whale sharks, marine sunfi sh, marlins, and tunas. It
is therefore important to distinguish between
individual - level rates, such as metabolic rate, which
scale with individual size, and population rates, such
as fecundity, mortality, and population growth rates,
which scale with a characteristic size at maturity or
maximum size.
264 Practical applications
mately linear negative relationship between the loga-
rithm of abundance and the logarithm of body mass
(Sheldon et al. 1972, 1977 ). The slopes of these so -
called size spectra (Fig. 21.1 ) are remarkably consist-
ent in environments ranging from the tropics to
temperate latitudes and the poles (Boudreau and Dickie
1992 ). The description and analysis of these patterns
by Sheldon and others was a great example of the
approach that was later described as “ macroecology. ”
When the size spectra are represented as the total
biomass in logarithmic size groups (e.g., placing all
individuals from 1 g to 10 g in one bin, from 10 g to
100 g in the next, and so on), the biomass in each
group is roughly constant or decreases weakly with
body size (Sheldon et al. 1972 ). This pattern is referred
to as the “ Sheldon hypothesis, ” and it implies that the
total biomass of predators is the same as, or a consist-
ent but relatively large fraction of, the total biomass of
their prey. Note that size spectra shown here are similar
to the ones commonly plotted for trees in forests, which
also consider size but not species identity (Enquist and
terrestrial food web analysis (Woodward et al. 2005a )
but are especially appropriate in the marine environ-
ment for at least three reasons. First, detailed aspects
of the life history and population dynamics are not well
known, especially in less accessible environments such
as the open ocean and deep sea. Second, the entire food
web in the sea is structured by size as much as by
species identity, with tiny phytoplankton accounting
for around 90% of primary production and the body
size of consumers broadly linked to their position in the
food chain (Duarte and Cebri á n 1996 ). Third, many
marine species grow by 5 – 6 orders of magnitude in
mass during their life cycle, performing a series of
trophic roles that are driven as much by size as their
taxonomic identity (Cushing 1975 ). Indeed, studies of
fi sh communities have shown that increases in trophic
level with body size are primarily driven by increases in
the trophic level of component species as they grow
(Jennings et al. 2001 ).
Studies of the size composition of individuals in
marine food webs have long demonstrated an approxi-
Figure 21.1 The generalized scaling of the total biomass of individuals in a marine food web sorted into logarithmic size
bins (e.g., 1 g – 10 g, 10 g – 100 g), illustrating the biomass “ size spectrum. ” Body mass spans many orders of magnitude from
primary producers to top predators and trophic level increases with body mass owing to size - based predation. The scaling
exponent is approximately − 0.1, so the total biomass of smaller organisms tends to be slightly higher than the biomass of
larger ones; but smaller organisms have relatively faster turnover times so their production is much higher. The ineffi cient
transfer of energy from prey to predators means that production falls by 80 – 90% at each step in the food chain. Owing to
their considerable scope for growth, the same species are found in several size - classes and feed at several trophic levels.
Drawings and photographs of organisms courtesy of R. Beckett, S. R. Jennings, and J. H. Nichols.
10−8
101
100.5
100
10−0.5
10−6 10−4 10−2 10010210 4106
100
102101
Relative turnover time (1/ P:B)
Relative biomass
Body mass (g)
Trophic level
10−1
123456
Marine ecology and fi sheries 265
impact on the slope of the size spectrum (Barnes et al.
2010 ). The relative constancy of predator – prey mass
ratios and transfer effi ciency and the absence of rela-
tionships with temperature or primary production help
explain why the slope of the size spectrum is also rela-
tively constant among ecosystems. However, the
“ height ” of the size spectrum does change with pro-
ductivity, refl ecting changes in the total numbers of
animals present with changes in the energy to support
them.
21.4 FOOD WEB COMPLEXITY
The broad characterizations of marine food webs based
on size spectra provide an appealing synthesis that
links process and structure, but the acceptable level of
abstraction depends on the questions being addressed.
For example, knowledge of processes governing bulk
energy fl ux is important for understanding and pre-
dicting community structures and ecological processes
in ecosystems that vary widely in productivity, biodi-
versity, and physical characteristics, and where knowl-
edge of the biota may not support consistently detailed
analyses. For these ecosystems, simple size - based rules
have been used to estimate global fi sh biomass and the
role of fi shes in biogeochemical processes, analyses
that would not have been tractable if approached on a
species - by - species basis (Jennings et al. 2008 ; Wilson
et al. 2009 ). Conversely, when questions about species
groups and species need to be addressed, for example
in regional and ecosystem - scale assessments, then
additional information has to be included.
Of course, the range of trophic ecologies in marine
food webs is more complex than implied by simple size -
based abstractions. Size spectra most accurately
describe communities of the open ocean where phyto-
plankton are the primary producers and almost all
energy transfer occurs in the pelagic environment. By
contrast, coastal and shelf communities are typically
more complex, with much of the energy captured and
processed by communities on the seafl oor. Seafl oor
animals can be herbivores and detritivores, which
obtain energy from benthic primary producers and
detritus exported from the pelagic food web. Owing to
the relatively shallow depths in coastal and shelf seas,
the production of seafl oor detritivores is accessible to
predators that may also forage in the open sea.
As well as being used for system - wide analysis, size
spectra can be compiled for different components of a
Bentley, Chapter 14 ). They differ from most of the
abundance – body mass relationships described else-
where in this book (Isaac, Carbone, and McGill, Chapter
7 ; Petchey and Dunne, Chapter 8 ; Hechinger, Lafferty,
and Kuris, Chapter 19 ), which are relationships
between log mean abundance and log mean body mass
where the data points represent single species or popu-
lations (e.g., of birds or mammals).
Explanations for the slopes of size spectra, where
slope refl ects the rate of change in abundance with
size, have been based on detailed process - based models
of predator – prey interactions and more simplistic
models based on fundamental ecological principles.
Almost all models are underpinned by the recognition
that the scaling of metabolism with body size accounts
for differences in the energy requirements of animals
in different size - classes.
In general terms, the slope of the size spectrum is a
function of the effi ciency of energy transfer from prey
to predators, dubbed “ trophic transfer effi ciency, ” and
the predator – prey size ratio (Borgmann 1987 ;
Hechinger, Lafferty, and Kuris, Chapter 19 ). Trophic
transfer effi ciency can be estimated directly or pre-
dicted by modeling the processes that account for
energy transfer, including the probability of encoun-
tering prey, the probability of prey capture, and the
gross growth effi ciency (Andersen et al. 2009 ; Petchey
and Dunne, Chapter 8 ). The scaling of maintenance
metabolism with size (Brown and Sibly, Chapter 2 ) does
not affect the relative abundance of animals with dif-
ferent maximum sizes in the size spectrum, and it can
be shown analytically that a change in the scaling from
M 3/4 to M 2/3 changes the slope of the resulting size
spectrum by only < 5%, an effect that is unlikely to be
detectable in data (Andersen and Beyer 2006 ).
Measurements of trophic transfer effi ciency in differ-
ent marine environments suggest that it is not affected
by changes in temperature and productivity among
ecosystems and typically ranges from 10% to 20%
(Ware 2000 ). Mean predator – prey mass ratios are
similarly unaffected and often range from 10
2 to 10
4 : 1
(Barnes et al. 2010 ). For this reason, patterns of energy
transfer from primary producers to fi sh are broadly
comparable in marine ecosystems, and high rates of
primary production tend to be translated into high
rates of fi sh production in all areas (Fig. 21.2 ).
There is some evidence that trophic transfer effi -
ciency decreases at higher trophic levels and predator –
prey mass ratios increase, but these changes appear to
counter one another as they do not have a discernible
266 Practical applications
biomass and log body size). Given that the rate of
metabolism scales as approximately M 3/4 , energy use
by animals in such communities is hypothesized to be
independent of body mass. This is the “ energy equiva-
lence hypothesis ” as fi rst proposed by terrestrial ecolo-
gists based on the scaling relationship between mean
abundance and mean size of different animals (Damuth
1981 ; see also Isaac, Carbone, and McGill, Chapter 7 ,
and Hechinger, Lafferty, and Kuris, Chapter 19 ). Their
argument is not entirely consistent with the argument
we are making here for a marine community that
shares the same energy source, but leads to the same
community. Differences in size - spectra slopes for differ-
ent components of the community refl ect differences
in feeding interactions. Several investigators have iden-
tifi ed differences between the slope of a size spectrum
in a predation - based community and in a community
in which energy is shared (Brown and Gillooly 2003 ;
Blanchard et al. 2009 ). In communities that share
energy (such as shellfi sh of different sizes that all eat
phytoplankton and detritus) the slope of the relation-
ship between log numbers and log body size tends to be
shallower than in predation - based communities and
approximates − 0.75 (equivalent to + 0.25 for log
Figure 21.2 The global distribution of phytoplankton primary production (upper panel) broadly refl ects the global
distribution of fi sheries catches (lower panel). Global phytoplankton production of 45 × 1 0 8 tonnes C per year is estimated to
be ≈ 90% of total marine primary production. From SeaWIFS data processed by Fr é d é ric M é lin and Rodney Forster (A) and
from Worm et al. (2009) , based on an analysis by Reg Watson (B).
Primary
production
(gC m−2yr−1)
Catches
In (t km−2yr−1)
1000
500
0
> 8
> 4
< 0
< −4
Marine ecology and fi sheries 267
21.5 FISHING IMPACTS
Fishing has wide - ranging impacts on populations,
communities, and ecosystems. The capacity of a
species ’ population to withstand fi shing mortality
depends on the rate of mortality and the life history of
the population. The faster life histories of smaller
species typically confer more resilience to fi shing mor-
tality. Fishing affects communities and ecosystems by
changing the relative abundance of species, with
knock - on impacts on food web interactions. The
impacts of fi shing on communities and ecosystems
have been an increasing focus of recent research,
refl ecting societal desire and political commitments to
meet environmental targets for ecosystems as well as
sustainable fi sheries. Analysis of these impacts poses
many new challenges for a scientifi c community that
has largely focused on fi shing impacts on a few well -
studied populations.
Susceptibilities of populations to fi shing mortality
are body - size dependent because smaller species have
greater capacity to sustain additional mortality (Fig.
21.3 ). Rates of fi shing mortality also tend to be size -
dependent because dif ferent fi shing gears target differ-
ent size - classes in populations and communities and
because the management system often defi nes regula-
tions for minimum mesh sizes to allow smaller indi-
viduals to escape. Consequently, in real fi sheries, larger
species are usually subject to higher mortality rates
and less able to sustain them. This leads to the differ-
ential depletion of larger individuals and species in
many communities, modifying the slopes of size spectra
that would be expected in unexploited systems (section
21.3 ). There are many examples of slopes of size
spectra becoming steeper with increased fi shing rates
(Rice and Gislason 1996 ; Bianchi et al. 2000 ). Changes
in the slope of the spectrum are due to the depletion of
large fi sh and the proliferation of small fi sh as their
larger predators are depleted (Dulvy et al. 2004 ; Daan
et al. 2005 ).
The relative impacts of fi shing on a community can
be quantifi ed by comparing the slopes of observed size
spectra with the predicted slopes of the spectra in the
absence of fi shing. In the North Sea, this approach was
used to predict that the biomass of large fi shes weigh-
ing 4 – 16 kg and 16 – 66 kg, respectively, was 97.4%
and 99.2% lower than would be expected in the
absence of fi sheries exploitation. In addition, because
the smaller fi shes that now dominate the size spectrum
have faster turnover times, the mean turnover time of
hypothesis (Jennings et al. 2007 ). In a predation - based
community, available energy supply decreases with
increasing body size, because the transfer of energy
from prey to predators is ineffi cient. This accounts for
the steepening of the size spectrum to slopes of − 1 to
− 1.2 or so (equivalent to 0 to − 0.2 for log biomass and
log body size in Fig. 21.1 ), rather than − 0.75.
If the food web is divided into food chains that are
based on energy sharing and predation then these will
be coupled to some extent. When coupling is stronger,
more energy from one food chain is transferred to the
other, either by predation or by consumption of detri-
tus. Models of the dynamic coupling between food
chains have shown that the strength of coupling
affects food web resilience. Resilience is a measure of
the rate at which a community returns to a “ steady
state ” following a change in primary production or
predation (Blanchard et al. 2011 ). Although just a
model - based analysis at present, the results imply that
deep pelagic oceans, where there is low coupling, are
more likely to be vulnerable to human and environ-
mental impacts than shallow coastal seas.
Species are a key focus of ecology, conservation, and
fi sheries but their identities are ignored in most treat-
ments of the size spectrum. Incorporating species in
the size spectrum is challenging when most fi sh eggs
have a size of 1 mg but fi sh species reach maturity at
sizes ranging from 1 g to 100 kg. In terrestrial ecology,
species are often characterized by a representative body
size, such as the size of an adult, but this approach
oversimplifi es marine food webs where individuals of
the same species can fulfi ll many roles in the food web
as they grow. From an analytical perspective, the large
scope for growth in fi sh requires that the whole size
structure of a species ’ population is resolved in food
web analysis, and the community size spectrum can be
decomposed into the size spectra for species with differ-
ent maximum size (Andersen and Beyer et al. 2006;
Pope et al. 2006 ; Fig. 21.3 ). Such approaches have
now been adopted to assess the effects of fi shing on
marine communities and ecosystems and are addressed
in section 21.5 . Species - specifi c differences in feeding
strategies and behavior can lead to interesting and
informative departures from the average tendency. For
example, fi lter - feeding sharks and whales, that can
“ feed down the food chain ” on smaller and more pro-
ductive size - classes of prey, have evolved a strategy that
gives them access to a greater resource base by using
their mobility to follow areas of abundant zooplankton
at cross - ecosystem scales.
268 Practical applications
the fi shed community fell from 3.5 to 1.9 years
(Jennings and Blanchard 2004 ; Fig. 21.4 ).
As well as the size - selective effects of fi shing on fi sh
communities, the physical contact between seabed
fauna and towed fi shing gears also results in size -
selective mortality. Towed gears may differentially kill
larger animals because smaller ones can be pushed
aside by the pressure wave in front of the gear (Gilkinson
et al. 1998 ). Consequently, benthic communities in
trawled areas tend to be dominated by smaller individu-
als and species, and the slopes of benthic invertebrate
size spectra become steeper in more heavily trawled
areas (Duplisea et al. 2002 ; Hiddink et al. 2006 ).
21.6 FISHERY ASSESSMENT AND
MANAGEMENT
Fisheries are managed because unregulated fi sheries
have consequences that society deems undesirable.
These consequences include collapses of fi sheries that
result in long - term loss of yield, with costly social and
Figure 21.4 The predicted slope of a size spectrum for an
unexploited North Sea fi sh community and the slope of the
size spectrum as determined from data for the exploited
North Sea fi sh community in 2001. The steeper slope for the
exploited community revealed the extent of decline in the
abundance of large fi shes following intensive fi shing.
Body mass (g)
Biomass
10
1
0.1
100
4.2 4.4 4.6
Trophic level
4.8 5.0
1,000
data
prediction
10,000
Figure 21.3 Biomass siz e spectra for a modeled community of six species spanning a range of maximum body sizes (A) that
is not fi shed and (B) where each species is fi shed with a trawl - type of size selection pattern. Both panels show the biomass of
all fi sh in the fi sh community (magenta lines) and of six species with varying asymptotic sizes (red lines) as a function of
individual weight. The dashed line is the theoretical spectrum (elevated for clarity), and the green line in (B) is the community
spectrum from (A). The calculations are based on the assumption that the total metabolism of an individual scales as M 3/4 ,
and fi shing impacts each species from 5% of the asymptotic weight and onwards with a fi shing mortality of 0.75 per year.
Fishing on all species with the same fi shing mortality has the largest impact on the bigger species in the community,
as the smaller species have a higher rate of production than the bigger species. Modeling approach based on
Anderson and Beyer (2006) .
AB
0.2
0.1
0.05
0.02
0.1 g 10 g 1 kg
Body mass
Relative biomass
100 kg 0.1 g 10 g 1 kg
Body mass
100 kg
Marine ecology and fi sheries 269
ers, targeted by fi shers, and regulated by managers.
Accordingly, fi sh communities have also been modeled
to predict the abundance of particular species in
defi ned size - classes. Such species - specifi c models are
based on the assumptions that growth dynamics can
be described by known parameters describing growth
rate, asymptotic size, size at maturity, and size -
dependent rates of reproduction. Feeding interactions
can be characterized by a binary diet matrix that
defi nes which species eat which other species, as in
Petchey and Dunne (Chapter 8 ). Mortality rates are the
sum of non - predation mortality, predation mortality,
and fi shing mortality. Predation mortality depends on
economic impacts, as well as unwanted changes to
marine habitats, species, and ecosystems that are seen
as a conservation issue.
To support management and to avoid the unwanted
impacts of fi shing, assessments of the rates of fi shing
that lead to high and sustainable catches have been
provided for decades. Developing processes to conduct
and improve these assessments is a central focus of
fi sheries science. Broadly, the assessment process
involves estimating population size and sustainable
rates of fi shing by accounting for inputs (growth,
recruitment) to and outputs (natural mortality, fish-
ing mortality) from the population. The assessment
process is currently focused on individual populations,
and is data intensive. Natural mortality is a parameter
that provides a link between single - species population
dynamics and the infl uence of the wider food web, and
may be estimated from life - history invariants and com-
munity models (Pope 1979 ; Andersen et al. 2009 ;
Gislason et al. 2010 ).
Despite the management focus on population assess-
ment, populations are embedded in communities and
ecosystems, and there is growing interest in under-
standing interactions among small groups of species
caught in the same fi sheries. To some extent there has
always been a push – pull in fi sheries and marine envi-
ronmental management between tractable abstrac-
tions of the ecosystem characterized by single - species
population analysis and community - based analyses
where mortality and growth are treated explicitly.
However, there has been a recent and sustained resur-
gence of interest in community - based analysis, not least
because management targets for community properties
such as size composition are now being considered in
some jurisdictions. Understanding the response of the
interacting populations to fi shing allows explicit exami-
nation of the trade - offs between the status of fi shed
populations and aspects of the community such as size
composition and trophic level. The disadvantages of
community - based analysis are the greater parameter
demands and diffi culties of formalizing transient and
complex interactions between species.
We have already discussed the development of size -
based models that incorporate species identity. This is
achieved by using maximum body size as a proxy for
species identity and has provided a generalized under-
standing of the processes structuring communities
and responses to fi shing (Fig. 21.3 ; Pope et al. 2006 ;
Andersen et al. 2008 ). However, real species matter in
fi sheries since they are known and desired by consum-
Figure 21.5 Modeled relationships between fi shery
catches, properties of a fi sh community, and the exploitation
rate. Lines denote the total catch from a fi sh community
(blue), the biomass of fi sh in the community (green), the
mean maximum size of fi shes in the community (a metric of
the life - history composition, yellow), and the number of
collapsed species (species are defi ned as collapsed when
biomass falls to < 10% of predicted biomass in the absence of
fi shing, red). MMSY is the maximum multispecies
sustainable yield that can be sustained by the community,
but achieving MSSY comes at costs to the larger, more
vulnerable species which are expected to collapse at these
exploitation rates. In this example, managers that seek to
reconcile demands for fi shing and conservation might aim
for exploitation rates around 0.2. At this exploitation rate
there is a small loss of potential yield to the fi shery but
changes to the community and the collapses of vulnerable
species are minimized. Simulations by J. S. Collie using the
size - based model of Hall et al. (2006) ; from Worm et al.
(2009) .
0.0 0.2
Percent of maximum
100
80
60
40
20
0
0.4
Rebuilding Overfishing
MMSY
0.6
Total catch
Total biomass
Mean Lmax
Collapsed species
Exploitation rate
0.8 1.0
270 Practical applications
the predator – prey size ratio and predator abundance,
while fi shing mortality depends on size, species iden-
tity, and the selectivity of fi shermen (Hall et al. 2006 ).
The strength of these models is that they allow explicit
consideration of the trade - offs between objectives for
fi sheries (e.g., high and sustainable yields or profi ts)
and conservation (e.g., maintaining viable populations
of vulnerable species and community properties) (Fig.
21.5 ).
Size - and species - based models have since been elab-
orated and extended to include dynamic food -
dependent growth, species - specifi c feeding behavior,
and an interaction matrix based on species spatial co -
occurrence). The new methods allow for both popula-
tion and overall community dynamics to be tracked
through time and for trade - offs between fi sheries yields
and environmental “ health ” indicators to be evaluated
across different species and size - specifi c harvesting
strategies. Alternate species and size - structured models
are also being used as part of simulation frameworks
to evaluate the performance of indicators in support-
ing achievement of conservation and management
objectives (Fulton et al. 2008 ). Size - and species - based
models can also be used to support single - species
assessments by better predicting natural mortality, one
of the most challenging parameters to measure directly.
Mortality rates can thus be seen as a consequence of
predators in the community rather than as intrinsic
life - history parameters related to lifespan (Andersen
et al. 2009 ). The models “ explain ” natural mortality as
a consequence of predator – prey interactions. In other
words, for one predator to satisfy its metabolic demands,
a corresponding number of prey have to die.
Interestingly, the balance between growth of preda-
tors, determined by their metabolic demands, and the
rate of death of smaller prey organisms, leads to mor-
tality scaling as M − 1/4 where M is the body mass of an
individual (Andersen et al. 2009 ).
21.7 CONCLUSIONS
The application of metabolic scaling theory to infer
population and ecosystem properties has a long history
in marine science, from the empirical determination of
relations between metabolism and population growth
rate to the role of metabolic scaling in predicting the
community size spectrum. Variations in metabolic rate
with body size and temperature drive patterns of
energy acquisition and have been shown to have con-
sequences for the life histories of marine plants and
animals and the structure and function of populations,
communities, and ecosystems. These insights into the
processes infl uencing marine communities have led to
present - day size - based analyses and dynamical com-
munity models to describe the effect of fi shing, climate,
and changes in primary production on the marine eco-
system. Further, the growing capacity to account for
species identity in size - based community models has
provided additional insights into life - history trade - offs
and food web structure, while supporting analysis and
advice on fi sheries management issues that are species
focused. Based on such insights we conclude that the
understanding of the importance of metabolism in
driving some of the main population, community, and
ecosystem processes in marine ecosystems is maturing.
The next steps are likely to involve moving beyond the
study of broad scaling patterns, to understand differ-
ences between metabolic predictions and observations,
at both the process and system levels. For example, an
intriguing new analysis of empirical data shows that
the scaling of natural mortality with body size varies
systematically with maximum size (Gislason et al.
2010 ). Drivers for the next steps in this research
remain strong, as fi shery managers increasingly
demand advice that places fi shery management in a
community and ecosystem context.
