Download Effects of biodiversity on ecosystem stability: distinguishing between

Nelson Valdivia
Effects of biodiversity on ecosystem stability
Distinguishing between number and composition of species
PhD thesis
University of Bremen
Germany, December 2008
Biologische Anstalt Helgoland
Alfred Wegener Institute for Polar and Marine Research
Marine Station
Ph.D. thesis
Effects of biodiversity on ecosystem
stability: distinguishing between number
and composition of species
Dissertation zur Erlangung des Doktorgrades der Naturwissenschaften
Vorgelegt dem Fachbereich Biologie/Chemie der Universität Bremen von
Nelson Valdivia
Gutachter:
1. Prof. Dr. Kai Bischof
2. Prof. Dr. Christian Wiencke
December 2008
Abstract
Declines in biodiversity have caused concern because of ethical and aesthetic reasons, but
also because of the consequences for the goods and services provided by natural ecosystems. Consequently, ecologists have focused for decades on testing the idea that systems
with more species are more stable. The results, however, have been complex and inconsistent. In particular, it is still unclear whether high stability in species-rich communities is due
to the number of species per se (species richness) or to the increased likelihood of including
particular species or functional types (species composition). In this thesis, I evaluated the
contribution of species richness and species identity to the stability of marine hard-bottom
communities. Combining observational and manipulative experimental methods, I conducted three field studies in intertidal and shallow subtidal habitats of Helgoland Island, NE
Atlantic. First, I conducted an observational study to test whether intertidal communities
containing many species are more stable (i.e. do vary less over time) than communities containing fewer species. Species covers were estimated every 6 months for 24 months and an
index of stability was calculated for total community cover across time (S = mean SD-1).
Second, I conducted a synthetic-assemblage experiment––in which I increased the diversity
of field-grown sessile suspension-feeding invertebrates––to determinate whether assemblages containing several functional groups consume a greater fraction of resources than is
caught by any of the functional types grown alone. (A functional group is a group of species
with the same effect on an ecosystem property.) Finally, I conducted a removal experiment
to test whether the loss of the canopy-forming alga Fucus serratus and mechanical disturbances that provide free substratum affect the temporal variability in cover of intertidal
communities. In the removal experiment, species covers were estimated every 3 months for
18 months and the temporal variance was analysed.
In general, the effects of the number of species and functional groups on ecosystem stability were weaker than those of species composition. In the observational study, stability
was a negative and curvilinear function of species richness, which probably resulted from
the dominance of few species. In accordance, the synthetic-assemblage experiment showed
that there was no relationship between resource consumption and functional group diversity
per se, but that different functional groups had idiosyncratic effects. On the other hand, the
removal of Fucus changed the physical environment by increasing temperature, irradiance,
and amount of sediment, which depressed the abundance of sensitive species like encrusting
algae and small sessile invertebrates, but raised the abundance of more tolerant species like
i
ephemeral green algae. This resulted in a significant increase in the variability of species
abundances, but not in that of communities. The negative covariances resulting from the
compensation between sensitive and tolerant species buffered the community stability
against the environmental disturbances. These patterns were consistent across two sites,
suggesting a consistent effect of canopies across the spatial variability of this system.
Species composition appears to be more important for ecosystem stability than taxonomic and functional richness. Yet, the occurrence of compensatory dynamics in the face of
environmental changes (i.e. the removal of Fucus) suggests that a variety of species with
differing environmental tolerances is needed to maintain the functioning of this ecosystem.
Therefore, predicting the consequences of species loss requires a detailed knowledge about
the effects of species on ecosystem functioning and their responses to the environment. Conservational managers should strive (i) in identifying species with disproportional effects on
ecosystem functioning, and (ii) in maintaining a redundancy of species with similar effects
on ecosystem functioning and a diversity of species with different sensitivities to a suite of
environmental conditions.
Keywords:
Biodiversity, ecosystem stability, species compensation, conservation
ii
Contents
Preface ....................................................................................................................................iv
Acknowledgments................................................................................................................... v
List of papers..........................................................................................................................vi
1
Introduction...................................................................................................................... 1
The context: the value of biodiversity ................................................................................... 1
Definitions ............................................................................................................................. 1
Theory ................................................................................................................................... 3
Size of ecosystem properties.............................................................................................. 4
Variance in species properties............................................................................................ 6
Observations and experiments .............................................................................................. 8
The model system: hard-bottom ecosystems ......................................................................... 9
Aims....................................................................................................................................... 9
2
Methods........................................................................................................................... 11
Study sites ............................................................................................................................ 11
Sampling and experimental designs.................................................................................... 11
3
Results and Discussion................................................................................................... 14
Species richness vs. species composition ............................................................................ 14
Species’ response traits influence community stability ....................................................... 17
The role of replication in biodiversity experiments............................................................. 17
Conclusion........................................................................................................................... 18
References.............................................................................................................................. 20
Glossary ................................................................................................................................. 25
Appendix ............................................................................................................................... 28
iii
Preface
This thesis reports the outcome of field-based experiments carried out during the last three
years and designed to explore the role of biological diversity in maintaining the stability of
coastal ecosystems. The experiments were designed to test theoretical predictions and
mechanisms that explain the effects of biodiversity. So, at the first glance, the scope of this
thesis might seem confined to the academic realm. However, the ultimate aim is to predict
the ecological consequences of anthropogenic impacts on biological species, and also to
predict the likely consequences for human welfare. This work bites a small piece of an immense puzzle.
The core of this thesis comprises four peer-review papers (I-IV) that can be found in the
Appendix section. The thesis summarises the major outcomes of the papers, and it is organised according to the IMRAD format (Introduction, Methods, Results, and Discussion). The
Introduction contains a review of the current knowledge about biodiversity and ecosystem
functioning. I was interested in illustrating mechanisms instead of describing patterns already described by others. The Methods section summarises briefly the characteristics of the
study sites, as well as the design, the set up, and the analysis of experiments. The results and
their interpretation are in the Result and Discussion section. Additionally, I provide a glossary of terms at the end of the thesis in order to help the reader to understand the mechanisms and processes mentioned in the text.
Paper I shows the results of an observational study where I compare the stability of intertidal communities with naturally differing number of species. I test the hypothesis that
stability is a positive function of species richness. In paper II, I evaluate the role of resource
complementarity as a mechanism explaining the effects of functional group richness on the
rate of resource consumption of subtidal organisms. In paper III, I test the interactive effects
of disturbances on the stability of intertidal communities. Finally, paper IV assesses the
level of replication needed to represent the number of species occurring in intertidal hardbottom communities, which may be important when analysing the relationship between diversity and ecosystem stability.
iv
Acknowledgments
Over the last three years, many friends and colleagues have contributed directly or indirectly
to this thesis. Those I have worked with in designing, setting up, and analysing experiments
have shared information and ideas. Summer interns have been enthusiast during long hours
of field work; editors and anonymous reviewers have brought part of this work to publication. Karin Boos helped me during the copy edition and printing process. Prof. Dr. Christian
Wiencke found always the way to fund materials, trips, and personnel (me). Andreas Wagner gave valuable technical assistance, and organised coffee breaks just in the best moment.
This thesis would not have been written without the constant support, encouragement, and
counsel of Dr. Markus Molis, who provided invaluable guidance and friendship. I thank all
these people.
I also thank my family for joining me on the adventure of moving to Helgoland. While
this work was being prepared, I was saddened by the loss of a member of my family:
mother-in-law Ingrid Wallberg. It is to her memory that I dedicate this contribution.
v
List of papers
This thesis is based on the following papers, which will be referred to in the text by their
roman numerals.
I
Valdivia N, Molis M (In press) Observational evidence of a negative biodiversity-
stability relationship in intertidal epibenthic communities. Aquatic Biology
II
Valdivia N, de la Haye K, Jenkins SR, Kimmance SA, Thompson R, Molis M (In
press) Functional composition, but not richness, affected the performance of sessile suspension-feeding assemblages. Journal of Sea Research
III
Valdivia N, Molis M (Under review) Species compensation buffers community stabil-
ity against the loss of an intertidal habitat-forming rockweed. Marine Ecology Progress Series
IV
Canning-Clode J, Valdivia N, Molis M, Thomason JC, and Wahl M (2008) Estimation
of regional richness in marine benthic communities: quantifying the error. Limnology and
Oceanography: Methods. 6: 580-590
vi
vii
Introduction
1
Introduction
The context: the value of biodiversity
Human activities are altering the global climate (IPCC 2007). In addition, destruction of
habitats, over harvesting, and introduction of exotic species are changing the local biodiversity of terrestrial and aquatic ecosystems (Dirzo & Raven 2003, Sax & Gaines 2003, Byrnes
et al. 2007). As a consequence, today’s species extinction rate is probably the highest in
Earth’s history (Dirzo & Raven 2003). The question therefore is not whether we are losing
species, but what the likely consequences of such biodiversity loss are. Seminal research
suggests that biodiversity influences the magnitude of and variability in ecosystems processes (reviewed by Cottingham et al. 2001, Stachowicz et al. 2007). In particular, the work
of MacArthur (1955) and Elton (1958) inspired the assertion that communities with many
interacting species are more stable than communities with fewer species. Ecologists therefore have raised the concern that changing biodiversity can impair ecosystem properties and
the goods and services provided by ecosystems, which in turn might have high societal costs
(Costanza et al. 1997, Armsworth & Roughgarden 2003).
It is not surprising therefore that the biodiversity-stability relationship had drawn the attention of ecologists for decades (Hooper et al. 2005), and that ecosystem stability had become an issue for policymakers (Christensen et al. 1996). However, the actual contribution
of biodiversity research to conservation is still under debate, because of the contrasting results of studies testing the idea that biodiversity begets ecosystem stability (Thompson &
Starzomski 2007). Specifically, there remains controversy over what constitutes a ‘richness
effect’ and how to untangle the effects on ecosystem functioning based on species richness
per se from the usually stronger effects of species identity and composition (Bruno et al.
2006).
Definitions
The exploration of biodiversity-stability relationships requires us to clarify the meaning of
biodiversity, stability, and other terms. Biodiversity is “the sum of all biotic variation in the
biosphere from the level of gene to ecosystem” (Purvis & Hector 2000). This includes, but
1
Biodiversity and stability
is not limited to, the number of species (species richness), the distribution of their abundances, and the presence or absence of key species.
The influence of biodiversity on ecosystem functioning depends on the suite of functional characteristics of the interacting species (Chapin et al. 2000). Functional traits are
those characteristics of species that influence ecosystem properties (functional effect traits)
or species’ responses to the environment (functional response traits). Functional groups are
therefore defined by either the effect of species on ecosystem functioning or their response
to the environment.
The term ecosystem functioning (or ecosystem performance) is a simple contraction for
‘how ecosystems work’, but encompasses complex mechanisms that regulate the transformation and transport of energy across the ecosystem. Ecosystem properties consist of sizes
of pools of materials like nutrients and carbon, and rates of processes like energy fluxes
across trophic levels (Christensen et al. 1996). Ecosystem goods and services are ecosystem
properties that contribute to human welfare both directly and indirectly. Food and materials
for construction are examples of ecosystem goods; nutrient cycling and buffering of coastal
erosion are examples of ecosystem services (Costanza et al. 1997). In this thesis, I use the
percent cover of benthic species as a surrogate for biomass, and filtration rates of sessile
suspension-feeding invertebrates as a surrogate for resource consumption and energy flux.
Stability has several meanings in ecology; indeed, a galaxy of definitions can be found in
the literature, and each of them can lead to a different conclusion about the biodiversitystability relationship (Grimm et al. 1992, Johnson et al. 1996). The six commonest definitions of stability are: the magnitude of disturbances a system can tolerate (domain of attraction, Holling 1973); how long a measure stays without change (persistence, Pimm 1991);
how much a measure changes by a disturbance (resistance, Pimm 1991); how long a measure needs to return to a specified fraction of its initial value (resilience, Pimm 1991); how
likely is that a system will continue functioning (reliability, Naeem 1998); and how much a
measure varies over time (variability, Pimm 1991). Pioneer biodiversity-stability researchers
explicitly considered stability to be related to temporal variability in ecosystem properties
(MacArthur 1955, Elton 1958). In accordance, I focus on the effects of biodiversity on the
temporal variance of ecosystem properties. For example, when comparing two temporal
2
Introduction
series of abundances, the “more stable” one will be that with the smallest fluctuations relative to its mean.
Theory
Temporal variability can be calculated using the variance in time series of species abundances. Because average abundances can differ, variance must be scaled relative to the mean
(Gaston & McArdle 1994). Usually, this is done using the coefficient of variation (CV = 100
V Pbeing V the standard deviation and P the mean), which decreases as stability increases. In this thesis, stability (S) is defined as S = PV(Tilman 1999). In contrast to CV,
the magnitude of S increases as stability increases; in addition, it approaches 0 when the
variation is large in relation to the mean.
On the other hand, the variance in and aggregate ecosystem property (e.g. total community abundance––the sum of the abundances of all of the species in the community) can be
expressed using a statistical rule (Schluter 1984, Doak et al. 1998):
V 2 ( xe )
V 2 ( 6 in x i )
6 in V 2 ( x i ) 2 6 in j V 2 ( x i , x j )
(1)
being xi the abundance of an individual species i, xe the aggregate community abundance
made by summing the abundance of all species, V2 the variance, V2 (xi, xj) the covariance
between species i and j, and n the number of species. Therefore, the variance of an aggregate ecosystem property depends on the sum of all species variances and the sum of all pairwise species covariances. If species vary independently, their covariance is zero and the
variance of the ecosystem property equals the summed species variances. However, when
species do not vary independently, their nonzero summed covariances cause the overall
variability to increase or decrease. Stability therefore will be defined as:
S
P
V
P
(2)
6Variance 26Covariance
In accordance with equation (2), stability will increase with increasing species richness
if the mean value of the property increases (P), or the summed variances decrease, or the
summed covariances decrease, or a combination of these occurs (Lehman & Tilman 2000).
Consequently, the factors influencing the size of ecosystem properties, the summed variances, and the summed covariances across gradients of species richness have been the focus
3
Biodiversity and stability
of theoretical work on biodiversity-stability relationships (Cottingham et al. 2001, Hooper et
al. 2005).
a
Resource A
Resource B
Resource C
800
600
400
200
0
Species a
Species b
c
2000
Mixture
of all species
Species c
b
Non-transgressive
overyielding
Transgressive
overyielding
1000
Mixture of all species
Average monoculture
Species c
Species b
Species a
Mixture of all species
Average monoculture
Species c
Species b
0
Species a
Total
consumption rate
Figure 1
Graphical depiction of resource
complementarity (a), transgressive
overyielding (b), and positive
sampling effects (c) in communities
containing 1 (species a, b, or c), and
3 (mixture) species. Total
consumption rate is the sum of the
consumption rates of all resources. If
species consume different resources
(a), positive species interactions
increase the performance of the
species-rich community relative to
any of the constituent species grown
alone (b). On the other hand, if the
species-rich community includes a
species with extreme functional
value, the performance of the
community will reflect the
performance of that species instead
of an average response of all of the
species in the community; in these
cases, the performance of the
species-rich community is larger
than that of the average
monoculture, but not larger than that
of the best-performing species
grown alone (c).
Consumption rate of resource i
1000
Size of ecosystem properties
Changes in the size of an aggregate ecosystem property can affect ecosystem stability (equation [2]). Theory predicts that ‘overyielding’, an increase in the size of an ecosystem property with increasing species richness, can result from resource partitioning or facilitation
(complementarity effect; Tilman et al. 1997a, Loreau 2000), or by the increased probability
that more diverse communities include species with extreme functional impacts (sampling
effect or positive selection effect; Huston 1997, Loreau et al. 2001). Complementarity effects (Fig. 1a) lead to the phenomenon called transgressive overyielding, in which produc4
Introduction
tivity or resource use of species-rich mixtures exceeds (transgress) that of the bestperforming species grown alone (Fig. 1b; Fridley 2001). This is because interspecific competition is reduced when species use different resources or use the same resource at different
moments or points in space. For example, sessile suspension-feeding invertebrates that consume small-sized plankton can coexist with others that consume larger particles (Gili &
Coma 1998). Species
1 species
CV =
0.3
200
150
150
100
100
50
50
0
0
60
CV =
0.19
200
150
40
100
20
0
40
50
3 species
0
5 species
CV
= 0.09
200
30
150
20
100
10
50
0
0
0
5
Total abundance
Figure 2
Simulation showing the effect of statistical
averaging on the variability in an aggregate
community property (total abundance,
heavy lines, right y-axis) made up by
summing the abundance of single species
(abundance of species i, left y-axis).
Fluctuations in species and community
abundances were simulated using Tilman’s
model (1999) where species abundances are
assumed to be independent, equally
abundant, and with the same coefficient of
variation (cv = 0.3). Average total
community abundance was held constant at
100 percent cover. Statistical averaging of
individual fluctuations dampens the
variability of the aggregate community
property as the number of species increases
(note decreasing CV for total abundance).
Abundance of species i
200
10
Time
15
20
of benthic algae can share the resource ‘space’ and minimise competition: seaweeds need a
relatively small area in the substratum in order to stay attached, but they are still able to develop a large canopy over an area where the substratum was monopolised by encrusting
forms or turf-forming algae (e.g. Connell 2003). In addition, canopies provide settlement
substratum for other smaller species, such as filamentous algae and sessile invertebrates.
5
Biodiversity and stability
Positive selection effects occur when the performance (e.g. resource use or productivity)
of the most efficient species explains that of the entire community (Fig. 1c; Ives et al. 2005).
For example, plant communities are usually dominated by individuals of the largest species
(e.g. Polley et al. 2007). Therefore, most of the biomass in a species-rich community may be
contributed by one or few dominant species and reflects the biomass of those species instead
of an average value of all species present in the community (Huston 1997).
100
cor = 0.95
CV
= 0.28
200
150
50
100
50
0
100
5
10
15
cor = 0
20
CV
= 0.21
200
150
50
100
50
0
0
0
100
Total abundance
0
Abundance of species i
Figure 3
Graphical depiction of the effect of
the covariance in species fluctuations
on the variability an aggregate
community property (total
abundance, heavy lines, right y-axis)
made up by summing the abundance
of single species (abundance of
species i, left y-axis) in communities
of 2 species (either solid or dashed
lines). Fluctuations in species
abundances were simulated by
generating 20 pairs of normal random
values with mean 50, coefficient of
variation 0.3, and specified
coefficient of correlation (cor = 0.95,
almost perfect positive correlation; 0,
no correlation, -0.95 almost perfect
negative correlation). Average total
community abundance was held
constant at 100 percent cover.
Negative pair-wise species
covariance dampens the variability of
the aggregate community property
(note decreasing CV for total
abundance).
0
5
10
15
cor = -0.95
20
CV
= 0.05
200
150
100
50
50
0
0
0
5
10
15
20
Time
Variance in species properties
As expressed in equations (1) and (2), stability in an aggregate ecosystem property like total
community abundance will be affected by the variances and covariances in species abun6
Introduction
dances. Statistical averaging, also called the portfolio effect, is a pivotal mechanisms leading
to a negative relationship between species richness and the variability of aggregate ecosystem properties (Doak et al. 1998, Tilman et al. 1998). When total community abundance is
the sum of the abundances of many species, each varying over time, then adding more species together will increase the probability that the fluctuations in these individual abundances will average out statistically (Fig. 2; Doak et al. 1998). This reduces the variability in
the aggregate property in relation to that of the average individual abundances. Accordingly,
whenever species fluctuations are not perfectly correlated, increasing species richness will
reduce the variability of the community mainly on statistical grounds.
However, because asynchrony among species results from differential environmental
tolerances, statistical averaging is due in part to ecological difference among species
(Cottingham et al. 2001). Moreover, the strength of statistical averaging seems to be
strongly affected by the relative abundance of species. For example, high dominance of few
species can dampen the richness-stability relationship (Doak et al. 1998), and lead to negative and curvilinear functions (Lhomme & Winkel 2002). Therefore, ecological processes
leading to heterogeneity and temporal asynchrony in species abundances influence the effect
of biodiversity on ecosystem stability.
These ecological processes affect also the covariances in species abundances, which in
turn influence the stability of the community (equations [1] and [2]). Pairs of species competing for the same resource or with differing abilities to respond to the environment should
show compensatory responses, such that when the abundance of one species increases and
that of the other decreases; the resulting negative covariance decreases the variability in the
total community abundance (Fig. 3; Doak et al. 1998, Yachi & Loreau 1999). If the variety
of environmental tolerances increases as species richness increases, then adding more species will increase the probability that some species compensate the function of other that
failed due to changes in the environment (Yachi & Loreau 1999, Ives et al. 2000). Therefore, maintenance of species with different functional response traits can be crucial for ecosystem stability. Surprising, conservation managers usually concentrate on rare species that
may be unable to compensate the loss of a dominant species (Thompson & Starzomski
2007).
7
Biodiversity and stability
Observations and experiments
Observational and manipulative experiments support the idea that increasing species richness leads to increasing ecosystem stability (e.g. McNaughton 1985, Tilman & Downing
1994, Naeem & Li 1997, Ptacnik et al. 2008), but most of these studies are confounded by
other variables (Hooper et al. 2005). For example, to demonstrate that biodiversity increases
drought resistance in grasslands, Tilman and Downing (1994) altered plant species richness
by using nutrient additions. However, increased resistance could have resulted either from
species compensation (Tilman 1996) or from differences in species composition caused by
the fertilisations (Huston 1997). In the observational study of McNaughton (1985) on Serengeti’s grasslands, the negative correlation between proportional diversity (H’) and community variability was probably influenced by differences in species composition and
abiotic conditions between sites.
In order to detect confounding effects of species richness and composition, investigating
the relationship between biodiversity and ecosystem functioning and stability should be
complemented by different experimental methods, including assembling communities in
controlled environment, manipulating diversity in the field, and observing patterns in nature
(Díaz et al. 2003). There is no single best method, as not all questions can be addressed
equally well by these three approaches. For example, the effects of species richness per se
are better addressed by synthetic-assemblages experiments, because of the greater control of
species composition across replicates and levels of species richness (e.g. Loreau & Hector
2001, Benedetti-Cecchi 2004). In these experiments, random selections of species or functional groups are assembled to generate different diversity treatments. However, because
such experiments do not represent how natural communities are assembled or dissembled––
species extinction are rarely random, for example––, the interpretation of studies using synthetic communities is difficult (Wardle et al. 2000, Loreau et al. 2001).
On the other hand, the effects of non-random species extinctions are better addressed by
removal experiments, in which target species or functional groups are removed from the
natural community (e.g. Wardle et al. 1999, O'Connor et al. 2008). As they are based on
naturally assembled communities, removal experiments allow incorporating the effects of
large-scale processes like variations in climate, disturbance regime, and biotic interactions
8
Introduction
on the regional species pool (Belyea & Lancaster 1999). However, the cost of manipulating
diversity in the field restricts the size and duration of removal experiments, which in turn
limits the interpretation of results to the local characteristics or context. Alternatively, observational studies allow using larger spatial and time scales, and broader ranges of species
usually used in manipulative experiments (e.g. Troumbis & Memtsas 2000). Observational
studies require to be carefully designed to account for the diversity of the sites under investigation. The number of species is related to sampling effort (Ugland et al. 2003), so proper
replication may be especially important in observational studies linking ecosystem stability
and species richness.
The model system: hard-bottom ecosystems
Intertidal and shallow subtidal rocky habitats offer potential to test the diversity-stability
relationship. In these habitats, abiotic stressors change and biological processes occur at
small temporal and spatial scales (Underwood & Chapman 1996). As a consequence, intertidal and shallow subtidal assemblages represent tractable experimental systems at the landscape scale and small-scale experiments are usually appropriate (Giller et al. 2004).
An example of the potential of rocky shores for biodiversity research is given by the relationship between community structure and canopy-forming species. These species modify
the environment so that it becomes more suitable for some species, but less suitable for others (e.g. Irving & Connell 2006, Lilley & Schiel 2006, Morrow & Carpenter 2008). In subtidal habitats, for example, canopy loss reduces the abundance of species adapted to shaded
conditions (e.g. encrusting coralline algae), yet it allows the increase of species adapted to
more exposed conditions (Irving & Connell 2006). In intertidal habitats, however, the response of understorey communities to canopy disturbances is still unclear (Lilley & Schiel
2006).
Aims
The aim of this study was to determinate the effects of species richness and species composition on the stability of intertidal and shallow subtidal hard-bottom communities. A combi-
9
Biodiversity and stability
nation of manipulative and observational approaches was used to address different but complementary hypotheses.
In an observational study, I tested the hypothesis that (1) ecosystem stability is positively
related to species richness; in a synthetic assemblage experiment, I tested the hypotheses
that (2) different functional groups use different resources and that (3) increasing number of
functional groups increases the efficiency of resource consumption of the assemblage (i.e.
functional richness leads to transgressive overyielding in filtration rate); in a removal experiment I tested the hypotheses that (4) the loss of a key canopy-forming species affects the
stability of the community and that (5) the effects of canopy removal depend on the presence of mechanical disturbances that provide free space. Finally, analysing species abundance data (here after referred to as the richness-estimation study), I tested whether the
sampling effort of the observational study was enough to represent the number of species of
the shores here studied. The results, reported as peer-review papers and referred to by their
roman numerals (I-IV, see Appendix), suggest that species composition and identity had far
stronger effects on ecosystem stability than species richness per se.
10
Methods
2
Methods
Study sites
Observational and manipulative studies were conducted at Helgoland Island, NE Atlantic,
between March 2006 and March 2008. The observational study, removal experiment, and
richness-estimation study (papers I, III, and IV) were conducted in the mid-low intertidal
zone. This zone is characterised by canopy-forming algae (e.g. Fucus spp. and Laminaria
digitata), turf-forming algae (e.g. Ceramium virgatum, Chondrus crispus, Cladophora
rupestris, Corallina officinalis, Mastocarpus stellatus), and encrusting algae (e.g. Phymatolithon spp.). Frequent sessile invertebrates are Dynamena pumila, Spirorbis spirorbis, and
Electra pilosa, and conspicuous mobile consumers include Carcinus maenas and several
species of periwinkles. Temporal patterns in community structure have an important seasonal component, as ephemeral algae like Ulva spp. and seasonal Cladophorales (e.g.
Cladophora sericea) become abundant during summer (Janke 1990). In the observational
study, stability was compared across five sites with naturally different number of species.
Each site was of ca. 200 m2 and adjacent sites were 100 m apart from each other. The removal experiment was replicated at two intertidal sites with differing degree of wave exposure in order to test for generality of findings. Finally, the richness-estimation study was
based on data from the northern intertidal area of Helgoland.
The synthetic-assemblage experiment (paper II) was conducted in the shallow subtidal
habitat, and sessile invertebrates growing on vertical surfaces were used as experimental
organisms. This assemblage is characterised by mussels (e.g. Mytilus edulis), ascidians (e.g.
Ciona intestinalis, Ascidiella aspersa, and Diplosoma listerianum), barnacles (e.g. Elminius
modestus and Balanus crenatus), and bryozoans like Cryptosula pallasiana, and Membranipora membranacea (Anger 1978, Wollgast et al. 2008).
Sampling and experimental designs
In the observational study, I compared community stability across the five sample sites.
Temporal variances in species abundances were calculated from repeated estimations of
species percent covers on plots of 0.25 m2 that were marked with stainless screws and sampled every 6 months for 24 months. The temporal variance and temporal mean of total
11
Biodiversity and stability
community cover (cover summed across all species in a sampling unit) were used to calculate the S index of stability (S = PV-1). On the other hand, species-accumulation curves were
used to calculate the number of species occurring at each sample site. I used regression
analyses to test the hypothesis that community stability is positively related to species richness. Additionally, I partitioned the temporal variance into the sum of all species variance
and the sum of all pair-wise species covariances as in equation (1). The summed covariances
were used as a measure of compensatory dynamics (see Theory section) and also to test
whether increases in the variance of species abundances are counterbalanced by increasingly
negative species covariances.
In the removal experiment (paper III), I tested the separate and interactive effects of the
removal of the canopy-forming alga Fucus serratus and mechanical disturbances on community stability. As in the observational study, temporal variances were obtained from repeated measures of species covers, but plots were of 0.09 m2 and sampled every 3 months
for 18 months. Fucus plants were removed used a knife and mechanical disturbance treatments consisted of a biomass removal with 50 % of the effort required to remove all organisms of the plot. I analysed two aspects of community stability: the temporal variability at
the community level (i.e. temporal variance of community total cover), and the temporal
variability at the species level (i.e. summed covariances and Bray-Curtis1 index).The effects
of canopy removal, disturbance, and site on all of the measures of temporal variability were
analysed using 3-way mixed analyses of variance (ANOVAs) with the factors Fucus canopy
(2 levels: present or removed) and disturbance (2 levels: undisturbed or disturbed) considered fixed and the factor site (2 levels, Nordostwatt or Westwatt) considered random.
The synthetic-assemblage experiment (paper II) was designed to test the effect of the
number of functional groups on filtration rates of suspension-feeding invertebrates, and to
separate this effect from that of functional identity. The design consisted of 3 functional
groups growing alone (i.e. mono-specific assemblages of mussels, colonial ascidians and
bryozoans, and barnacles) and one complete mixture containing all groups. Organisms used
1
The Bray-Curtis index (BC) measures the variability between samples in terms of species abundances. The
advantage of this method is that ignores the ‘double zeros’; i.e., it downplays the similarity between samples in
which the same species is absent. BC works well on ecological data, which are usually plagued by zeros.
12
Methods
to construct the assemblages were obtained by exposing artificial substrata (3 × 25 × 25 mm
PVC tiles) to colonisation in the water column for 7 months, starting on May 2006 to match
with the main recruitment period of epibenthic species. After that, tiles containing one functional group each were used to construct the experimental assemblages (Stachowicz et al.
2002).
The synthetic assemblages were maintained in the field (1 m depth), but filtration rate
assays were conducted in the laboratory. Filtration rate was measured as the volume of water cleared per unit of time from a mixed-culture microalgal suspension. Four microalgae of
different size were used as food in order to allow for resource complementarity in terms of
particle size. A cytometric technique allowed the identification of each species of microalgae, and so for testing whether functional types consumed different resources. ANOVA and
planned contrasts were used to tease apart the effects of functional richness (richness effect)
from those of each functional group (identity effect). The hypothesis that functional richness
leads to overyielding was tested using a planned contrast between the filtration rate of the
mixture and the average filtration rate of the monocultures (see Fig. 1b-c). The hypothesis
that functional richness leads to transgressive overyielding was tested by comparing the
performance of the mixture and the best-performing functional group in monocultures (see
Fig. 1b-c). Permutational multivariate analysis of variance (PERMANOVA) was used to test
for resource partitioning among functional groups on the basis of consumer-specific changes
in the multivariate structure of prey (Fig. 1a).
For the richness-estimation study (paper IV), I quantified the abundance of species occurring on fifty-two 0.04 m2 replicate plots in spring 2006. Species-accumulation curves
were used to calculate the number of species in the maximum number of quadrats. Then, the
probable regional richness was estimated by fitting a curvilinear growth model that provides
the asymptotic number of species as the number of replicates approaches infinity (Morgan et
al. 1975).
13
Biodiversity and stability
3
Results and Discussion
Biodiversity, broadly defined, significantly influenced the magnitude and variability of ecosystem properties such as community biomass (measured as percent cover) and resource
consumption (measured as filtration rate). Nevertheless, the effects of species composition
seemed to be more important than those of species richness. Contrarily to our predictions,
the observational study showed a negative and curvilinear diversity-stability relationship. In
the synthetic-assemblage experiment (paper II), filtration rates differed significantly among
functional groups grown alone, but their average filtration rate did not differ from that of the
high-diversity mixtures––i.e. the presence of more functional groups did not increase filtration rate. Finally, the removal of the canopy-forming alga Fucus serratus increased the variability of species without affecting the variability of communities (paper III). Compensatory
dynamics, such that the abundance of some species increases while that of other decreases,
buffered the community-level stability against the environmental changes caused by the
canopy removal––such patters were consistent across two sites.
Collectively, these results agree with biodiversity studies on marine macroalgal (Bruno
et al. 2006), terrestrial plant (Hooper et al. 2005), and freshwater communities (Downing
2005, Weis et al. 2008). These previous studies have shown that richness effects are actually
subtle and that compositional effects are strong. The loss or gain of particular species therefore may have a stronger effect on ecosystem stability than species richness per se. Therefore, predicting the consequences of biodiversity loss remains complicated, because it requires an accurate knowledge of the system and natural life history and should be drawn
from sound experimental evidence, not from generalised models.
Species richness vs. species composition
Overyielding, or an increase in an ecosystem property with increasing species richness, was
detected in the observational study (paper I). Theory predicts that overyielding is due to
resource complementarity (Tilman et al. 1997b), which may occur in benthic communities.
Epibenthic species can partition the available space by forming multilayered spatial structures––the experiments conducted on the intertidal areas (papers I, III, and IV) showed that
14
Results and discussion
encrusting forms, turf-forming, and canopy-forming algae formed three biotic layers, and
the observational study showed that this layering tended to increase across the richness gradient. Multilayered structure is a characteristic of benthic communities made up of numerous species (Bruno et al. 2003), so resource complementarity in terms of differential use of
the space might be widespread among these communities.
However, overyielding might well have resulted from the increased probability that species-rich communities included particular (“key”) species with strong effects on community
abundance––i.e. positive sampling effects. Indeed, few species dominated the communities,
so the apparent pattern of the community abundance could well have reflected those of the
dominant species instead of an average response of all species in the community. Moreover,
one of the dominant species, the canopy former Fucus serratus, had significant effects on
the species composition and stability (paper III); so the relationships between site species
richness, community abundance, and stability (paper I) would have been strongly influenced
by changes in the abundance of this species. On the other hand, the negative and curvilinear
richness-stability relationship (paper I) may have resulted from the dominance of species––
experiments have shown that the stability of communities dominated by few species is
driven by these particular taxa (Steiner et al. 2005, Polley et al. 2007), and simulations suggest that strong heterogeneity among species abundances may lead to negative and nonlinear relationships between species richness and stability (Lhomme & Winkel 2002).
In consequence, even when complementarity in the use of space may be common in
natural benthic communities, there is a fair chance that selection effects also operate within
these assemblages. Both, selection effects and positive species interactions (including complementarity and facilitation) can act simultaneously or sequentially (Hooper et al. 2005,
Bruno et al. 2006). The challenge is therefore to develop analytical tools that allow quantifying the relative contribution of each of these mechanisms to ecosystem function (e.g. Loreau
& Hector 2001, Fox 2005).
The synthetic-assemblage experiment (paper II), provided the opportunity to test
whether resource complementarity occurs within subtidal suspension feeders. The experiment was replicated at two locations in NE Atlantic coasts, and the results from both locations suggest that complementarity was actually absent: the high efficiency of mussels in
filtrating most of the phytoplankton species suggests that filtration rate of the mixtures was
mostly due to the activity of this functional group, which may have prevented resources
15
Biodiversity and stability
complementarity and led to no richness effects. These results, complemented with the correlative evidence of the observational study, suggest that species identity and composition
may have strong effects on the functioning and stability of natural community and that the
consequences of biodiversity changes can not be predicted from the number of species that
are loss or gained. Identity effects may be common within benthic communities, as suggested by recent experiments conducted on intertidal (O'Connor et al. 2008) and shallow
subtidal communities (Bruno et al. 2006, O'Connor & Bruno 2007), as well as reviews and
meta analyses of published datasets (Cardinale et al. 2006, Stachowicz et al. 2007).
Because the synthetic-assemblage experiment did not represent how natural communities are assembled, its interpretation in a real scenario of biodiversity change may become
difficult. For example, whether species are assembled as larvae and juveniles or adults can
influence the outcome of synthetic experiments (Garnier et al. 1997). The removal experiment (paper III) tested in natural conditions what happened with the community and species
stability when a particular species went locally extinct. Interesting, the results of this ‘natural’ experiment also suggest that the functional characteristics of species affect stability, albeit it was not designed to tease richness effects apart from identity effects.
In the removal experiment, a single species had strong effects on composition and stability of species. The removal of Fucus serratus significantly influenced the physical surroundings of the remaining species, as shown in other communities where canopies ameliorate
stressors like temperature and water evaporation (Bertness et al. 1999, Moore et al. 2007),
and also where canopies reduce sedimentation (Kennelly & Underwood 1993). In my experiment, these changes had negative effects on species sensitive to sedimentation and osmotic stress, such as encrusting algae and small sessile invertebrates, yet they had positive
effects on species more tolerant, such as ephemeral green algae. Therefore, disturbance can
have differing effects on species, which may play an important role in maintaining the stability of the community (Micheli et al. 1999).
16
Results and discussion
Species’ response traits influence community stability
The compositional changes caused by the canopy removal reduced the stability of species,
but, these severe disturbances did not affect the stability of communities. Probably, the
negative covariance resulting from the compensation between sensitive and tolerant species
maintained the stability of the community. This also may explain why the additional mechanical disturbances significantly decreased species stability only in removal plots, without
affecting the community stability. Mechanical disturbances can have significant effects on
species richness and composition (Valdivia et al. 2008), species coexistence (Shea et al.
2004), and stability (Bertocci et al. 2007). So, species compensation can maintain the stability of communities in the face of strong environmental disturbances. The role of species
compensation in buffering community stability against stochastic change has been shown in
mathematical simulations (Fig. 3; Doak et al. 1998, Yachi & Loreau 1999) and field observations (Ernest & Brown 2001).
The importance of negative covariances and compensatory dynamics was also noted in
the observational study (paper I), where the lack of correlation between the summed covariances and species richness contributed to the negative richness-stability relationship found–
–theory predicts that increasingly negative covariances should offset the increases in species
variances as more species are present (equation [2]; Tilman et al. 1998). Therefore, even
when the number of species seemed to have little effects on ecosystem functioning, the results from the intertidal experiments agree with the assertion that the presence of a variety of
responses to the environment is fundamental in maintaining community stability (Walker
1992, Yachi & Loreau 1999).
The role of replication in biodiversity experiments
In the richness-estimation study (paper IV), by extrapolating species-accumulation curves
we predicted a probable regional richness similar to the maximum number of species quantified in the observational study (65 vs. 72). This suggests that the sampling effort in the latter
was enough to represent the number of species occurring on mid-low intertidal areas of Helgoland. On the other hand, comprehensive inventories of species suggest that 53 (Janke
1986, Reichert & Buchholz 2006) sessile and 39 macroalgal (I. Bartsch, unpubl. data) spe17
Biodiversity and stability
cies can occur in these shores. According to these values, our extrapolations of the speciesaccumulation curves are clearly below the actual number of species. However, these extensive inventories included probably species occurring in different year and habitats, so these
species do not necessarily coexist at the local scale.
In observational studies linking species richness and the variability in ecosystem properties, the level of replication may be particularly important: first, assuring a proper replication can be critical for reducing unwanted variability derived from spatial and temporal
patchiness in species distributions (Cottingham et al. 2001). Second, account of rare species
might be important when rare species have disproportional effects on ecosystem properties
(e.g. keystone species; Lyons et al. 2005). Therefore, the ability of an experimental design to
detect compositional effects on ecosystem function can depend on sample size (Allison
1999).
Conclusion
Species composition seemed to be more important for the stability of this ecosystem than
the number of species and functional groups. Consequently, predicting the consequences of
the widespread human-driven changes in biodiversity needs an accurate knowledge on life
history and biology of species. So, descriptive work on basic life history traits is fundamental in this context. On the other hand, we should not assume that mechanisms predicted by
theory to lead to positive richness-stability (and functioning) relationships are unimportant
in the systems here studies. Resource complementarity influences species coexistence
(Ricklefs 1990), and the impact of species richness on ecosystem properties can grow
stronger through succession (Cardinale et al. 2004, Cardinale et al. 2007).
Further research should address the influence of the relative abundance of species (i.e.
evenness) and different types of disturbances on the relationship between biodiversity and
ecosystems stability, in addition to the occurrence of species compensation under different
levels of environmental stress. We still need to unravel the relationship between taxonomic
and functional diversity. Identifying those traits of species that influence ecosystem properties and species’ responses to the environment requires us to assess the impacts of environ18
Results and discussion
mental disturbances at the levels of communities, populations, and organisms, and to investigate the variations in physiological traits across geographical scales (e.g. Dahlhoff &
Menge 1996, Chown & Gaston 2008).
Management of natural communities is generally based on the conservational status of
species; that is, species are usually managed if they are endangered or introduced. However,
conservation managers only rarely consider the functional effects of species (Thompson &
Starzomski 2007). According to the results of this thesis, conservational efforts should be
directed to identify the functional traits that make species important for the functioning and
stability of ecosystems. Key functional traits should be conservation priorities. Finally, managers should assure that natural communities contain many species with different functional
responses and also many species with similar functional effects. This will allow species
compensation in the face of rapid environmental changes.
19
Biodiversity and stability
References
Allison GW (1999) The implications of experimental design for biodiversity manipulations. Am Nat 153:2645
Anger K (1978) Development of a subtidal epifaunal community at the island of Helgoland. Helg Mar Res
31:457-470
Armsworth PR, Roughgarden JE (2003) The economic value of ecological stability. P Natl Acad Sci USA
100:7147-7151
Belyea LR, Lancaster J (1999) Assembly rules within a contingent ecology. Oikos 86:402-416
Benedetti-Cecchi L (2004) Increasing accuracy of causal inference in experimental analyses of biodiversity.
Funct Ecol 18:761-768
Bertness D, Leonard G, Levine J, Schmidt P, Ingraham A (1999) Testing the relative contribution of positive
and negative interactions in rocky intertidal communities. Ecology 80:2711-2726
Bertocci I, Vaselli S, Maggi E, Benedetti-Cecchi L (2007) Changes in temporal variance of rocky shore organism abundances in response to manipulation of mean intensity and temporal variability of aerial exposure. Mar Ecol Prog Ser 338:11-20
Bruno JF, Lee SC, Kertesz JS, Carpenter RC, Long ZT, Duffy JE (2006) Partitioning the effects of algal species identity and richness on benthic marine primary production. Oikos 115:170-178
Bruno JF, Stachowicz JJ, Bertness D (2003) Inclusion of facilitation into ecological theory. Trends Ecol Evol
18:119-125
Byrnes JE, Reynolds PL, Stachowicz JJ (2007) Invasions and extinctions reshape coastal marine food webs.
PLoS ONE 2:e295
Cardinale BJ, Ives AR, Inchausti P (2004) Effects of species diversity on the primary productivity of ecosystems: extending our spatial and temporal scales of inference. Oikos 104:437-450
Cardinale BJ, Srivastava DS, Duffy JE, Wright JP, Downing AL, Sankaran M, Jouseau C (2006) Effects of
biodiversity on the functioning of trophic groups and ecosystems. Nature 443:989-992
Cardinale BJ, Wrigh JP, Cadotte MW, Carroll IT, Hector A, Srivastava DS, Loreau M, Weis JJ (2007) Impacts
of plant diversity on biomass production increase through time because of species complementarity. P
Natl Acad Sci USA 104:18123-18128
Chapin FS, Zavaleta ES, Eviner VT, Naylor RL, Vitousek PM, Reynolds HL, Hooper DU, Lavorel S, Sala OE,
Hobbie SE, Mack MC, Diaz S (2000) Consequences of changing biodiversity. Nature 405:234-242
Chown SL, Gaston KJ (2008) Macrophysiology for a changing world. P R Soc B 275:1469-1478
Christensen NL, Bartuska AM, Brown JH, Carpenter S, D'Antonio C, Francis R, Franklin JF, MacMahon JA,
Noss RF, Parsons DJ, Peterson CH, Turner MG, Woodmansee RG (1996) The report of the Ecological
Society of America committee on the scientific basis for ecosystem management. Ecol Appl 6:665691
Connell SD (2003) The monopolization of understorey habitat by subtidal encrusting coralline algae: a test of
the combined effects of canopy-mediated light and sedimentation. Mar Biol 142:1065-1071
Costanza R, dArge R, deGroot R, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, ONeill RV, Paruelo J,
Raskin RG, Sutton P, van den Belt M (1997) The value of the world's ecosystem services and natural
capital. Nature 387:253-260
Cottingham KL, Brown BL, Lennon JT (2001) Biodiversity may regulate the temporal variability of ecological
systems. Ecol Lett 4:72-85
20
References
Dahlhoff EP, Menge BA (1996) Influence of phytoplankton concentration and wave exposure on the ecophysiology of Mytilus californianus. Mar Ecol Prog Ser 144:97-107
Díaz S, Symstad AJ, Chapin FS, Wardle DA, Huenneke LF (2003) Functional diversity revealed by removal
experiments. Trends Ecol Evol 18:140-146
Dirzo R, Raven PH (2003) Global state of biodiversity and loss. Annu Rev Env Resour 28:137-167
Doak DF, Bigger D, Harding EK, Marvier MA, O'Malley RE, Thomson D (1998) The statistical inevitability
of stability-diversity relationships in community ecology. Am Nat 151:264-276
Downing AL (2005) Relative effects of species composition and richness on ecosystem properties in ponds.
Ecology 86:701-715
Elton SC (1958) The ecology of invasions by animals and plants. The University of Chicago Press, Chicago,
USA
Ernest SKM, Brown JH (2001) Homeostasis and compensation: the role of species and resources in ecosystem
stability. Ecology 82:2118-2132
Fox JW (2005) Interpreting the 'selection effect' of biodiversity on ecosystem function. Ecol Lett 8:846-856
Fridley JD (2001) The influence of species diversity on ecosystem productivity: how, where, and why? Oikos
93:514-526
Garnier E, Navas M-L, Austin MP, Lilley JM, Gifford RM (1997) A problem for biodiversity-productivity
studies: how to compare the productivity of multispecific plant mixtures to that of monocultures?
Acta Oecologica 18:657-670
Gaston KJ, McArdle BH (1994) The temporal variability of animal abundances - measures, methods and patterns. Philos T Roy Soc B 345:335-358
Gili JM, Coma R (1998) Benthic suspension feeders: their paramount role in littoral marine food webs. Trends
Ecol Evol 13:316-321
Giller PS, Hillebrand H, Berninger U-G, Gessner MO, Hawkins S, Inchausti P, Inglis C, Leslie H, Malmqvist
B, Monaghan MT, Morin PJ, O'Mullan G (2004) Biodiversity effects on ecosystem functioning:
emerging issues and their experimental test in aquatic environments. Oikos 104:423-436
Grimm V, Schmidt E, Wissel C (1992) On the application of stability concepts in ecology. Ecol Model 63:143161
Holling CS (1973) Resilience and stability of ecological systems. Annu Rev Ecol Syst 4:1-23
Hooper DU, Chapin I, F. S., Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM, Loreau M,
Naeem S, Schmid B, Setälä H, Symstad AJ, Vandermeer J, Wardle DA (2005) Effects of biodiversity
on ecosystem functioning: a consensus of current knowledge. Ecol Monogr 75:3-35
Huston MA (1997) Hidden treatments in ecological experiments: re-evaluating the ecosystem function of biodiversity. Oecologia 110:449-460
IPCC (2007) Climate Change 2007: Synthesis Report. Cambridge University Press, Cambridge
Irving A, Connell S (2006) Predicting understorey structure from the presence and composition of canopies: an
assembly rule for marine algae. Oecologia 148:491-502
Ives AR, Cardinale BJ, Snyder WE (2005) A synthesis of subdisciplines: predator-prey interactions, and biodiversity and ecosystem functioning. Ecol Lett 8:102-116
Ives AR, Klug JL, Gross K (2000) Stability and species richness in complex communities. Ecol Lett 3:399-411
Janke K (1986) Die Makrofauna und ihre Verteilung im Nordost-Felswatt von Helgoland. Helg Mar Res 40:155
Janke K (1990) Biological interactions and their role in community structure in the rocky intertidal of Helgoland (German Bight, North Sea). Helg Mar Res 44:219-263
21
Biodiversity and stability
Johnson KH, Vogt KA, Clark H, Schmitz O, Vogt D (1996) Biodiversity and the productivity and stability of
ecosystems. Trends Ecol Evol 11:372-377
Kennelly SJ, Underwood AJ (1993) Geographic consistencies of effects of experimental physical disturbance
on understorey species in sublittoral kelp forests in central New South Wales. J Exp Mar Biol Ecol
168:35-58
Lehman CL, Tilman D (2000) Biodiversity, stability, and productivity in competitive communities. Am Nat
156:534-552
Lhomme JP, Winkel T (2002) Diversity-stability relationships in community ecology: Re-examination of the
portfolio effect. Theor Popul Biol 62:271-279
Lilley S, Schiel D (2006) Community effects following the deletion of a habitat-forming alga from rocky marine shores. Oecologia 148:672-681
Loreau M (2000) Biodiversity and ecosystem functioning: recent theoretical advances. Oikos 91:3-17
Loreau M, Hector A (2001) Partitioning selection and complementarity in biodiversity experiments. Nature
412:72-76
Loreau M, Naeem S, Inchausti P, Bengtsson J, Grime JP, Hector A, Hooper DU, Huston MA, Raffaelli D,
Schmid B, Tilman D, Wardle DA (2001) Biodiversity and ecosystem functioning: current knowledge
and future challenges. Science 294:804-808
Lyons KG, Brigham CA, Traut BH, Schwartz MW (2005) Rare species and ecosystem functioning. Conserv
Biol 19:1019-1024
MacArthur RH (1955) Fluctuations of animal populations and a measure of community stability. Ecology
36:533-536
McNaughton SJ (1985) Ecology of a grazing ecosystem: the Serengeti. Ecol Monogr 55:259-294
Micheli F, Cottingham KL, Bascompte J, Bjornstad ON, Eckert GL, Fischer JM, Keitt TH, Kendall BE, Klug
JL, Rusak JA (1999) The dual nature of community variability. Oikos 85:161-169
Moore P, Hawkins SJ, Thompson RC (2007) Role of biological habitat amelioration in altering the relative
responses of congeneric species to climate change. Mar Ecol Prog Ser 334:11-19
Morgan PH, Mercer LP, Flodin NW (1975) General model for nutritional responses of higher organisms. P
Natl Acad Sci USA 72:4327-4331
Morrow K, Carpenter R (2008) Shallow kelp canopies mediate macroalgal composition: effects on the distribution and abundance of Corynactis californica (Corallimorpharia). Mar Ecol Prog Ser 361:119-127
Naeem S (1998) Species redundancy and ecosystem reliability. Conserv Biol 12:39-45
Naeem S, Li SB (1997) Biodiversity enhances ecosystem reliability. Nature 390:507-509
O'Connor NE, Bruno JF (2007) Predatory fish loss affects the structure and functioning of a model marine
food web. Oikos 116:2027-2038
O'Connor NE, Grabowski JH, Ladwig LM, Bruno JF (2008) Simulated predator extinctions: Predator identity
affects survival and recruitment of oysters. Ecology 89:428-438
Pimm SL (1991) The balance of nature?: ecological issues in the conservation of species and communities.
The University of Chicago Press, Chicago
Polley HW, Wilsey BJ, Derner JD (2007) Dominant species constrain effects of species diversity on temporal
variability in biomass production of tallgrass prairie. Oikos 116:2044-2052
22
References
Ptacnik R, Solimini AG, Andersen T, Tamminen T, Brettum P, Lepisto L, Willen E, Rekolainen S (2008) Diversity predicts stability and resource use efficiency in natural phytoplankton communities. P Natl Acad
Sci USA 105:5134-5138
Purvis A, Hector A (2000) Getting the measure of biodiversity. Nature 405:212-219
Reichert K, Buchholz F (2006) Changes in the macrozoobenthos of the intertidal zone at Helgoland (German
Bight, North Sea): a survey of 1984 repeated in 2002. Helg Mar Res 60:213-223
Ricklefs RE (1990) Ecology. W. H. Freeman and Company, New York
Sax DF, Gaines SD (2003) Species diversity: from global decreases to local increases. Trends Ecol Evol
18:561-566
Schluter D (1984) A variance test for detecting species associations, with some example applications. Ecology
65:998-1005
Shea K, Roxburgh SH, Rauschert SJ (2004) Moving from pattern to process: coexistence mechanisms under
intermediate disturbance regimes. Ecol Lett 7:491-508
Stachowicz JJ, Bruno JF, Duffy JE (2007) Understanding the effects of marine biodiversity on communities
and ecosystems. Annu Rev Ecol Evol S 38:739-766
Stachowicz JJ, Fried H, Osman RW, Whitlatch RB (2002) Biodiversity, invasion resistance, and marine ecosystem function: Reconciling pattern and process. Ecology 83:2575-2590
Steiner CF, Long ZT, Krumins JA, Morin PJ (2005) Temporal stability of aquatic food webs: partitioning the
effects of species diversity, species composition and enrichment. Ecol Lett 8:819-828
Thompson R, Starzomski BM (2007) What does biodiversity actually do? A review for managers and policy
makers. Biodivers Conserv 16:1359-1378
Tilman D (1996) Biodiversity: population versus ecosystem stability. Ecology 77:350–363
Tilman D (1999) The ecological consequences of changes in biodiversity: a search for general principles. Ecology 80:1455–1474
Tilman D, Downing JA (1994) Biodiversity and stability in grasslands. Nature 367:363-365
Tilman D, Knops J, Wedin D, Reich P, Ritchie M, Siemann E (1997a) The influence of functional diversity and
composition on ecosystem processes. Science 277:1300-1302
Tilman D, Lehman CL, Bristow CE (1998) Diversity-stability relationships: Statistical inevitability or ecological consequence? Am Nat 151:277-282
Tilman D, Lehman CL, Thomson KT (1997b) Plant diversity and ecosystem productivity: Theoretical considerations. P Natl Acad Sci USA 94:1857-1861
Troumbis AY, Memtsas D (2000) Observational evidence that diversity may increase productivity in Mediterranean shrublands. Oecologia 125:101-108
Ugland KI, Gray JS, Ellingsen KE (2003) The species-accumulation curve and estimation of species richness.
J Anim Ecol 72:888-897
Underwood AJ, Chapman MG (1996) Scales of spatial patterns of distribution of intertidal invertebrates.
Oecologia 107:212-224
Valdivia N, Stehbens JD, Hermelink B, Connell SD, Molis M, Wahl M (2008) Disturbance mediates the effects
of nutrients on developing assemblages of epibiota. Austral Ecol 33:951-962
Walker BH (1992) Biodiversity and ecological redundancy. Conserv Biol 6:18-23
Wardle DA, Bonner KI, Barker GM, Yeates GW, Nicholson KS, Bardgett RD, Watson RN, Ghani A (1999)
Plant removals in perennial grassland: vegetation dynamics, decomposers, soil biodiversity, and ecosystem properties. Ecol Monogr 69:535-568
23
Biodiversity and stability
Wardle DA, Huston MA, Grime JP, Berendse F, Garnier E, Laurenroth WK, Setälä H, Wilson SD (2000) Biodiversity and ecosystem function: an issue in ecology. Bull Ecol Soc Am 81:235-239
Weis JJ, Madrigal DS, Cardinale BJ (2008) Effects of algal diversity on the production of biomass in homogeneous and heterogeneous nutrient environments: a microcosm experiment. PLoS ONE 3:e2825
Wollgast S, Lenz M, Wahl M, Molis M (2008) Effects of regular and irregular temporal patterns of disturbance
on biomass accrual and species composition of a subtidal hard-bottom assemblage. Helg Mar Res
62:309-319
Yachi S, Loreau M (1999) Biodiversity and ecosystem productivity in a fluctuating environment: The insurance hypothesis. P Natl Acad Sci USA 96:1463-1468
24
Glossary
Glossary
Aggregate ecosystem property: a property that is calculated by summing that property
across the species living in the ecosystem.
Biodiversity: the sum of all biotic variation in the biosphere from the level of gene to ecosystem.
Compensation, species compensation, compensatory dynamics: a decrease in the abundance of one species that is accompanied by a compensatory increase in the abundance of
other.
Complementarity effect: an increase in an ecosystem property due to resource complementarity or facilitation among species in a species-rich community.
Ecosystem: the level of biological organisation that includes animals and plants in association, together with the physical variables of their surroundings. Ecological interactions between species regulate the transformation and transport of energy across the ecosystem.
Such transformations include the assimilation of carbon dioxide into organic carbonic compounds by plants and the consumption of plants by grazers and animals by carnivorous.
Ecosystem functioning, performance: a simple contraction for ‘how ecosystems work’
and encompasses ecosystem properties, goods, and services.
Ecosystem goods and services: the ecosystem properties that contribute to human welfare
both directly and indirectly.
Ecosystem properties: 1 the sizes of pools of materials like nutrients and carbon. 2 the
rates of processes like energy fluxes across trophic levels.
Functional traits: the characteristics of species that influence ecosystem properties.
Functional effect traits: the characteristics of species that influence ecosystem properties;
e.g., biomass, size, rate of nutrient uptake.
Functional group: a classification of species according to either their effect to ecosystem
properties or their response to the environment.
25
Biodiversity and stability
Functional response traits: the characteristics of species that define how species respond
to the environment; e.g., ranges of tolerance to salinity, temperature, solar radiation, or other
environmental variables.
Observational study: a study in which a variable is measured across individuals, populations, or higher levels of organisation. No attempt is made to affect the response of the observational units––no treatment is applied, for example.
Overyielding: an increase in the magnitude of an ecosystem property (e.g. community biomass) as species richness increases.
Positive selection effects: the increased probability that more diverse communities include
species with extreme functional values. The performance of the community represents then
the performance of these particular species instead of the average response of all of the species in the community.
Removal experiment: an experiment in which the individuals of a species or functional
group is removed from a community.
Resource complementarity, partitioning, niche partitioning: the capacity of species to
use different resources or use them in different points of time or space.
Richness effect: an increase in an ecosystem property in a species-rich community relative
to a species-poor one. The property in the species-rich community is larger than the average
property calculated across the constituent species grown alone (monocultures).
Species accumulation curve: a graphical method used to estimate the species richness in
areas where the observer is unable to sample all of the species. The procedure consists of
calculating and plotting the average number of species (and its standard deviation) of the
smallest sample size (1). Then all combinations of the next sample size are randomised and
the mean cumulative species richness is plotted. This procedure is repeated for all sample
sizes.
Species richness: the number of species living in a given area.
Stability: the state of being not likely to change or fail.
26
Glossary
Statistical averaging, portfolio effect: a reduction in the variability in an ecosystem property as species richness increases. This occurs because the ecosystem property is the sum of
that property across (temporally) fluctuating species; adding more species increases the
probability that these fluctuations will ‘average out’.
Summed covariances: the sum of all of the pair-wise species covariances (calculated
throughout time) within a biological community
Summed variances: the sum of all of the species variances (calculated throughout time)
within a community.
Synthetic-assemblage experiment: an experiment in which the researcher constructs communities by placing together individuals of several species into an experimental unit. The
selection of species is usually done at random from a subset of the local species pool.
Transgressive overyielding: the phenomenon in which the productivity or resource use of
species-rich mixtures exceeds (transgress) that of the best-performing species grown alone.
The presence of transgressive overyielding suggests that positive species interactions may
be responsible for richness effects.
27
Biodiversity and stability
Appendix
Papers I – IV
28
Paper I
Paper I
Nelson Valdivia*, Markus Molis (In press) Observational evidence of a negative
biodiversity-stability relationship in intertidal epibenthic communities. Aquatic Biology
Biologische Anstalt Helgoland, Alfred Wegener Institute for Polar and Marine Research,
Section Seaweed Biology, Kurpromenade 201, 27498 Helgoland, Germany
* Corresponding author
Email [email protected]
Tel. ++49(0)47258193294
Fax ++49(0)47258193283
1
Appendix
2
Paper I
ABSTRACT: The idea that diversity begets the functioning and stability of ecosystems has
been intensely examined in terrestrial habitats, yet these relationships remain poorly studied
in the marine realm. Theoretical and empirical work suggests that diversity enhances the
stability of communities, but decreases the stability of populations. This is because
compensatory dynamics, such that when one species decreases while other increases,
stabilise the community as long as species richness increases the variety of responses to the
environment. In an observational field study, the temporal variability in species abundance
was used as a measure of stability that was compared among five intertidal sites of naturally
different species richness. Percent covers of macrobenthic species were estimated every 6
months for 2 years. Stability in total community cover was a negative but curvilinear
function of species richness. In addition, the stability of single populations (averaged over
all species) fluctuated across the species richness gradient, without showing the predicted
negative pattern. We found no evidence for increasing compensatory dynamics with
increasing species richness, suggesting that the variety of responses to environmental
changes was unrelated to diversity. Diversity-stability relationships in natural communities
may be more complex than those predicted by theory and manipulative experiments.
KEY WORDS: Diversity-stability hypothesis · Hard-bottom · Intertidal · Marine · Portfolio
effect · Species compensation · Species richness · Statistical averaging · Temporal variability
3
Appendix
INTRODUCTION
The effects of biodiversity on ecosystem processes have received considerable attention
because of the concern that loss of biodiversity can impair the functioning of ecosystems
(reviewed by Hooper et al. 2005, Stachowicz et al. 2007). Greater species diversity
represents more adaptive responses to environmental fluctuations (MacArthur 1955, Elton
1958). By this, the probability that some species maintain functioning when other species
fail ensures the persistence of ecosystem properties under variable environmental conditions
(Walker 1992, Yachi & Loreau 1999). Indeed, influential research in terrestrial habitats has
shown that diversity is beneficial for the functioning and stability of ecosystems (e.g.
Tilman 1996, Hector et al. 1999, Loreau & Hector 2001, Tilman et al. 2006). However,
these ideas remain poorly examined in aquatic ecosystems, for which there is also a need for
understanding the ecological consequences of species loss (Gessner et al. 2004, Hooper et
al. 2005). Considering the differences between terrestrial and aquatic ecosystems (Giller et
al. 2004), generalisations obtained from terrestrial habitats may not apply for marine
habitats.
Stability has several meanings in ecology, including the resistance to and the resilience
from disturbances, the resistance to invasions, and the temporal variability in an community
property (Johnson et al. 1996, Shea & Chesson 2002). In this study we focus on temporal
variability, expressed as the temporal variance in total species cover of intertidal epibenthic
communities, and on the role of statistical averaging (also called the portfolio effect) and
overyielding as two mechanisms by which community variability decreases with increasing
diversity. Statistical averaging occurs when an aggregate community property (e.g. total
species abundance) is calculated by adding that property across species. If the temporal
variations of species are asynchronous, adding more species will increase the probability
4
Paper I
that those fluctuations are averaged out and the variability in total abundance will decrease
merely on statistical grounds (Doak et al. 1998). Nevertheless, because asynchrony among
species can result from the different abilities of species to tolerate environmental changes,
statistical averaging is due in part to ecological differences among species (Cottingham et al.
2001).
Asynchrony in species fluctuations lead to compensatory dynamics, such that the
abundance of one species decreases while that of other increases; the resulting negative
covariance buffers the community stability (e.g. Vasseur & Gaedke 2007). If species
richness increases the variety of responses to the environment, then the presence of more
species increases the probability that some species compensate the loss of others (Yachi &
Loreau 1999, Ives et al. 2000). Increasing compensatory dynamics with increasing diversity
will tend to stabilise the community but will cause individual populations to be more
unstable (Lehman & Tilman 2000).
The strength of statistical averaging effects depends on the relative abundance of species
(Steiner et al. 2005). When species contribute unequally to the community abundance, the
negative effect of diversity on community variability is dampened (Doak et al. 1998), as
shown in terrestrial plant communities (Polley et al. 2007). Moreover, high species
dominance can inverse the diversity-stability relationship, and may lead to a non-linear
diversity-stability relationship (Lhomme & Winkel 2002).
Overyielding, increases in the mean of an aggregate property with species richness, is the
second mechanisms that influences the diversity-stability relationship. Overyielding comes
from differences among species–if many species compete for several resources, then
coexistence results in a greater proportion of space covered by the community (Tilman et al.
1997). A more diverse assemblage stands for greater variety of species traits, which can
cause the average community property to increase in comparison to the property of the
5
Appendix
average population. This overyielding effect will temporally stabilise the community as
species richness increases (Lehman & Tilman 2000).
Experimental manipulation of marine epibenthic diversity shows that diversity enhances
community stability (reviewed by Stachowicz et al. 2007). Spatial models based on
observational data, however, predict the contrary (Dunstan & Johnson 2004, 2006). This is
possible when species produce aggregate structures (e.g. aggregations of conspecifics or
colonies), as result from differential use of the space among species. These structures raise
spatial refuges, leading to enhanced probabilities of survival and to more stable
communities at low diversity sites (Dunstan & Johnson 2004, 2006). Contrarily, theory
predicting a positive diversity-stability relationship is based on the assumption of wellmixed communities, where aggregations of conspecifics are almost absent (Dunstan &
Johnson 2006).
In an observational study, we tested the relationships between species richness and
community stability. Observational studies permit the inspection of broader ranges of
species richness and more realistic environmental conditions than those usually present in
manipulative experiments (Stachowicz et al. 2007). We tested whether species richness is
positively related to community stability (temporal variability in cover summed across all
species in a sampling unit), but negatively related to population stability (temporal
variability in the cover of individual epibenthic species). In addition, we investigated
whether species richness is positively related to average community cover (i.e. overyielding
effect), and whether species richness is positively related to the occurrence of species
compensation (i.e. whether species covariances become more negative as species richness
increases).
6
Paper I
MATERIAL AND METHODS
Study sites. The study was conducted at 5 intertidal sites of naturally differing species
richness on the rocky shore of Helgoland Island, NE Atlantic. Each site was ca. 200 m2 in
area and adjacent sites were 100 m apart from each other. Two sites, ‘Barren Ground’
(BG) and ‘Semi-sheltered Fucus Bed’ (SFB), were located on the moderately exposed
north-eastern shore, which is partly sheltered from wave action by a 250 m long concrete
jetty running from north to south. The mid-intertidal at BG was formerly dominated by the
blue mussel Mytilus edulis and fucoid seaweeds (Bartsch & Tittley 2004). Today, the
community at BG is dominated by encrusting coralline algae (Phymatolithon spp.) and high
densities of the periwinkle Littorina littorea, while mussels and fucoid seaweeds have
almost disappeared. During September and November 2007, the average densities of L.
littorea were 227 and 281 ind. m-2 at BG, but 16 and 90 ind. m-2 at SFB (M. Molis, unpubl.
data). At SFB, the canopy-forming brown seaweed Fucus serratus extensively covers the
substrate from the lower intertidal to the upper subtidal, where the understorey is dominated
by Phymatolithon spp. and the turf-forming algae Cladophora rupestris, Chondrus crispus,
and Corallina officinalis (Bartsch & Tittley 2004). The third site, ‘Exposed Fucus Bed’
(EFB) was located at the western wave-exposed rocky shore of Helgoland. Here, the dense
F. serratus canopy had been gradually replaced by the red algae C. crispus and
Mastocarpus stellatus (Bartsch & Tittley 2004). The fourth and fifth sites were located on
concrete harbour walls at south-eastern Helgoland. ‘Exposed Mole’ (EM) is a wave-exposed
site that is dominated by dense turfs of C. rupestris, patches of the barnacle Verruca
stroemia, and Phymatolithon spp. The fifth site, ‘Sheltered Harbour’ (SH) is a wavesheltered site that is dominated by a number of red algae such as Phyllophora spp.,
7
Appendix
Ceramium virgatum, and Bonnemaisonia hamifera (Trailliella-phase). In addition F.
serratus and the encrusting bryozoan Electra pilosa exist here in high abundance.
Community sampling. During March 2006, fifteen 0.5 × 0.5 m plots were randomly
positioned and permanently marked with stainless screws at each site. All sites were
sampled every six months between March 2006 and March 2008, except that the final
sampling of SH was delayed by one month. Due to time constrains, a random sub-sample of
nine fixed plots was followed throughout time. In species accumulation curves, seven or
eight plots were enough to represent the number of species at each site (Appendix 1,
available as Supplementary Material). Over the two-year study period, two plots were lost at
SH and EFB and one plot at EM.
For each plot, percent cover of each macrobenthic species was estimated to the nearest 1
%. Species with <1 % cover in a plot were uniformly recorded with 0.5 % abundance. Due
to the multilayered structure of the assemblages, total community cover could well exceed
100 %. Taxa were identified to the lowest possible taxonomic level in the field. For
ambiguous taxa, sub-samples collected from adjacent areas were identified in the laboratory.
Some taxa were identified to genus level, such as Phymatolithon spp., Porphyra sp., and
Ulva spp. Small burrowing spionids were grouped as family Spionidae and Ectocarpales
were identified to order level (Appendix 2).
Data analysis. Because species richness did vary over time, the gradient of species
richness was defined by using species accumulation curves that were generated separately
for each site, using the data of all sample dates. The maximum number of species obtained
from each curve corresponded to the site-specific richness used in the analyses. Species
occurring in less than three out of the five sample dates or contributing <1 % to total
community cover were omitted from all analyses, except for rare species with a consistent
8
Paper I
seasonal pattern (defined as the occurrence of a species during the same season across
years).
The PRIMER Similarity-Percentages routine, SIMPER, was used to identify the species
with larger contribution to the multivariate structure of each site. Bray-Curtis (BC)
similarities (1 – BC) were calculated between all pairs of samples in the entire data set. The
average similarities between all pairs of within-site samples were then broken down into
separate contributions from each species to the structure of each site (Clarke & Warwick
2001).
The P V-1 ratio (temporal stability, S) was used as a measure of community stability,
where P is the temporal mean community total cover for a time period and V is its temporal
standard deviation over the same interval (Tilman 1999). In comparison to the frequently
used coefficient of variability (100 V P-1), which approaches zero as stability decreases, S is
advantageous because its magnitude increases with stability. The stability of the ith species,
Si, was calculated by dividing its mean cover by its standard deviation. Population stability
was then calculated for each plot by averaging Si across all species (Tilman et al. 2006).
The temporal variance in total community cover was partitioned into the sum of all (N) of
the species variances and that of species covariances. This was done by calculating an N × N
covariance matrix across time for each plot; the sum of the diagonal corresponds to the
summed species variances and the sum of the off diagonals to the summed species
covariances. The sum of the full covariance matrix corresponds to the net variance (i.e.
summed variances plus summed covariances). The summed covariances were used as a
measure of compensatory dynamics–if species compensation increases, then the summed
covariances become more negative.
Regression analyses of the relationship between diversity and stability were conducted
using R environment version 2.7.2 (R Development Core Team 2008). We conducted
9
Appendix
orthogonal polynomial regressions to assess curvilinear patterns of diversity-stability
relationships. We tested up to the fourth-order fit (one minus the number of richness levels)
and we used the procedure described by Sokal and Rohlf (1995), in which the significance
of each polynomial regression is tested as part of the ANOVA table. All curves were fitted
using least squares regression and analyses of variance used the general linear models
routines. All measures of stability were Loge transformed due to their patchy statistical
distribution. The transformation assured normality and allowed the use of general linear
models.
Regression analyses were also used to investigate the relationship between richness and
(1) the average total community cover (averaged over the five sample dates), (2) the sum of
all species variances, (3) the sum of all pair-wise species covariances, and (4) the net
variance in total community cover. Analysis 1 was done to test whether increasing species
richness leads to overyielding and analyses 2, 3, and 4 to test whether increases in the
variance of species abundances are offset by increasingly negative species covariances.
Statistical averaging effects depend on the way in which the temporal variance in the
abundance of a species changes with the temporal mean (Tilman et al. 1998). The general
tendency of the variance V2 to increase with the mean P is described with Taylor’s power
function, V2 = c Pz, where c is a constant and z is the scaling coefficient (Taylor 1961). The
value of z affects the strength of the statistical averaging, with 1 < z < 2 meaning that
diversity dampens the community variability but increases the population variability
(Tilman et al. 1998, Tilman 1999). The logarithmic transformation of V2 = c Pz results in a
linear equation in the form of log (V2) = c + z log (P). We fitted this regression to the most
important species identified by SIMPER routines and to the entire data set, combining all
species.
10
Paper I
RESULTS
Seventy-three taxa were identified during the study; 52 were included in the analyses
(Appendix 2). Site species richness was of 30 at BG, 34 at EFB, 36 at EM, 40 at SFB, and
43 at SH. The total community cover averaged over the five sample dates (± SEM) ranged
from 119 ± 7 % (BG) to 211 ± 6 % (SFB). The taxa contributing most to the community
structure at each site were identified using SIMPER routines (Table 1). At BG, EM, and
SFB, 3 to 4 species contributed the 90 % of the communities; at EFB and SH, 6 and 8
species respectively. The taxa with the highest and most consistent contribution to withinsite similarities were Phymatolithon spp., Fucus serratus and Cladophora rupestris (Table
1). These three species represented 61 % of the sum of all of the species abundances from
the five sample sites.
Contrary to our predictions, community stability was a negative and curvilinear function
of species richness (Fig. 1). Accordingly, both the linear and cubic models significantly fit
these data (Fig. 1, Table 2). Highest community stability values were found at BG, while
lowest values were found at EFB and SH. Population stability showed large fluctuations
over the species richness gradient, and no clear trend to decrease was observed.
Consequently, the linear model was insignificant, whereas the quadratic and cuartic models
explained significant portions of the population stability data (Fig. 1, Table 2). Population
stability was highest at EM, and lowest at EFB and SH.
The average total community cover significantly increased with site species richness (y =
–54.10 + 5.95×, R2 = 0.5, F1, 38 = 38, p < 0.0001). In addition, total community cover
increased with site diversity at each of the sample dates (separate regressions performed at
each sample date, p 0.004).
11
Appendix
The summed variances showed an oscillating pattern across the species richness gradient,
and a significant trend to increase (Fig. 2, Table 3). On the other hand, the summed
covariances were independent of species richness (Fig. 2, Table 3). Summed covariances
were on average less than zero (–766.2 ± 188.5, one sample t-test, p 0.001). When
analysing each site separately, however, we found that summed covariances were less than
zero at BG, EM, and SH (one sample t-tests, p 0.03), but not at EFB and SFB (one sample
t-test, p 0.09). As a consequence of the insignificant relationship between the summed
covariances and diversity, the net variance (i.e. summed variances plus summed
covariances) followed a similar pattern to that of the summed variances, showing an
irregular increase over the species richness gradient (Fig 2, Table 3).
The fitted z-value (± SEM) for the three taxa with the highest contribution to the withinsite similarities were 1.26 ± 0.14 for Phymatolithon spp., 1.38 ± 0.06 for Fucus serratus,
and 1.12 ± 0.1 for Cladophora rupestris; the fitted z-value for the entire data set was 1.34 ±
0.01. According to their z-values, the stability of these taxa should have decreased with
species richness; but, the individual regressions showed differing patterns. The stability of
Phymatolithon spp. tended to decrease with increasing species richness, while that of Fucus
serratus and Cladophora rupestris showed large departures from linearity that resulted in a
significant cuartic fit for both species (Fig. 3, Table 4).
12
Paper I
DISCUSSION
Community stability
These observations suggest that community stability decreased as the number of species
increased, in contrast to what most theoretical and empirical work predicts (reviewed by
Hooper et al. 2005, Stachowicz et al. 2007). In addition, the patterns of community and
population stability were highly complex. In this study, the average total community cover
significantly increased with species richness (i.e. overyielding) and the variance scaled with
the mean cover with 1 < z < 2–overyielding and such variance-mean rescaling should have
led to a positive diversity-stability relationship (Tilman et al. 2006, van Ruijven & Berendse
2007). Yet, increasing stability with increasing diversity also requires increasingly negative
species covariances and an even distribution of species abundances.
On average, summed covariances were significantly less than zero. At both sites
dominated by the canopy forming Fucus serratus, however, covariances were equal or
larger than zero. Positive covariances in these sites may have resulted from the positive
effect of F. serratus on obligate understorey species (N. Valdivia, unpublished data).
Moreover, persistent removals of the F. serratus canopy caused compensatory dynamics of
species with different environmental tolerances; the resulting negative covariances buffered
the community stability but reduced the population stability (N. Valdivia unpublished data).
Therefore, the covariance in the species responses to environmental disturbances can
strongly influence the stability of the here studied shores. In the present study, the
insignificant relationship between the species covariances and species richness probably
prevented a positive effect of diversity on stability.
13
Appendix
The relationship between species richness and stability was also influenced by the relative
abundance of species. In this experiment, three taxa explained ca. the 60 % of the sum of all
of the species covers. When few taxa numerically dominate the system, community stability
can be driven by fluctuations of these components (Steiner et al. 2005, Polley et al. 2007).
In addition, large differences among species abundances can result in negative and
curvilinear richness-stability relationships when z = 1.2 (Lhomme & Winkel 2002). In our
case, the z-values were close to 1.2 (e.g. 1.26 ± 0.14 for Phymatolithon spp., but 1.35 ± 0.01
for all species), suggesting that large heterogeneity among species abundances may also
explain the negative and complex pattern of community stability.
Overyielding probably resulted from the multilayered structure of macrobenthic
assemblages, which allows single species to expand by differential use of the available
space. Erect life forms use little space of primary substratum but can expand above the
substratum and thus increase in abundance. This causes the total percent cover to exceed
100 %. For instance, seaweeds can develop and expand a canopy in an area where the
primary substratum was monopolised by encrusting forms (Connell 2003). Such a spatial
structure was apparent in this study, as encrusting, turfing, and canopy-forming algae
formed 3 layers of biota. This suggest that if we would have focused on one layer of species
(i.e. do not allow total percent covers > 100 %), probably we would have found no
overyielding. On the other hand, large spatial structures may also have caused community
stability to decrease with species richness, because such structures raise refugee and
increase the probability of survival in communities dominated by few species (Dunstan &
Johnson 2004, 2006).
Population stability
14
Paper I
We detected significant fluctuations in the pattern of population stability across the
species richness gradient, but we did not find a clear trend to decrease. Population stability
should decrease with increasing diversity if the latter is positively related to the number of
potential competitive interactions or to the variety of adaptive responses to the environment
(Ives et al. 2000). In this study, the absence of a negative diversity-covariance relationship
suggests that both the variety of environmental tolerances and the number of competitive
interactions were independent of diversity. Differential use of the space could have
alleviated competition at high diversity sites, reducing the probability of compensatory
changes that cause individual populations to be more variable.
According to their z-values, the stability of single species should have decreased with
species richness (Tilman 1999). However, individual species tended to be more stable at
sites where they were more abundant. Similarly, a recent experiment in which species
abundances varied across the diversity gradient showed that the stability of single species
performed differently than expected based on the variance-mean rescaling (van Ruijven &
Berendse 2007). Because constancy in species abundance is an assumption of statistical
averaging (Doak et al. 1998), this mechanisms may be well supported by manipulative
experiments, but probably not by observational studies.
Our observations agree with studies conducted in multitrophic systems showing no clear
diversity effect on population stability (McGrady-Steed & Morin 2000, Steiner et al. 2005),
but contradict studies conducted on single trophic levels that show negative relationships
(Tilman et al. 2006, van Ruijven & Berendse 2007). In our case, primary producers
dominated the assemblages in terms of abundance, but 57 % of taxa were invertebrates. On
the other hand, keystone consumers can strongly control the community structure (Paine
1966). Therefore, the high stability of Phymatolithon spp. at the species-poor site Barren
Ground might be related to the large density of the periwinkle Littorina littorea observed at
15
Appendix
the study site. Epibenthic grazers like L. littorea control the recruitment of algae, thereby
affecting the structure of macrobenthic assemblages (McQuaid 1996). The grazing activity
of L. littorea at Barren Grounds may be an important factor in depressing species richness
and simultaneously promoting the persistence of encrusting algae like Phymatolithon spp. at
high abundances. Manipulative experiments are necessary to address the role of trophic
interactions on relationship between diversity and stability.
In conclusion, we observed a negative and curvilinear pattern in community stability and
a complex pattern in population stability. Probably, putative positive effects of overyielding
and variance-mean rescaling on community stability were offset by strong heterogeneity
among species abundances and invariant species covariances across the species richness
gradient. The observational evidence presented here is not unequivocal, since we did not
control for factors that might have covaried with species richness, such as wave exposure or
nutrient levels. In addition, ecosystem properties such as fluxes of nutrients and carbon were
not assessed. Because different ecosystem properties can have different responses to
changes in diversity (Jiang et al. 2008), experiments that explore multiple ecosystem
properties can provide a more comprehensive view of the functional role of diversity. Even
though, we suggest that the relative abundance of species and ecological interactions
influencing the covariances among species may play a pivotal role in the relationship
between diversity and ecosystem stability.
Acknowledgements. We are grateful to a number of friends and colleagues who
enthusiastically helped during long hours of field work, including J. Ellrich, A. Engel, M.
Honens, M. Marklewitz, A. Wagner, and H.Y. Yun. A. Knox polished the English of an
early version of the manuscript. This work forms part of the MarBEF responsive model
16
Paper I
project BIOFUSE. Financial support by the Alfred-Wegener-Institute for Marine and Polar
Research to N.V. is acknowledged.
17
Appendix
References
Bartsch I, Tittley I (2004) The rocky intertidal biotopes of Helgoland: present and past. Helg
Mar Res 58:289-302
Clarke KR, Warwick RM (2001) Change in marine communities: an approach to statistical
analysis and interpretation PRIMER-E Ltd, Plymouth
Connell SD (2003) The monopolization of understorey habitat by subtidal encrusting
coralline algae: a test of the combined effects of canopy-mediated light and
sedimentation. Mar Biol 142:1065-1071
Cottingham KL, Brown BL, Lennon JT (2001) Biodiversity may regulate the temporal
variability of ecological systems. Ecol Lett 4:72-85
Doak DF, Bigger D, Harding EK, Marvier MA, O'Malley RE, Thomson D (1998) The
statistical inevitability of stability-diversity relationships in community ecology. Am
Nat 151:264-276
Dunstan PK, Johnson CR (2004) Invasion rates increase with species richness in a marine
epibenthic community by two mechanisms. Oecologia 138:285-292
Dunstan PK, Johnson CR (2006) Linking richness, community variability, and invasion
resistance with patch size. Ecology 87:2842-2850
Elton SC (1958) The ecology of invasions by animals and plants. The University of Chicago
Press, Chicago, USA
Gessner MO, Inchausti P, Persson L, Raffaelli DG, Giller PS (2004) Biodiversity effects on
ecosystem functioning: insights from aquatic systems. Oikos 104:419-422
Giller PS, Hillebrand H, Berninger U-G, Gessner MO, Hawkins S, Inchausti P, Inglis C,
Leslie H, Malmqvist B, Monaghan MT, Morin PJ, O'Mullan G (2004) Biodiversity
effects on ecosystem functioning: emerging issues and their experimental test in
aquatic environments. Oikos 104:423-436
Hector A, Schmid B, Beierkuhnlein C, Caldeira MC, Diemer M, Dimitrakopoulos PG, Finn
JA, Freitas H, Giller PS, Good J, Harris R, Högberg P, Huss-Danell K, Joshi J,
Jumpponen A, Körner C, Leadley PW, Loreau M, Minns A, Mulder CPH,
O'Donovan G, Otway SJ, Pereira JS, Prinz A, Read DJ, Scherer-Lorenzen M,
Schulze ED, Siamantziouras ASD, Spehn EM, Terry AC, Troumbis AY, Woodward
FI, Yachi S, Lawton JH (1999) Plant diversity and productivity experiments in
european grasslands. Science 286:1123-1127
Hooper DU, Chapin I, F. S., Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge
DM, Loreau M, Naeem S, Schmid B, Setälä H, Symstad AJ, Vandermeer J, Wardle
DA (2005) Effects of biodiversity on ecosystem functioning: a consensus of current
knowledge. Ecol Monogr 75:3-35
Ives AR, Klug JL, Gross K (2000) Stability and species richness in complex communities.
Ecol Lett 3:399-411
Jiang L, Pu Z, Nemergut DR (2008) On the importance of the negative selection effect for
the relationship between biodiversity and ecosystem functioning. Oikos 117:488-493
Johnson KH, Vogt KA, Clark H, Schmitz O, Vogt D (1996) Biodiversity and the
productivity and stability of ecosystems. Trends Ecol Evol 11:372-377
Lehman CL, Tilman D (2000) Biodiversity, stability, and productivity in competitive
communities. Am Nat 156:534-552
Lhomme JP, Winkel T (2002) Diversity-stability relationships in community ecology: Reexamination of the portfolio effect. Theor Popul Biol 62:271-279
Loreau M, Hector A (2001) Partitioning selection and complementarity in biodiversity
experiments. Nature 412:72-76
18
Paper I
MacArthur RH (1955) Fluctuations of animal populations and a measure of community
stability. Ecology 36:533-536
McGrady-Steed J, Morin PJ (2000) Biodiversity, density compensation, and the dynamics
of populations and functional groups. Ecology 81:361-373
McQuaid CD (1996) Biology of the gastropod family Littorinidae .2. Role in the ecology of
intertidal and shallow marine ecosystems. Oceanography and Marine Biology, Vol
34 34:263-302
Paine R (1966) Food web complexity and species diversity. Am Nat 100:65-75
Polley HW, Wilsey BJ, Derner JD (2007) Dominant species constrain effects of species
diversity on temporal variability in biomass production of tallgrass prairie. Oikos
116:2044-2052
R Development Core Team (2008) R: A language and environment for statistical
computing, R Foundation for Statistical Computing, Vienna
Shea K, Chesson P (2002) Community ecology theory as a framework for biological
invasions. Trends Ecol Evol 17:170-176
Sokal RR, Rohlf RJ (1995) Biometry. Freeman and Company, New York
Stachowicz JJ, Bruno JF, Duffy JE (2007) Understanding the effects of marine biodiversity
on communities and ecosystems. Annu Rev Ecol Evol S 38:739-766
Steiner CF, Long ZT, Krumins JA, Morin PJ (2005) Temporal stability of aquatic food
webs: partitioning the effects of species diversity, species composition and
enrichment. Ecol Lett 8:819-828
Taylor LR (1961) Aggregation, variance and the mean. Nature 189:732-735
Tilman D (1996) Biodiversity: population versus ecosystem stability. Ecology 77:350–363
Tilman D (1999) The ecological consequences of changes in biodiversity: a search for
general principles. Ecology 80:1455–1474
Tilman D, Lehman CL, Bristow CE (1998) Diversity-stability relationships: Statistical
inevitability or ecological consequence? Am Nat 151:277-282
Tilman D, Lehman CL, Thomson KT (1997) Plant diversity and ecosystem productivity:
Theoretical considerations. P Natl Acad Sci USA 94:1857-1861
Tilman D, Reich PB, Knops JMH (2006) Biodiversity and ecosystem stability in a decadelong grassland experiment. Nature 441:629-632
van Ruijven J, Berendse F (2007) Contrasting effects of diversity on the temporal stability
of plant populations. Oikos 116:1323-1330
Vasseur DA, Gaedke U (2007) Spectral analysis unmasks synchronous and compensatory
dynamics in plankton communities. Ecology 88:2058-2071
Walker BH (1992) Biodiversity and ecological redundancy. Conserv Biol 6:18-23
Yachi S, Loreau M (1999) Biodiversity and ecosystem productivity in a fluctuating
environment: The insurance hypothesis. P Natl Acad Sci USA 96:1463-1468
19
Appendix
Table 1. Dominant taxa at sites with a naturally differing number of species. Decomposition
of within-site Bray-Curtis similarities into contribution of taxa to the structure of each site
(Contributioni). Taxa contributions are also expressed as percent (%i). A value of the ratio
Contributioni/SD 1 indicates that the contribution of taxon i to the within-site similarity is
consistent across all pairs of samples. Percent cover (averaged over plots and all sample
dates) of each taxon is given (%-coveri). Site-specific species richness is given in brackets.
Taxa cumulating up to 90 % of the contribution to the with-site similarities are shown.
Taxon
%-coveri
Contributioni (%i)
Contributioni/SD
70.78
49.31
78.90
2.95
Littorina littorea
6.67
3.96
6.33
1.92
Hildenbrandia rubra
7.92
2.62
4.19
0.56
12.32
2.56
4.09
0.31
BG, Barren Ground (30)
Phymatolithon spp.
Haemescharia hennedyi
EFB, Exposed Fucus Bed (34)
Phymatolithon spp.
42.67
16.13
40.35
1.27
Fucus serratus
32.37
9.08
22.71
0.75
Chondrus crispus
17.89
5.27
13.19
0.81
Corallina officinalis
8.63
2.09
5.23
0.61
Mastocarpus stellatus
9.24
1.96
4.91
0.50
11.66
1.70
4.26
0.38
Ulva spp.
20
Paper I
EM, Exposed Mole (36)
Cladophora rupestris
86.68
49.02
73.15
3.10
Phymatolithon spp.
20.56
6.28
9.37
1.07
Verruca stroemia
22.75
5.88
8.78
0.94
SFB, Semi-sheltered Fucus Bed (40)
Fucus serratus
76.27
29.30
40.33
2.43
Phymatolithon spp.
66.51
26.89
37.02
3.69
Cladophora rupestris
38.40
12.52
17.24
1.96
Ceramium virgatum
31.79
7.63
23.60
0.67
Fucus serratus
30.69
6.24
19.28
0.59
Electra pilosa
22.29
4.85
15.00
0.73
Bonnemaisonia hamifera
23.57
4.03
12.47
0.39
Phyllophora spp.
12.14
2.34
7.24
0.62
Chondrus crispus
9.97
2.31
7.13
0.73
Ulva spp.
8.87
1.30
4.02
0.56
10.81
1.14
3.53
0.31
SH, Sheltered Harbour (43)
Ectocarpales
21
Appendix
Table 2. Results of orthogonal polynomial regressions of species richness on community
and population stability.
Source
df
MS
F
p
Community stability
Species richness, N
4
1.81
9.40
<0.0001
Nlinear
1
3.44
17.82
0.0002
Nquadratic
1
0.79
4.08
0.0511
Ncubic
1
2.63
13.61
0.0008
Ncuartic
1
0.40
2.09
0.1569
35
0.19
Species richness, N
4
0.59
7.56
0.0002
Nlinear
1
0.15
1.94
0.1725
Nquadratic
1
0.48
6.17
0.0179
Ncubic
1
0.30
3.89
0.0566
Ncuartic
1
1.42
18.25
0.0001
35
0.08
Residual
Population stability
Residual
22
Paper I
Table 3. Results of orthogonal polynomial regressions of species richness on summed
variances, summed covariances, and net variance (summed variances + summed
covariances).
Source
df
MS
F
p
Summed variances
Species richness, N
4
8490828
9.98
<0.0001
Nlinear
1
7589923
8.92
0.0051
Nquadratic
1
142765
0.17
0.6846
Ncubic
1
19102565
22.45
<0.0001
Ncuartic
1
7128060
8.38
0.0065
35
850863
Species richness, N
4
1564017
1.11
0.3662
Nlinear
1
426974
0.30
0.5851
Nquadratic
1
555457
0.40
0.5337
Ncubic
1
4858397
3.46
0.0715
Ncuartic
1
415240
0.30
0.5902
35
1405830
7.31
0.0002
Residual
Summed covariances
Residual
Net variance (summed variances + summed covariances)
Species richness, N
4
5137088
23
Appendix
Nlinear
1
11617286
16.54
0.0003
Nquadratic
1
135017
0.19
0.6638
Ncubic
1
4693597
6.68
0.0141
Ncuartic
1
4102454
5.84
0.0210
35
702542
Residual
24
Paper I
Table 4. Results of orthogonal polynomial regressions of species richness on the stability of
each of the three dominant taxa
Source
df
MS
F
p
Phymatolithon spp.
Species richness, N
4
4.70
12.54
<0.0001
Nlinear
1
3.75
10.00
0.0032
Nquadratic
1
1.14
3.03
0.0904
Ncubic
1
12.09
32.25
<0.0001
Ncuartic
1
1.82
4.86
0.0341
35
0.37
Species richness, N
4
3.05
7.97
0.0002
Nlinear
1
4.88
12.72
0.0013
Nquadratic
1
1.83
4.79
0.0372
Ncubic
1
3.53
9.22
0.0051
Ncuartic
1
1.97
5.15
0.0312
28
0.38
Species richness, N
4
10.88
21.02
<0.0001
Nlinear
1
0.28
0.54
0.4659
Residual
Fucus serratus
Residual
Cladophora rupestris
25
Appendix
Nquadratic
1
20.80
40.19
<0.0001
Ncubic
1
1.30
2.51
0.1225
Ncuartic
1
21.14
40.84
<0.0001
34
0.52
Residual
26
Paper I
Figure captions
Fig. 1. Relationship between species richness and stability in percent cover of epibenthic
species. Stability, S, was calculated as the quotient between the temporal mean in cover, P,
and its standard deviation, V, over the same time period. (a) Stability of total community
cover. (b) Stability of cover of single species averaged across 52 species. Each circle
represents the stability of a 0.25 m2 plot that was followed over time. Codes for sites are as
follows. BG: Barren Ground, EFB: Exposed Fucus Bed, EM: Exposed Mole, SFB: Semisheltered Fucus Bed, and SH: Sheltered Harbour.
Note: regression parameters of site species richness (N) are as follows. Community stability
= 227.25 – 18.53N + 0.50N2 – 0.004N3. Population stability = 31450 – 346N + 14N2 –
0.25N3 + 0.0017N4
Fig. 2. Relationship between site species richness and (a) summed variances, (b) summed
covariances, and (c) net variance (summed variances plus summed covariances). Codes for
sites as in Fig. 1
Note: regression parameters of site species richness (N) are as follows. Summed variances =
– 7465000 + 810900N – 32790N2 + 585N3 – 4N4. Net variance = – 5522000 + 602900N –
24530N2 + 441N3 – 3N4
Fig. 3. Patterns of stability in percent cover of the 3 taxa with the highest and most
consistent contributions to the within-site similarities. Codes for sites as in Fig. 1
Note: regression parameters of site species richness (N) on stability (S) are as follows.
SPhymatolithon spp. = – 3014 + 346.5N – 14.83N2 + 0.27N3 – 0.002N4. SFucus serratus = – 7531 +
841N – 0.35N2 + 0.6N3 – 0.004N4. SCladophora rupestris = 1980 – 1324N + 54.51N2 – 0.9N3 +
0.007N4
27
Appendix
Community stability [log(PV-1)]
4
R2 = 0.4891
(a)
R2 = 0.4636
(b)
3
2
1
Population stability [log(PV)]
0
3
2
1
0
-1
30
Sites: BG
35
EFB
40
EM
SFB
SH
Site species richness
Fig. 1. Valdivia and Molis
28
Paper I
(a) Summed variances, R2 = 0.5328
6000
4000
2000
0
(b) Summed covariances, NS
Variance
2000
0
-2000
-4000
6000
(c) Net variance (variances + covariances)
R2 = 0.4552
4000
2000
0
30
Sites: BG
35
EFB
40
EM
SFB
SH
Site species richness
Fig. 2. Valdivia and Molis
29
Appendix
6
(a) Phymatolithon spp., R2 = 0.5889
4
2
Species stability [log(PV)]
0
6
(b) Fucus serratus, R2 = 0.5323
4
2
0
5
(c) Cladophora rupestris, R2 = 0.7120
0
30
Sites: BG
35
EFB
40
EM
SFB
SH
Site species richness
Fig. 3. Valdivia and Molis
30
Paper I
The following appendix accompanies the article
Observational evidence of a negative biodiversity-stability relationship in intertidal
epibenthic communities
Nelson Valdivia*, Markus Molis
Biologische Anstalt Helgoland, Alfred Wegener Institute for Polar and Marine Research,
Section Seaweed Biology, Kurpromenade 201, 27498 Helgoland, Germany
*Email [email protected]
Appendix 1. Species accumulation curves of data from the first sample date on March 2006
at each site. We used a random method that finds the mean species accumulation curve and
its standard deviation (vertical bars) from random permutations of the data
31
Appendix
Semi-sheltered Fucus bed
Barren ground
30
30
20
20
10
10
0
0
2
4
6
8
10
12
14
2
6
8
10
12
14
8
10
12
14
Sheltered harbour
Exposed Fucus bed
Number of species
4
30
30
20
20
10
10
0
0
2
4
6
8
10
12
14
6
8
10
12
14
Exposed mole
30
20
10
0
2
4
Number of plots
32
2
4
6
19 Desmarestia aculeata1
20 Dumontia contorta
Rhodophyta
11 Ahnfeltia plicata
12 Bonnemaisonia hamifera
13 Ceramium virgatum
14 Chondrus crispus
15 Cladostephus
spongiosus1
16 Corallina officinalis
17 Cystoclonium purpureum
18 Delesseria sanguinea1
Phaeophyceae
5 Ascophyllum nodosum1
6 Ectocarpales
7 Fucus serratus
8 Fucus vesiculosus
9 Laminaria digitata
10 Sargassum muticum
Chlorophyta
1 Cladophora rupestris
2 Rhizoclonium tortuosum
3 Ulva spp.
4 Unidentified green
microalgae1
Taxon
-
0.1
0.1
-
0.4
2.1
2.1
2.0
-
0.3
1.2
-
0.4
5.4
2.5
-
1.4
0.6
-
0.2
0.3
-
0.3
-
3.8
-
<0.1
7.3
0.6
-
0.8
-
Barren Ground
2
3
4
-
9.5
0.3
-
0.4
0.3
-
1
0.6
4.2
2.3
6.3
6.3
1.6
-
0.3
2.3
-
5
4.2
10.8
-
2.9
2
-
3.4 2.3
16.4 11.6
1.0
-
Exposed Fucus Bed
2
3
4
5
91
-
1
9.8
0.2
-
5.7
-
5.0
6.8
33
13.7 10.9
0.2
17.5
1.9 0.1
0.8 0.1
14.9 20.6 30.9 24.6 10.0
<0.1
-
0.1
0.1
-
0.4
-
12.3
2.9
33.9 54.4 18.5 30.6 22.6
0.1
0.1 0.8 <0.1
-
4.6
1.6
-
1
temporal distribution, ‘-‘ = absent. See ‘Results’ in main article for details.
5
Semi-sheltered Fucus Bed
1
2
3
4
5
0.1
0.1
-
0.1
9.0
-
0.1
0.3
-
0.1
0.5
<0.1
-
0.2
3.0
-
<0.1
0.1
-
0.2
-
0.1
4.6
-
0.1
-
<0.1
0.2
-
0.1
2.0
-
0.2
0.8
-
0.8
-
1.2
-
0.4
0.1
-
0.1
2.7
-
70.5 88.6
0.2 0.1
-
2.1
0.2
0.4
<0.1
0.2
4.9
-
13
45
0.4
-
2.0
-
2.1
9.9
-
4.9
85.6
6.1
-
1.1
29.4
-
1.6
1.3
-
1.1
4.0
-
Sheltered Harbour
2
3
4
0.4
4.0
-
5
0.1
16.1 8.3 7.4 1.7 3.2
20.5 41.8 29.1 30.4 15.0
<0.1
11.9 0.4 1.9 0.6 9.6
0.4
0.1 0.9 <0.1
0.6
4
-
1
2.8
-
0.1
-
0.1
-
-
0.1
<0.1
-
0.4
-
<0.1 0.5 0.9 0.6 0.3 <0.1
25.3 26.9 19.8 11.8 33.2
0.3 38.6 7.3 33.1 10.4 89.4
8.2 7.7 19.1 5.6 7.3 2.5
<0.1
0.4
-
93
9.7
-
87.8 84.5 85.2 89.4 18.7 33.1 42.0 57.0 46.0
<0.1
3.1
2.4 <0.1
1.0 5.3 6.6 0.6
-
Exposed Mole
2
3
4
2007, 4 = September 2007, and 5 = March 2008. 1 marks taxa that were excluded from the analyses due to low contribution or patchy spatio-
Appendix 2. Average percent cover of each taxon observed at each site and sampling date (1 = March 2006, 2 = September 2006, 3 = March
Paper I
Dynamena pumila
Laomedea flexuosa
Sagartia spp.
Sagartiogeton laceratus
Tubularia indivisa1
Lanice conchilega
Pomatoceros triqueter1
Spionidae spp.
Spirorbis spirorbis
Crustacea
46 Carcinus maenas
42
43
44
45
Polychaeta
41 Janua pagenstecheri1
36
37
38
39
40
-
<0.1
0.8
1.8
-
Demospongiae
32 Halichondria panicea
33 Leucosolenia botryoides
34 Sycon ciliatum1
Cnidaria
35 Actinia equina1
-
31 Rhodomela confervoides
-
0.4
0.1
0.1
0.1
-
-
-
-
<0.1
0.4
<0.1
0.1
-
-
-
-
-
0.1
0.5
-
<0.1 <0.1
0.1
0.5
-
<0.1 <0.1
-
-
-
-
0.1
0.4
0.2
0.1
-
0.4
-
-
-
0.4
0.2
0.1
0.4
-
-
0.1
0.1
0.2
0.3
0.1
<0.1
-
0.1
-
0.1
-
<0.1 <0.1
0.3
-
0.1
0.2
0.1
-
-
<0.1
-
34
-
0.3
<0.1
0.1
0.1
-
-
<0.1
-
-
0.2
0.2
0.2
0.1
-
-
2.0
-
0.4
0.1
0.5
0.1
-
0.2
1.4
-
0.6
-
-
0.1
0.3
0.1
0.1
-
0.3
-
1.6
-
0.1
0.2
0.5
<0.1
-
0.1
0.2
-
0.8
-
<0.1
<0.1
0.2
4.4
-
0.2
0.1
-
3.6
-
-
0.1
0.2
9.9
0.4
0.2
-
0.3
-
0.1
-
0.6
0.5
6.4
0.3
0.2
-
0.6
-
-
<0.1
<0.1
0.2
1.1
0.1
0.1
-
-
0.1
4.3
0.4
-
0.5
<0.1
-
0.1
0.1
-
0.1
0.3
0.8
0.4
-
0.2
-
<0.1 <0.1 <0.1
0.1
0.1
-
-
0.1
0.1
0.1
0.5
-
0.1
0.2
2.6
0.2
-
38.6 0.1 25.1 0.8 1.5 5.3 1.4 0.7 0.7 0.3 0.2 0.7 0.2 0.2 1.6 1.6 1.2 1.9 1.4 0.1
12.9 1.2 11.4 1.1 13.6 7.7 4.8 5.0 4.1 6.1 <0.1 0.1
1.1 1.9 2.7 1.8 1.6
2.5
9
1.2 5.3 1.5
9
6.2
5 10.4 11.1 8.1 12.1 8.8 11.8 10.5 0.3
0.9 0.2 0.4 0.8 2.4 <0.1
<0.1 <0.1
<0.1
0.1 0.3 0.8 0.4 2.4
9.5
74.6 48.9 74.4 62.6 79.9 42.1 33.3 55.1 61.1 53.1 25.1 7.9 44.8 22.5 4.2 80.1 63.1 59.8 64.6 64.1 3.9
0.1
0.1
-
21
22
23
24
25
26
27
Haemescharia hennedyi
Hildenbrandia rubra
Mastocarpus stellatus
Membranoptera alata
Phyllophora spp.
Phymatolithon spp.
Plocamium
cartilagineum
28 Plumaria plumosa
29 Polysiphonia spp.
30 Porphyra spp.1
Appendix
0.4
-
1.7
0.1
0.8
-
0.1
0.2
0.2
4.3
-
0.6
0.1
8.9
2.7
-
-
-
0.2
0.4
<0.1
0.2
0.1
-
4.5
0.3
-
-
0.1
1.2
0.6
-
0.4
<0.1
-
1.3
0.7
-
<0.1
-
0.2
0.4
-
0.1
0.1
0.1
2.0
-
<0.1 <0.1
0.6 0.9 0.5
<0.1
0.1
15.2 16
14
2.3 4.7 1.6
0.1 <0.1 <0.1
Elminius modestus
Pagurus bernhardus
Semibalanus balanoides
Verruca stroemia
Lepidochitona cinerea
Littorina littorea
Littorina obtusata
Mytilus edulis
Nucella lapillus1
Cryptosula pallasiana
Electra pilosa
Flustrellidra hispida
Membranipora
membranacea1
Ascidiacea
69 Alcyonidium
gelatinosum1
Echinodermata
67 Amphipholis squamata
68 Echinus esculentus1
63
64
65
66
Bryozoa
62 Bicellariella ciliata1
57
58
59
60
61
Mollusca
53 Crassostrea gigas
54 Elysia viridis
55 Gibbula cineraria
56 Lacuna vincta1
52 Pycnogonum littorale
1
Pantopoda
51 Nymphon brevirostre1
47
48
49
50
-
-
0.1
0.2
-
-
-
0.2
-
-
-
0.1
0.4
-
-
-
-
-
-
0.2
-
-
-
-
-
-
-
0.1
0.6
0.2 0.3 <0.1 <0.1
<0.1 0.1
0.1
4.5 10.6 7.1 5.5
0.5 0.4 0.3 0.3
4.0 1.9 3.3 1.3
-
-
-
<0.1
-
-
-
-
0.2
7.5
0.1
0.7
-
-
-
0.1
-
-
0.1
-
0.6
0.1
-
<0.1
0.1
0.7
0.5
-
-
-
-
-
0.1
-
0.1
-
0.2
0.1
1.8
0.6
-
-
-
0.1
-
-
<0.1
<0.1
-
-
-
-
-
-
-
-
-
<0.1
-
0.6
-
35
-
-
0.1
-
-
-
0.9
0.8
-
-
-
0.4
1.0
0.2
-
0.1
0.1
0.1
0.1
-
-
0.1
-
-
-
0.3
0.5
38
0.1
-
-
-
0.1 1.2
1.4 12.4
0.3 0.5
<0.1
-
<0.1
0.1
-
-
-
0.7 0.2
1.0 0.3
52.3 16.1 10.8
<0.1 <0.1
<0.1 <0.1 <0.1
0.1 <0.1 <0.1
1.4 1.1 1.6 0.1
0.5 0.4 0.4 <0.1
<0.1
0.1
<0.1
<0.1
-
-
-
-
0.2
0.4
0.5
-
-
0.1
8.5
0.7
-
0.1
0.1
0.1
-
-
-
-
0.6
0.4
-
0.3
0.1
1.2
<0.1
-
-
<0.1 <0.1
0.8
0.5
10.8
-
0.1
-
0.3
0.4
0.1
-
0.4
0.1
0.4
1.6
-
-
-
0.2
0.5
0.3
-
-
-
<0.1
-
0.1
9.4
-
-
0.1
-
-
-
0.8
7.4
1.2
-
-
-
0.2
1.8
0.7
-
-
-
0.1
3.3
-
<0.1 0.4 0.3 <0.1
<0.1 <0.1 <0.1
0.5 0.1 0.1 0.1
0.6 0.4 0.5
0.3
-
-
-
0.1
0.1 0.1 <0.1
0.2 0.5 0.3
<0.1 <0.1 <0.1
-
-
0.2
0.1
0.1
-
-
-
<0.1
0.3 <0.1
17.5
-
0.2
<0.1
<0.1
<0.1
<0.1
0.1
-
-
-
0.1
0.4
5.8
-
0.1
-
-
<0.1
<0.1
-
0.1
-
-
0.3
-
0.2
0.3 <0.1
0.2
1.8 <0.1
27.8 18.9 40.6 16.7
2.8 2.1 3.7
0.1
-
0.2
0.1
0.1
-
-
-
0.1
16.7
Paper I
-
72 Didemnum maculosum1
73 Diplosoma listerianum
-
70 Botryllus schlosseri
71 Clavelina lepadiformis1
Appendix
-
-
-
-
-
-
-
-
-
36
-
-
-
-
-
-
-
-
-
-
0.1
-
-
0.1
0.2
0.1
-
0.1
0.1
<0.1
-
-
-
Paper II
Nelson Valdivia a, *, Kate L. de la Haye b, c, Stuart R. Jenkins b, d, Susan A. Kimmance e,
Richard Thompson c, Markus Molis a (In press) Functional composition, but not richness,
affected the performance of sessile suspension-feeding assemblages. Journal of Sea Research
a
Biologische Anstalt Helgoland, Alfred Wegener Institute for Polar and Marine Research,
Section Seaweed Biology, Kurpromenade 201, 27498 Helgoland, Germany
b
Marine Biological Association of the United Kingdom, The Laboratory, Citadel Hill,
Plymouth PL1 2PB United Kingdom
c
Marine Biology and Ecology Research Centre, School of Biological Sciences, University of
Plymouth, Plymouth PL4 8AA United Kingdom
d
School of Ocean Sciences, Bangor University, Menai Bridge, Anglesey LL59 5AB United
Kingdom
e
Plymouth Marine Laboratory, Prospect Place, The Hoe, Plymouth PL1 3DH, UK
* Corresponding author
Email [email protected]
Tel. ++49(0)47258193294
Fax ++49(0)47258193283
Appendix
2
Paper II
Abstract
The efficiency by which communities capture limiting resources may be related to the
number of species or functional types competing therein. This is because species use different
resources (i.e. complementarity effect) or because species-rich communities include species
with extreme functional traits (positive selection effect). We conducted two manipulative
studies to separate the effects of functional richness and functional identity on the feeding
efficiency (i.e. filtration rate) of suspension-feeding invertebrates growing on vertical surfaces.
In addition, one experiment tested whether the density of organisms influences the effect of
functional diversity. Monocultures and complete mixtures of functional types were fed with a
solution of microalgae of different sizes (6 Pm – 40 Pm). Experiments conducted at two
locations, Helgoland and Plymouth, showed that functional identity had far larger effects on
filtration rate than richness. Mixtures did not outperform the average monoculture or the bestperforming monoculture and this pattern was independent on density. The high efficiency of
one of the functional types in consuming most microalgae could have minimised the resource
complementarity. The loss or gain of particular species may therefore have a stronger impact on
the functioning of epibenthic communities than richness per se.
3
Appendix
Keywords
Biodiversity-Ecosystem Functioning; Complementarity Effect; Density; Ecosystem
functioning; Filtration rates, Selection Effects; Resource Consumption; Suspension feeders
4
Paper II
1. Introduction
Present rates of species invasion and extinction (Thomas et al., 2004; Byrnes et al., 2007)
have stimulated research linking biodiversity and ecosystem functioning (hereafter BEF),
because of the potential loss of ecosystem ‘goods and services’ (Chapin et al., 2000). A
substantial number of theoretical and empirical studies suggest that taxonomic and functional
diversity are linked to the performance of ecosystems (Hooper et al., 2005; Stachowicz et al.,
2007). However, to gain a mechanistic understanding of BEF relationships requires us to
accurately distinguish changes in ecosystem processes due to the number of species (richness
effect) from changes attributable to the usually stronger effect of species identity or
composition (Schwartz et al., 2000).
Energy transfer in an ecosystem underlies services such as food production and nutrient
cycling (Christensen et al., 1996). In coastal ecosystems, assemblages of sessile suspensionfeeding invertebrates mediate the coupling and energy transfer between two major habitats––
the water column and the benthos (Gili and Coma, 1998). Suspension feeders can directly
control pelagic primary production by capturing large amounts of plankton, and indirectly
regulate secondary production in coastal trophic webs by supporting populations of mobile
predators (Navarrete et al., 2005; Nielsen and Maar, 2007). Assemblages of sessile suspension
feeders provide therefore an opportunity to examine the functional consequences of biodiversity
change.
Positive richness effects on ecosystem properties can result from resource complementarity
or facilitation (complementarity effect; Tilman et al., 1997; Loreau, 2000). This theory was
developed from observations on terrestrial plant communities, but it may also apply to aquatic
5
Appendix
ecosystems (Bruno et al., 2005; Griffin et al., 2008). For example, suspension feeders able to
partition a resource on the basis of particle size can form complex assemblages made up of
numerous species (Lesser et al., 1994): species feeding on smaller particles can coexist in the
same community with other species feeding on larger prey. At the same time, higher diversity
results in higher spatial complexity, which alters patterns in near-bottom currents and increases
particle capture of individuals; this allows a species-rich assemblage to capture more food than
any of its constituent species grown alone (Cardinale et al., 2002). These observations suggest
that resource complementarity and facilitation may occur within suspension-feeding
assemblages; however, few experiments have been designed to test the role of these processes
in linking suspension feeder diversity and energy transfer (e.g. Cardinale and Palmer, 2002;
Cardinale et al., 2002).
On the other hand, richness effects can result from the increased probability that species-rich
communities include species with extreme functional values (sampling effect or positive
selection effect; Huston, 1997; Loreau et al., 2001). This mechanism is based on the assumption
that competitive ability and functional impact of species are positively correlated; thereby the
performance of a mixture of species is explained by that of the best-performing species in
monoculture (Tilman et al., 1997). However, when competitive abilities and functional impact
are weakly or negatively correlated, negative selection effects can offset the resource
partitioning and dampen, or even reverse, a positive BEF relationship (Bruno et al., 2006; Jiang
et al., 2008).
Most BEF experiments suggest that sampling effects are the major mechanisms explaining
richness effects (Cardinale et al., 2006; Stachowicz et al., 2007). However, recent evidence
suggests that complementarity operates in natural communities, and also that its expression is
6
Paper II
influenced by other factors (Cardinale et al., 2007; Griffin et al., 2008). The density of
organisms within assemblages influences the strength of intraspecific and interspecific
competition, which in turn can affect the relationship between species richness and ecosystem
properties (Cardinale et al., 2004; Weis et al., 2007). Despite recognition of its importance,
empirical tests of the potential role of density in mediating the richness effects are still rare (but
see Griffin et al., 2008).
While BEF research has largely focused at the species level (Hendriks and Duarte, 2008),
grouping species into functional types allows for a more tractable study of complex systems:
patterns are more consistent because the variability among functionally similar species is
averaged out (Hooper et al., 2005). Here we present the results of field manipulative studies in
which we tested the hypotheses that (1) the number of functional types increases the feeding
rate of assemblages of suspension-feeding invertebrates, (2) these effects are influenced by the
density of the assemblage, and (3) suspension feeders show different preferences for food
particles of different size. In order to accurately test these hypotheses, we separated the effects
of functional richness from the effects of functional identity. We used field-grown epibiota,
grouped together according to functional traits related to their abilities to deplete limiting
resources. A mixture of differently sized microalgae was used as food in order to allow for
resource complementarity on the basis of size-specific filtration rates. Selective feeding in
terms of particle size has been shown in barnacles (Crisp, 1964), bryozoans (Pascoe et al.,
2007), ascidians (Petersen, 2007), and mussels (Rouillon and Navarro, 2003). In addition, we
sought to maximise the ecological relevance of this work by replicating the experiment at two
locations in north-eastern Atlantic coasts.
7
Appendix
2. Methods
2.1. Experimental design
The experiments were conducted at Helgoland Island, south-eastern North Sea and at
Plymouth, in the western English Channel. The original design encompassed the replication of
identical experiments at both locations. Settlement patterns, however, were different between
locations, as the abundance and diversity of settlers were larger at Helgoland than at Plymouth.
Therefore, we did not conduct identically replicated experiments but we used the data from
both locations to address different but complementary hypotheses. The data from Helgoland
were used to test hypothesis 1 (the number of functional types increases the feeding rate of
assemblages of suspension-feeding invertebrates) and hypothesis 2 (density influences the
effect of functional richness). The data from Plymouth were used to test hypotheses 1 (see
above) and 3 (suspension feeders show different preferences for food particles of different
size).
Experiments included three functional types in monoculture and a complete mixture
containing all functional types. Each treatment was replicated five times at Helgoland and three
times at Plymouth. At Helgoland, all treatments were crossed with the factor density (either
high or low) to test for density-dependent diversity effects on filtration rates. Our experimental
design, in which overall density was kept constant across the different diversity treatments (i.e.
a replacement design), allowed a clear partitioning of the effects of functional richness and
identity and detection of non-transgressive and transgressive overyielding (Loreau, 1998; Bruno
et al., 2006). Non-transgressive overyielding occurs when the mixture performance exceeds that
8
Paper II
of the average of its component species in monocultures, while transgressive overyielding
occurs when the mixture performance exceeds that of the best-performing monoculture
(Fridley, 2001).
Species were selected in accordance with natural patterns in the distribution and abundance
of epibiota on vertical surfaces at both locations. Barnacles, bryozoans, ascidians, and mussels
characterise these assemblages (Anger, 1978). We defined the functional types on the basis of
morphological differences that may lead to resource partitioning in terms of particle size (Table
1). For example, barnacles use thoracic appendages (the cirri) to catch and handle food; a
variety of setae present in the cirri allow barnacles to catch plankton across a large size range
from flagellates to small crustaceans (Chan et al. 2008). Ascidians use ciliary pumps to drive
water through a mucus-net that retains the suspended food particles. Bryozoans use lateral cilia
to drive the water toward a filter formed by a band of stiff cilia. Mussels use gill filaments with
lateral cirri that beat against the current (reviewed by Riisgård and Larsen, 2000).
We also considered the individual size as a functional trait, because the morphology of the
organisms influences the physical habitat they occupy and may influence flow patterns and
delivery of food. For example, in assemblages dominated by small and opportunistic species
like bryozoans and colonial ascidians, increasing abundance of massive forms (e.g. mussels)
increases spatial complexity, flow retention, and thus the efficiency of resource capture of the
community (Gili and Coma, 1998). In this study, barnacles, bryozoans, colonial ascidians, and
mussels were used as functional types, but bryozoans and colonial ascidians were classified as
the same type (‘colonials’) because of similar size spectra and modularity (Table 1).
Additionally, at Helgoland we included tiles without suspension feeders but covered with
encrusting algae, which are common in these epibenthic assemblages (Wollgast et al., 2008).
9
Appendix
This was done to provide open settlement space and thus allow for potential effects of early
colonisers on filtration rates. ‘Open space’ was included as monoculture and also within the
mixture assemblages.
The organisms were obtained by exposing settlement tiles (25 × 25 mm Polyvinyl chloride)
to natural colonisation in the water column for 7-8 months. One side of each tile was roughened
to constant texture, on which larvae and propagules were allowed to settle. The other side of the
tile had a Velcro strip used to attach the tile onto a 100 × 100 mm PVC plate. Sixteen
(Helgoland) or nine (Plymouth) tiles were attached onto each plate so that each experimental
unit consisted of a flat grid with 16 or nine interchangeable subunits. The plates were vertically
submerged in the water during May 2006, and recovered on 14 December 2006 (Helgoland)
and 31 January 2007 (Plymouth) to assemble the experimental assemblages.
The tiles were detached from the plates and those tiles covered by a single functional type
were placed into the respective categories. Each monoculture was assembled by reattaching a
group of tiles covered by the same functional type on a plate. For both monocultures and
mixtures, tiles were randomly allocated within each plate. In order to facilitate attachment via
byssal threads, mussels were transplanted onto the tiles by either laying mussels for few days
on the tiles or gluing a short piece of fishing thread onto one of the valves of each mussel and
knotting it to the tile. By keeping the number of tiles constant across functional types, we
assured that replicates for the mixture treatment had the same functional composition.
We manipulated the factor density at Helgoland by varying the number of tiles on each plate:
the high density treatment consisted of plates with 16 tiles covered with organisms, while the
low density treatment consisted of plates with eight tiles covered with organisms plus eight
uncovered tiles. Tiles were suspended in the water column from 20-litre buoys or pontoons so
10
Paper II
that they were constantly immersed at a depth of 1 m below the sea surface and at least 4 m
above the seabed on low spring tides.
2.2. Laboratory determination of clearance rates
Grazing experiments were conducted in laboratory conditions to estimate the filtration rate of
the assemblages growing on the settlement plates. At Helgoland, two filtration rate assays were
conducted on 28 December 2006 and 10 January 2007. At Plymouth we conducted one assay
on 2 April 2007 because it took longer for the assemblages to develop. A second assay at
Plymouth was not possible because of deterioration of the organisms. Filtration rate was
measured as the volume of water cleared per unit of time from a mixed-culture microalgal
suspension. We used four algal strains from the Plymouth Algal Culture Collection:
Cryptomonas rostrella PLY405, Isochrysis galbana PLY680, Prorocentrum micans PLY97A,
and Tetraselmis suecica PLY305. These species will be hereafter referred to by genus.
Microalgae differed in size: Isochrysis (6 Pm max length) < Tetraselmis (15 Pm max length) <
Cryptomonas (25 Pm max length) < Prorocentrum (40 Pm max length). Isochrysis and
Tetraselmis, however, were considered as a single small-sized group because it was impossible
to accurately distinguish them in the suspension. Microalgae were cultured on Provasoli
medium in a climate-controlled room (15° C) in 12:12 h light: dark cycle of 50 Pmol quanta s-1
m-2 (PAR).
For each assay, frames were taken out from the water and the plates were cut loose for
examination of filtration rates in the laboratory. Each plate was submerged for 1 h in an opaque
plastic container (1-litre volume) filled with 800 ml of microalgal suspension. Four containers
11
Appendix
that received 800 ml of the microalgal suspension but no plate served as consumer-free
controls. To measure the filtration rate of each epibenthic assemblage, 25 ml of macroalgal
suspension was taken before and after the assay.
In Helgoland the chlorophyll-a concentration of each suspension sample was measured with
a non-destructive fluorometric technique (BBE Cuvette Fluorometer, BBE Moldaenke GmbH
Germany). This instrument is equipped with light-emitting diodes to excite chlorophyll-a
fluorescence, which in turn is linearly related to the concentration of the pigment. Chlorophylla concentrations measured with fluorometric methods are highly correlated to those measured
with HPLC (Beutler et al., 2002) and counts of cells (Lürling and Verschoor, 2003). In
Plymouth, we used a FACSort flow cytometer (Becton Dickinson, Oxford, UK) equipped with
a 15 mW laser exciting at 488 nm and with a standard filter set up. Specific phytoplankton
groups were discriminated by differences in side scatter and red/orange fluorescence. Flow
rates of the flow cytometer were calibrated daily using quality control beads (0.5 µm,
Polysciences) of a known concentration. Samples were analysed at high flow rate (~ 142 Pl
min-1) to quantify algal size distribution and thus assess size-specific filtration rates. Average
initial Chlorophyll-a concentration at Helgoland was 1.41 Pg l-1 and average initial number of
cells (pooling all phytoplankton species) at Plymouth was ca. 14000 cells ml-1 (Table 1). In
preliminary assays at Helgoland, this Chlorophyll-a concentration enabled us to detect
differences in filtration rate among functional types––lower concentrations were totally
consumed in one hour.
Filtration rates (m, ml h-1 g-1) based either on Chlorophyll-a concentration or number of cells
were calculated using the equation of Fox et al. (1937): m = M (b – a) (n t)-1, where M is the
volume of suspension (ml), a is the logarithmic decrement in the controls (i.e. LN(Ac, t=0) –
12
Paper II
LN(Ac, t=1), Ac, t=0 and Ac, t=1 being the concentration of suspension initially and after time t), b in
the logarithmic decrement in the test suspension (LN(AT, t=0) – LN(AT, t=1)), n is the wet biomass
(g) of each plate (wet weight of the entire experimental unit minus that of the PVC plate and
tiles), and t is the duration of the experiment (1 hour).
2.3. Statistical analysis
Analyses of variance (ANOVAs) were used to test the hypotheses that functional richness
increases feeding rate and that density influences the effect of richness. Planned contrasts were
used to separate the effects of functional richness from identity. Each of the two assays from
Helgoland was analysed using a 2-way ANOVA with the factors assemblage (5 levels: open
space, barnacles, colonials, mussels, or mixture) and density (2 levels: high or low) considered
fixed. Data from Plymouth were analysed using a 1-way ANOVA with the factor assemblage (4
levels: barnacles, colonials, mussels, or mixture) considered fixed. After the ANOVAs, the sum
of squares (SS) of the factor assemblage was partitioned into a planned contrast between the
mixture and the average response of the functional types in monoculture (richness effect). The
residual SS corresponded to the differences among the assemblages made up of a single
functional type and tested the effects of the functional type identity (Bruno et al., 2005).
Transgressive overyielding was then tested using planned contrasts between the mixture and the
best-performing functional type in monoculture. We tested the hypothesis that suspension
feeders have different patterns of resource consumption––i.e. there are differences in the
relative abundance of microalgal species consumed by the three functional types analysed at
Plymouth––using permutational multivariate analysis of variance (PERMANOVA; Anderson,
13
Appendix
2001). In this way, we tested for resource partitioning among functional types on the basis of
changes in the multivariate structure of prey (Griffin et al., 2008). At Plymouth, one replicate
was lost for the ‘colonials’ treatment because of severe weather and so we replaced this missing
value with the mean for the remaining plates of the treatment and one degree of freedom was
subtracted from the residual (Underwood, 1997). Data were square-root transformed to achieve
homogeneity of variances (tested with Levene’s test).
14
Paper II
3. Results
On average, final phytoplankton concentrations were 0.67 Pg l-1 (Chlorophyll-a) at
Helgoland, and ca. 5500 cells ml-1 at Plymouth (Table 1). At both locations, mussels consumed
most of the phytoplankton in one hour; at Helgoland this group totally depleted the resource,
but 20 % of the decrease in phytoplankton concentration was probably due to cell precipitation
or cell damage (Table 1). At Helgoland, ca. 15 % of the microalgae was consumed by
suspension feeders that colonised the open tiles (Table 1).
The results from both assays conducted at Helgoland showed that the effects of functional
richness were negligible in comparison to the significant effects of identity (Table 2). In
addition, there was no support for transgressive overyielding, because consumption rates of the
mixtures were never higher than those of the best-performer monoculture (mussels). In the first
assay conducted at Helgoland, the clearance rate of the mixture was statistically equivalent to
that of mussels (F1, 46 = 0.08, p = 0.78; from planned contrast), while in the second assay the
performance of the mixture was significantly lower than the performance of mussels (F1, 46 =
18.12, p < 0.01; from planned contrast).
Density effects were significant only in the first Helgoland assay (Helgoland 1 in Table 2).
The filtration rates of the high density assemblages were higher than those of the low density
assemblages (Fig. 1). A non-significant interaction term indicated that the effects of functional
richness and identity were independent from those of density (Table 2).
The results from the work conducted at Plymouth confirmed the trends from the experiment
at Helgoland and showed that the effect of functional richness was weak compared to that for
identity (Fig. 2, Table 2). Planned contrasts between the mixture and the best-performing
15
Appendix
monoculture showed no significant difference in filtration rates (i.e. we found no evidence for
transgressive overyielding; F1, 9 = 0.3, p = 0.59). Multivariate analyses revealed distinct patterns
in resource use among functional types (Fig 2, unbalanced PERMANOVA, pseudo-F2, 5 =
23.66, p < 0.01). Barnacles tended to prefer the larger microalgal species, while mussels
showed maximum filtration rates on microalgae from all three size classes.
16
Paper II
4. Discussion
Results from Helgoland and Plymouth suggest that efficiency in resource use was strongly
affected by the identity but not by the richness of functional types. Our findings are similar to
those from studies in terrestrial and marine habitats showing that identity and compositional
effects are common, but that richness effects and transgressive overyielding are more evasive
phenomena (Bruno et al., 2005; Hooper et al., 2005; O'Connor and Crowe, 2005; Bruno et al.,
2006). Our results might have important ecological implications, because they suggest that the
loss or gain of particular species has the strongest impact on the functioning of epibenthic
ecosystems instead of richness per se.
The high efficiency of the blue mussel in removing the majority of algae may have
minimised complementarity in the mixed assemblage and prevented a richness effect. Also, the
filtration rate of mussels probably explained that of the mixture assemblages, because the
performance of both treatments tended to be similar––i.e. a positive selection effect. However,
the assessment of the relative contribution of complementarity and selection effects (either
positive or negative) requires measuring the performance of each species in the mixture
(Stachowicz et al., 2007). Experiments using such techniques have shown that putative positive
selection effects (because mixtures yielded the same than the best-performing species) were
actually the outcome of resource complementarity being offset by strong negative selection of
the best-performing species (Bruno et al., 2005; Reusch et al., 2005; Bruno et al., 2006).
Negative selection effects occur when the competitive ability and functional impact of species
are weakly or negatively correlated (Loreau and Hector, 2001; Jiang et al., 2008). Such tradeoffs between competitive ability and functional impact are widespread in nature (e.g. keystone
17
Appendix
species; Lyons et al., 2005) and negative selection effects may be particularly frequent for other
than biomass functions (Jiang et al., 2008). Selection effects and positive species interactions
(including resource complementarity and facilitation) can operate simultaneously or
sequentially (Hooper et al., 2005).
On the other hand, we found that the effects of functional identity were independent of
density. In addition, in one assay density increased the consumption rate of monocultures and
mixtures, suggesting that there were enough resources to prevent strong consumptive
competition even at a high density of suspension feeders––if phytoplankton concentration
would have been a limiting factor, then increased density would have decreased the per capita
(per gram of tissue) consumption rates in monocultures because of consumptive competition
among conspecifics. The absence of strong competition, in terms of resource use, probably
contributed to the insignificant effects of functional richness on filtration rates. For example,
Griffin and collaborators (2008) showed that the effects of predator richness were detectable
only at high predator density where competitive interactions were intensified. Similarly, early
studies showing significant richness effects have been conducted usually in nutrient-poor
systems (e.g. Tilman et al., 2001), where consumptive competition was probably high.
Collectively, the results from these experiments stress the importance of competition in linking
diversity and ecosystem functioning.
Our experiments should be considered in context with natural assemblages. Even when we
sought to maximise the ecological relevance of the study by replicating the experiment at two
locations of the NE Atlantic, the actual impact of functional types on filtration rates may differ
from our results. For instance, predators were excluded by suspending the set up in the water
column. If predators are present, richness effects may be stronger due to selective predation on
18
Paper II
the dominant species (Duffy, 2002). Also, we used a maximum of six species, while sessile
epibenthic assemblages can contain more that 30 species (Wollgast et al., 2008). The richness
effect may be stronger at higher levels of species richness, as shown by several studies on
mobile grazers (Stachowicz et al., 2007); yet, richness effects at low number of species have
been shown by experiments on primary producers (Loreau et al., 2001; Tilman et al., 2001),
invertebrate predators (Griffin et al., 2008), and invertebrate suspension feeders (Cardinale et
al., 2002). Whether the effect of richness is greater at higher richness levels is still an open
question.
Despite these limitations, we have presented a novel approach to assess the role of resource
partitioning in the relationship between diversity and ecosystem functioning. The ultimate goal
is to predict the ecological consequences of the widespread human-driven alterations of biota;
our work therefore represents a small piece of a much larger puzzle.
Acknowledgements
Thanks are due to J. Ellrich and M. Stillfried for their support during the clearance rate
assays, to A. Kraberg and A. Wagner for helping in culturing the algal, and to K. Boos for
valuable discussion. Two anonymous reviewers improved an early version of this manuscript.
This work benefited from discussions with members of the MarBEF responsive mode project
BIOFUSE. Financial support by the Alfred-Wegener-Institute for Marine and Polar Research to
N.V. is acknowledged. Thanks to the University of Plymouth Diving and Marine Centre for
facilitating in-water trails in Plymouth.
References
19
Appendix
Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance.
Austral Ecol. 26, 32-46.
Anger, K., 1978. Development of a subtidal epifaunal community at the island of Helgoland.
Helg. Mar. Res. 31, 457-470.
Beutler, M., Wiltshire, K.H., Meyer, B., Moldaenke, C., Luring, C., Meyerhofer, M., Hansen,
U.P., Dau, H., 2002. A fluorometric method for the differentiation of algal populations
in vivo and in situ. Photosynth. Res. 72, 39-53.
Bruno, J.F., Boyer, K.E., Duffy, J.E., Lee, S.C., Kertesz, J.S., 2005. Effects of macroalgal
species identity and richness on primary production in benthic marine communities.
Ecol. Lett. 8, 1165-1174.
Bruno, J.F., Lee, S.C., Kertesz, J.S., Carpenter, R.C., Long, Z.T., Duffy, J.E., 2006. Partitioning
the effects of algal species identity and richness on benthic marine primary production.
Oikos 115, 170-178.
Byrnes, J.E., Reynolds, P.L., Stachowicz, J.J., 2007. Invasions and extinctions reshape coastal
marine food webs. PLoS ONE 2, e295.
Cardinale, B.J., Palmer, M.A., 2002. Disturbance moderates biodiversity-ecosystem function
relationships: Experimental evidence from caddisflies in stream mesocosms. Ecology
83, 1915-1927.
Cardinale, B.J., Palmer, M.A., Collins, S.L., 2002. Species diversity enhances ecosystem
functioning through interspecific facilitation. Nature 415, 426-429.
Cardinale, B.J., Ives, A.R., Inchausti, P., 2004. Effects of species diversity on the primary
productivity of ecosystems: extending our spatial and temporal scales of inference.
Oikos 104, 437-450.
Cardinale, B.J., Srivastava, D.S., Duffy, J.E., Wright, J.P., Downing, A.L., Sankaran, M.,
Jouseau, C., 2006. Effects of biodiversity on the functioning of trophic groups and
ecosystems. Nature 443, 989-992.
Cardinale, B.J., Wrigh, J.P., Cadotte, M.W., Carroll, I.T., Hector, A., Srivastava, D.S., Loreau,
M., Weis, J.J., 2007. Impacts of plant diversity on biomass production increase through
time because of species complementarity. P. Natl. Acad. Sci. USA 104, 18123-18128.
Chapin, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds, H.L.,
Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C., Diaz, S., 2000.
Consequences of changing biodiversity. Nature 405, 234-242.
Christensen, N.L., Bartuska, A.M., Brown, J.H., Carpenter, S., D'Antonio, C., Francis, R.,
Franklin, J.F., MacMahon, J.A., Noss, R.F., Parsons, D.J., Peterson, C.H., Turner, M.G.,
Woodmansee, R.G., 1996. The report of the Ecological Society of America committee
on the scientific basis for ecosystem management. Ecol. Appl. 6, 665-691.
Crisp, D.J., 1964. An assessment of plankton grazing by barnacles. In: Crisp, D.J. (Ed.),
Grazing in terrestrial and marine environments. Blackwell Scientific Publications,
Oxford.
Duffy, J.E., 2002. Biodiversity and ecosystem function: the consumer connection. Oikos 99,
201-219.
Fox, D.L., Sverdrup, H.U., Cunningham, J.P., 1937. The rate of water propulsion by the
California mussel. Biol. Bull. 72, 417-438.
Fridley, J.D., 2001. The influence of species diversity on ecosystem productivity: how, where,
and why? Oikos 93, 514-526.
20
Paper II
Gili, J.M., Coma, R., 1998. Benthic suspension feeders: their paramount role in littoral marine
food webs. Trends Ecol. Evol. 13, 316-321.
Griffin, J.N., De la Haye, K.L., Hawkins, S.J., Thompson, R.C., Jenkins, S.R., 2008. Predator
diversity and ecosystem functioning: Density modifies the effect of resource
partitioning. Ecology 89, 298-305.
Hendriks, I.E., Duarte, C.M., 2008. Allocation of effort and imbalances in biodiversity
research. J. Exp. Mar. Biol. Ecol. 360, 15-20.
Hooper, D.U., Chapin, I., F. S., Ewel, J.J., Hector, A., Inchausti, P., Lavorel, S., Lawton, J.H.,
Lodge, D.M., Loreau, M., Naeem, S., Schmid, B., Setälä, H., Symstad, A.J.,
Vandermeer, J., Wardle, D.A., 2005. Effects of biodiversity on ecosystem functioning: a
consensus of current knowledge. Ecol. Monogr. 75, 3-35.
Huston, M.A., 1997. Hidden treatments in ecological experiments: re-evaluating the ecosystem
function of biodiversity. Oecologia 110, 449-460.
Jiang, L., Pu, Z., Nemergut, D.R., 2008. On the importance of the negative selection effect for
the relationship between biodiversity and ecosystem functioning. Oikos 117, 488-493.
Lesser, M.P., Witman, J.D., Sebens, K.P., 1994. Effects of flow and seston availability on scope
for growth of benthic suspension-feeding invertebrates from the gulf of maine. Biol.
Bull. 187, 319-335.
Loreau, M., 1998. Separating sampling and other effects in biodiversity experiments. Oikos 82,
600-602.
Loreau, M., 2000. Biodiversity and ecosystem functioning: recent theoretical advances. Oikos
91, 3-17.
Loreau, M., Hector, A., 2001. Partitioning selection and complementarity in biodiversity
experiments. Nature 412, 72-76.
Loreau, M., Naeem, S., Inchausti, P., Bengtsson, J., Grime, J.P., Hector, A., Hooper, D.U.,
Huston, M.A., Raffaelli, D., Schmid, B., Tilman, D., Wardle, D.A., 2001. Ecology Biodiversity and ecosystem functioning: Current knowledge and future challenges.
Science 294, 804-808.
Lürling, M., Verschoor, A.M., 2003. F0-spectra of chlorophyll fluorescence for the
determination of zooplankton grazing. Hydrobiologia 491, 145-157.
Lyons, K.G., Brigham, C.A., Traut, B.H., Schwartz, M.W., 2005. Rare species and ecosystem
functioning. Conserv. Biol. 19, 1019-1024.
Navarrete, S.A., Wieters, E.A., Broitman, B.R., Castilla, J.C., 2005. Scales of benthic-pelagic
coupling and the intensity of species interactions: From recruitment limitation to topdown control. P. Natl. Acad. Sci. USA 102, 18046-18051.
Nielsen, T.G., Maar, M., 2007. Effects of a blue mussel Mytilus edulis bed on vertical
distribution and composition of the pelagic food web. Mar. Ecol. Prog. Ser. 339, 185198.
O'Connor, N.E., Crowe, T.P., 2005. Biodiversity loss and ecosystem functioning:
distinguishing between number and identity of species. Ecology 86, 1783-1796.
Pascoe, P.L., Parry, H.E., Hawkins, A.J.S., 2007. Dynamic filter-feeding responses in fouling
organisms. Aquat. Biol. 1, 177-185.
Petersen, J.K., 2007. Ascidian suspension feeding. J. Exp. Mar. Biol. Ecol. 342, 127-137.
Reusch, T.B.H., Ehlers, A., Hammerli, A., Worm, B., 2005. Ecosystem recovery after climatic
extremes enhanced by genotypic diversity. P. Natl. Acad. Sci. USA 102, 2826-2831.
21
Appendix
Riisgård, H.U., Larsen, P.S., 2000. Comparative ecophysiology of active zoobenthic filter
feeding, essence of current knowledge. J. Sea. Res. 44, 169-193.
Rouillon, G., Navarro, E., 2003. Differential utilization of species of phytoplankton by the
mussel Mytilus edulis. Acta. Oecol. 24, S299-S305.
Schwartz, M.W., Brigham, C.A., Hoeksema, J.D., Lyons, K.G., Mills, M.H., van Mantgem,
P.J., 2000. Linking biodiversity to ecosystem function: implications for conservation
ecology. Oecologia 122, 297-305.
Stachowicz, J.J., Bruno, J.F., Duffy, J.E., 2007. Understanding the effects of marine
biodiversity on communities and ecosystems. Annu. Rev. Ecol. Evol. S. 38, 739-766.
Thomas, J.A., Telfer, M.G., Roy, D.B., Preston, C.D., Greenwood, J.J.D., Asher, J., Fox, R.,
Clarke, R.T., Lawton, J.H., 2004. Comparative losses of British butterflies, birds, and
plants and the global extinction crisis. Science 303, 1879-1881.
Tilman, D., Lehman, C.L., Thomson, K.T., 1997. Plant diversity and ecosystem productivity:
Theoretical considerations. P. Natl. Acad. Sci. USA 94, 1857-1861.
Tilman, D., Reich, P.B., Knops, J., Wedin, D., Mielke, T., Lehman, C., 2001. Diversity and
productivity in a long-term grassland experiment. Science 294, 843-845.
Underwood, A.J., 1997. Experiments in ecology. Their logical design and interpretation using
analysis of variance. Cambridge University Press, Cambridge, 504 pp.
Weis, J.J., Cardinale, B.J., Forshay, K.J., Ives, A.R., 2007. Effects of species diversity on
community biomass production change over the course of succession. Ecology 88, 929939.
Wollgast, S., Lenz, M., Wahl, M., Molis, M., 2008. Effects of regular and irregular temporal
patterns of disturbance on biomass accrual and species composition of a subtidal hardbottom assemblage. Helg. Mar. Res. 62, 309-319.
22
Paper II
Figure legends
Fig. 1. Results of two assays conducted at Helgoland to test the effects of functional richness,
identity, and density on the feeding efficiency of epibiota. Outcomes of statistical tests are
given in Table 2. Values are given as mean ± SEM (n = 5).
Fig. 2. Results of the experiment at Plymouth to investigate the rate of filtration of each
microalga for individual suspension feeders and a mixed assemblage. Each replicate containing
either a monoculture or a mixture was supplied with a mixed suspension containing all
microalgae in equal proportions. Outcomes of statistical tests are given in Table 2. Values are
given as mean ± SEM (n = 3).
23
Barnacles
Colonials
Mussels
Mixture
B
Mussels
Barnacles
Colonials
A
Functional
type
1.41
1.36
1.43
1.43
0.92
0.54
0.00
0.49
35
61
100
65
24
Cirri beat
Cilia pump water into a
mucus-net
Ciliary sieving
Gill filaments with lateral
cilia
Feeding mode1
14189
13311
14914
13113
10455
3150
547
729
26
76
96
94
Plymouth (cell concentration [cells ml-1])
%-removed
t0
t1
<10 mm
>10 mm
Colonials
Solitary
Helgoland (Chlorophyll-a concentration [Pg l-1])
%-removed
t0
t1
<10 mm
<10 mm
Solitary
Colonials
Elminius modestus, Balanus crenatus
Colonial ascidians: Botryllus schlosseri,
Diplosoma sp.
Bryozoans: Cryptosula pallasiana
Mytilus edulis
Size of
individuals
Modularity
Species
precipitation or cell damage). NI: not investigated. 1 Source: Riisgård and Larsen (2000)
encrusting algae. Controls were used to quantify the changes in microalgal concentration that can not be explained by filtration activity (e.g. cell
concentration consumed in one hour by each functional type and mixtures is given (%-removed). Open space corresponds to tiles covered by
Average concentration of phytoplankton initially (t0) and one hour after (t1) the filtration rate assays (B). The percent from initial phytoplankton
Species of suspension-feeding invertebrates grouped as functional types according to modularity, size of individuals, and feeding mode (A).
Table 1
Appendix
Open space
Controls
1.49
1.37
0.98
1.09
34
20
25
NI
14309
NI
12591
NI
12
Paper II
Appendix
Table 2
ANOVAs of three experiments testing the effect of invertebrate functional richness and
identity on community-level filtration rates. The factor assemblage included three or four
functional types of suspension feeders in monocultures and one mixed assemblage. At
Helgoland, density (either low or high) was included as an orthogonal factor. Richness and
identity effects were tested with orthogonal planned contrasts (shown indented). Data were
square-root transformed before the analysis to ensure homogeneity of variances (Levene’s
test, P > 0.05)
Source of variation
SS
df
MS
F
P
Helgoland 1
Assemblage
Richness
Identity
Density
Density × Assemblage
Density × Richness
Density × Identity
Residual
15.663
1.559
14.104
1.970
1.575
0.475
1.099
18.124
4
1
3
1
4
1
3
40
3.916
1.559
4.701
1.970
0.394
0.475
0.366
0.453
8.642
2.092
10.376
4.347
0.869
1.049
0.809
<0.0001
0.1546
<0.0001
0.0435
0.4910
0.3118
0.4965
Helgoland 2
Assemblage
Richness
Identity
Density
Density × Assemblage
Density × Richness
Density × Identity
Residual
22.504
0.070
22.434
0.230
0.663
0.347
0.317
8.053
4
1
3
1
4
1
3
40
5.626
0.070
7.478
0.230
0.166
0.347
0.106
0.201
27.945
0.107
37.149
1.140
0.824
0.518
0.525
<0.0001
0.7450
<0.0001
0.2920
0.5179
0.4755
0.6679
36.398
1.383
41.545
<0.0001
0.2669
0.0001
Plymouth
Assemblage
149.913
3
49.971
Richness
19.544
1
19.544
Identity
130.369
2
65.185
Residual
10.983
7a
1.569
a
Degrees of freedom were corrected for missing data
26
8
Average monoculture
Assay 1
Mixture
Mussels
10
Colonials
10
Barnacles
Open space
Clearance rate [ml h-1 g-1]
Paper II
Low density
High density
6
4
2
0
Assay 2
8
6
4
2
0
Fig. 1. Valdivia et al
27
Appendix
Clearance rate [ml h-1 g-1]
100
80
Isochrysis-Tetraselmis complex
Cryptomonas
Prorocentrum
60
40
20
Average monoculture
Mixture
Mussels
Colonials
Barnacles
0
Fig. 2. Valdivia et al
28
Paper III
Nelson Valdivia*, and Markus Molis (Under review) Species compensation buffers community
stability against the loss of an intertidal habitat-forming rockweed. Marine Ecology Progress
Series
Biologische Anstalt Helgoland, Alfred Wegener Institute for Polar and Marine Research,
Section Seaweed Biology, Kurpromenade 201, 27498 Helgoland, Germany
* Corresponding author
Email [email protected]
Tel. ++49(0)47258193294
Fax ++49(0)47258193283
Appendix
2
Paper III
ABSTRACT: Most of the research on ecosystem stability has focused on the relationship
between biodiversity and community temporal variability, yet the role of biotic interactions in
maintaining stability has received less attention. Compensatory changes in species populations,
such that the abundances of some species increase while those of others decrease when the
environment changes, can maintain a steady state between resource supply and resource
consumption. In a fluctuating environment, therefore, the variability of species abundances may
be larger than the variability of the community abundance. Here, we show that removal of a key
structural component of hard-bottom communities, the canopy-forming alga Fucus serratus,
significantly increased the temporal variability of populations but not of communities. Results
from factorial experiments replicated at two shores of Helgoland Island, NE Atlantic,
consistently suggest that environmental changes resulting from the canopy removal triggered
compensatory dynamics of species with differing functional traits. The provision of additional
open substratum (i.e. additional mechanical disturbances) did not influence these patterns.
These results agree with previous studies involving the removal of canopy-forming species, but
contradict recent analyses suggesting that species compensation is rare in several ecosystems.
We suggest that compensatory dynamics will have a critical role in maintaining the stability of
systems where biological habitat amelioration has both positive and negative effects on other
species.
Key words: bioengineering, canopy, disturbance, Fucus serratus, intertidal ecology, temporal
variability, species compensation, stability
3
Appendix
INTRODUCTION
Stability is a fundamental aspect of ecological systems (Worm & Duffy 2003). In particular,
ecosystem stability can be highly valuable because of the societal dependence on the good and
service that ecosystems provide to mankind (Costanza et al. 1997, Armsworth & Roughgarden
2003). Accordingly, it is not surprising that the stability of ecological processes–such as the
temporal persistence of communities and their resistance to environmental change–had drawn
the attention of ecologists and policymakers for decades (e.g. MacArthur 1955, Christensen et
al. 1996). Studies on ecosystem stability have focused on the effects of decreasing diversity on
the temporal variability of community and population properties, such as biomass and resource
use (e.g. Tilman & Downing 1994, McCann 2000, Ptacnik et al. 2008). However, the effects of
biotic interactions on ecosystem stability have received less attention (reviewed by Hooper et
al. 2005).
Models on the effect of biotic interactions on ecosystem processes predict that stability is
maintained via compensatory population dynamics, such that the contribution of some species
to ecosystem properties decreases while that of others increases (e.g. Austin & Cook 1974, Ives
et al. 1999). This occurs because species respond differently to environmental changes, thus the
role of stressed or disturbed species may be assumed by unharmed species if the latter provide
lost functional traits (Yachi & Loreau 1999). Species compensation can be revealed by a high
temporal variability of species abundances relative to the variability of the community
abundance, and also by prevalent negative covariances within the community (Schluter 1984,
Micheli et al. 1999). These ideas however have been tested by relatively few empirical studies
4
Paper III
(e.g. Ernest & Brown 2001, Bai et al. 2004, Vasseur & Gaedke 2007), and recent analyses
suggest that species compensation is actually rare in real communities (Houlahan et al. 2007).
Habitat-forming species provide an opportunity to test whether species compensation occurs
within natural communities. Canopy-forming algae are key structural elements on temperate
rocky shores, where they modify the environment and facilitate or suppress other species (e.g.
Irving & Connell 2006, Lilley & Schiel 2006, Morrow & Carpenter 2008). Under the stressful
conditions that characterise the intertidal zone, canopy-forming algae can ameliorate the habitat
by shading, reducing desiccation, and buffering temperatures (Bertness et al. 1999, Lilley &
Schiel 2006). In addition, canopies can also reduce the accumulation of sediments (Kennelly &
Underwood 1993). If the canopy-mediated habitat amelioration facilitates some species but
suppress others, then species compensation may be recurrent within these assemblages.
However, these effects of canopies on community stability may interact with those of other
factors. Space available for settlement is often a limiting resource for sessile epibenthic
organisms (Connell 1961, Connolly & Muko 2003), but it can be provided by mechanical
disturbances that remove biomass from the community (Shea et al. 2004). Since canopies may
limit the subset of species that are able to colonise the substratum (Jenkins et al. 2004), the
effects of disturbances on the understorey may be stronger after the removal of the canopyforming species. On the other hand, the putative effects of canopy removal can be exacerbated
by additional provisions of settlement space. Therefore, the effect of canopy removal on
community stability may depend on disturbance and vice versa.
Here, we present the results of a manipulative field study on the effect of the canopy-forming
algae Fucus serratus (here after Fucus) on the stability of intertidal hard-bottom communities.
We explored (1) whether the removal of Fucus affects the temporal variability of community
5
Appendix
abundance and species populations, and (2) whether these effects depend on mechanical
disturbances that provide additional space. Factorial experiments were replicated at two
intertidal sites of different wave exposure to test for the generality of patterns. We found that
the removal of Fucus consistently increased the variability of species populations without
affecting the variability of communities, because of compensatory dynamics of species with
different functional traits.
MATERIALS AND METHODS
Study sites. The experiment was replicated on two differently wave-exposed rocky shores in
the natural reserve of Helgoland Island (North Sea, NE Atlantic). ‘Westwatt’ is exposed to
strong prevailing south-westerly winds, while ‘Nordostwatt’ is protected by a 250 m long
concrete jetty. Losses in dry mass of domes (Ø at base = 62 cm) made of Plaster of Paris that
were deployed during two 3-day periods (20 – 23 April 2007 and 8 – 11 June 2007) were
significantly higher at Westwatt than at Nordostwatt (repeated measures ANOVA: F1, 8 =
13.29, P < 0.01), irrespectively of the period (F1, 8 = 1.92, p = 0.2).
At both sites, dense Fucus stands extend from the lower intertidal to the shallow sublittoral.
The understorey of the Fucus beds is dominated by encrusting coralline algae (mostly
Phymatolithon spp.), and the turf-forming algae Cladophora rupestris, Chondrus crispus, and
Corallina officinalis (Bartsch & Tittley 2004). The most frequent sessile invertebrates are
Dynamena pumila, Spirorbis spirorbis, and Electra pilosa, while conspicuous mobile
consumers include Carcinus maenas and several species of periwinkles (Reichert & Buchholz
2006). During spring and summer, foliose and filamentous ephemeral algae like Ulva spp.,
6
Paper III
Dumontia contorta, Ectocarpales, and seasonal Cladophorales are abundant in gaps between
Fucus patches (Janke 1990).
Experimental design and set-up. During March 2006, twenty 0.3 × 0.3 m² plots with Fucus
cover of ca. 90 % were permanently marked with stainless screws at each site. This small plot
size was used to minimise the impact of our manipulations on the natural Fucus population.
Nevertheless, plots of the same size were also used in other experiments involving the removal
of fucoid canopies (Moore et al. 2007). In each site, holdfasts and erect fronds of Fucus were
removed from 10 randomly selected plots using a knife although avoiding damage the
understorey. Edge effects in the removal plots were avoided by trimming all Fucus plants
within a margin of ca. 40 cm wide along each plot. Fucus recruits were regularly removed
throughout the 18-month study. Mechanical disturbance treatments, consisting of a biomass
removal with 50 % of the effort required to remove all organisms of the plot, were randomly
applied to half of the canopy-present plots and half of the canopy removal plots. Thirty-six
passages of a 2 cm wide chisel were needed to remove all organisms (excluding encrusting
algae and organisms occurring in small crevices) from a 0.09 cm2 area. Eighteen passages of
the chisel were applied on each disturbed plot therefore. The cover of bare rock, measured 1-3
days after applying the disturbances, significantly increased as result of the treatment (F1, 32 =
5.04, p = 0.03). All manipulations were conducted during low tide.
Sampling. The percent cover of each macrobenthic species was estimated per plot to the
nearest 1 % by the same observer before and 1-3 days after the manipulations in March 2006.
Species covers were subsequently estimated every 3 months following the initial sampling. Due
to the multilayered structure of the assemblages, total percent cover could well exceed 100 %.
Species with <1 % cover were uniformly recorded with 0.5 % abundance. Using the same
7
Appendix
method, we quantified the percent cover of particulate material (sediment) on each plot. Visual
estimation of percent surface cover of sediment is one of the principal methods used to quantify
sedimentation in rocky intertidal habitats (Airoldi 2003). Taxa were identified to the lowest
possible taxonomic level in the field, but ambiguous taxa were identified in the laboratory using
samples collected from areas adjacent to the plots. Four taxa were identified to genus level:
Porphyra sp., Phymatolithon spp., Sagartia spp., and Ulva spp. Small burrowing spionids were
grouped as family Spionidae and Ectocarpales were identified to order level. For some
analyses, species were grouped into five functional types: (1) ‘encrusting algae’, comprising
Haemescharia hennedyi, Hildenbrandia rubra, and Phymatolithon spp.; (2) ‘ephemeral algae’,
dominated by Ulva spp., Cladophora sericea, Dumontia contorta, and Ectocarpales; (3) ‘turfforming algae’, dominated by Chondrus crispus, Cladophora rupestris, Corallina officinalis,
and Mastocarpus stellatus; (4) ‘sessile invertebrates’, dominated by Dynamena pumila,
Sagartiogeton laceratus, Semibalanus balanoides, Spionidae spp., and Spirorbis spirorbis; and
(5) ‘mobile consumers’, dominated by Littorina obtusata and Littorina littorea. The congeneric
species C. sericea and C. rupestris were classified as different functional types because of
different life histories. Cladophora sericea is a seasonal species, while C. rupestris is perennial
(Bartsch & Tittley 2004).
Statistical analyses. We tested the separate and interactive effects of canopy removal and
mechanical disturbance on two facets of community variability: the variability of community
abundance and the variability of species populations. Both types of variance were calculated
from the 7 repeated measures of cover of each taxon over the course of the experiment (18
months). A single value per type of variance was calculated for each plot and these values were
then considered independent in the analyses. We calculated the temporal variance in total
8
Paper III
community cover, which in turn can be expressed as the sum of all of the species variances plus
the sum of all of the pair-wise species covariances (Schluter 1984). This partition provides a
useful tool for detecting compensatory dynamics within communities (Tilman 1999, Houlahan
et al. 2007). We used the variance and summed variance as measures of community-level
variability, and the summed covariance as a measure of population-level variability and
compensation. If species undergo compensatory dynamics, then the summed covariance of all
species-pairs will be negative. In addition, we explored the temporal variability in the identity
and relative abundance of the component species (i.e. species composition). For each plot, we
calculated a Bray-Curtis dissimilarity matrix across all sample dates. The average of the BrayCurtis matrix provided a single value of compositional variability per plot. The effects of
canopy removal and disturbance on each of the measures of temporal variability were analysed
using 3-way mixed ANOVA with the factor site (2 levels: Nordostwatt or Westwatt) considered
random, and the factors canopy (2 levels: present or removed) and disturbance (2 levels:
undisturbed or disturbed) both considered fixed. Homogeneity of variances was graphically
explored and tested using Cochran’s test; when necessary data were ln-transformed to meet the
assumptions. Tukey HSD post-hoc test was used for unplanned comparisons.
We showed the temporal changes in species composition using non-metric multivariate
(nMDS) ordination plots. For each site, we separately plotted centroids of the 28 time (7 levels)
× canopy (2 levels) × disturbance (2 levels) cells, because there were too many observation
points to view in a single ordination. Centroids were computed as the averages of principal
coordinates calculated from Bray-Curtis matrices. nMDS ordinations were then plotted on the
basis of the Euclidean distance between each pair of centroids (see Terlizzi et al. 2005 for
details).
9
Appendix
Finally, we conducted a redundancy analysis (RDA) to show the correlation between pairs of
functional types in function of the experimental factors. In addition, we included the percent
cover of sediment as response variable. RDA focuses on the relationship between two sets of
variables (i.e. the matrix of functional types and the matrix of factors) as summarised in a
matrix of regression coefficients. The results of the RDA were visualised in a correlation biplot, in which each response variable (i.e. each functional type and sediment) was standardised
to zero mean and unit variance (Ter Braak & Looman 1994). All statistical analyses were
conducted using the R environment version 2.7.2 (R Development Core Team 2008). nMDS
ordinations were plotted using PRIMER version 5.
RESULTS
The variability at the community level was neither affected by canopy removal nor by
disturbance (Fig. 1a, b; Table 1). The variance of total community cover (ln-transformed) was
independent of the factor site, while the summed variance was significantly higher at
Nordostwatt than at Westwatt (Fig. 1a, b; Table 1).
Canopy removal strongly affected the variability of species populations. On the one hand, the
summed covariance became significantly more negative as result of the canopy removal,
irrespectively of site and disturbance (Fig. 1c, Table 2). On the other hand, the variability of
species composition (Bray-Curtis dissimilarities) was influenced by canopy removal and site
(Fig. 1d, Table 2). Separate 2-way ANOVAs conducted for each site showed that canopy
removal significantly increased the Bray-Curtis dissimilarities at Nordostwatt (Fig. 1d; F1, 16 =
10.69, p < 0.01) irrespectively of disturbance (F1, 16 < 0.01, p = 0.94). At Westwatt, however, a
10
Paper III
significant canopy by disturbance interaction (Fig. 1d, F1, 16 = 6.13, p = 0.02) indicated that the
destabilising effect of disturbance was significant only on the canopy removal plots (Fig 1d, p =
0.03; from Tukey HSD multiple comparisons).
The nMDS plots show the strong influence of canopy removal on the multivariate structure
of these assemblages (Fig. 2). At both sites, temporal changes of species composition were
larger on canopy-removed plots than on canopy-present plots. These patterns showed a seasonal
component, as communities tended to diverge during summer months (labels 2 and 6 in Fig. 2),
but to converge during the winter and early spring months (labels 1, 4 and 5 in Fig. 2).
The response of species composition to the canopy removal came from the differing patterns
of functional types and it was also related to changes in the physical environment (Fig. 3). The
patterns were generally similar between sites (Fig. 3). The RDA bi-plots show that the highest
abundances of encrusting algae and sessile invertebrates occurred on the canopy-present plots,
but the lowest on the removal plots. Conversely, highest abundances of ephemeral algae and
sediments occurred on the removal plots, but the lowest on the canopy-present plots.
Accordingly, there was a strong negative correlation between encrusting and ephemeral algae,
and there were strong positive correlations between encrusting algae and sessile invertebrates
and between ephemeral algae and sediments (Fig. 3). This analysis also detected a seasonal
pattern: the largest abundances of encrusting algae and sessile invertebrates occurred during
early spring, autumn, and winter (labels 1, 3, and 4 in Fig. 3), while the largest abundances of
ephemerals occurred during summer and the spring 2007 (labels 5 and 6 in Fig. 3). Turfforming algae and mobile consumers were not clearly related to the canopy treatments. On the
other hand, there was no consistent relationship between the disturbance treatments and the
11
Appendix
abundance of functional types and sediment. For example, mobile consumers increased during
the first summer in the disturbed plots, but only at one site (label 2 in Fig. 3a).
DISCUSSION
Here we have shown that the removal of a key dominant species, Fucus serratus, significantly
increased the variability in species populations without affecting the variability in community
abundance. Negative covariances were persistent within the communities, and they became
more negative due to the canopy removal. This indicates that compensatory dynamics, such that
the abundance of some species increases while that of others decreases, were strengthened by
the canopy removal. The removal of Fucus encouraged ephemeral algae to proliferate, but it
discouraged encrusting algae and sessile invertebrates. Additional provisions of free space
(mechanical disturbances) had limited effects on community and population variability.
Experiments replicated at two sites showed virtually the same patterns of variability, suggesting
that the effects of canopy removal can be consistent across the spatial variability of this system.
These results suggest that compensatory dynamics maintain the community stability when
bioengineering has differing (positive and negative) effects on other species.
Compensatory responses to biological habitat amelioration
Species compensation was due to the canopy-mediated changes in the physical habitat. The
removal of Fucus increased the cover of sediments, which in turn can be an important source of
stress for hard-bottom communities (Airoldi 2003). In the intertidal rocky shores of Helgoland,
12
Paper III
additionally, the removal of Fucus can increase the understorey temperature in 9° C and the
amount of irradiance in 400 % in relation to areas covered by canopies (D. Kohlmeier & K.
Bischof, unpubl. data). Moreover, the canopy-mediated amelioration of osmotic stressors like
temperature and water evaporation is well documented in the literature (e.g. Bertness et al.
1999, Lilley & Schiel 2006, Moore et al. 2007). Increased light and sedimentation led to
blooms of ephemeral algae, as shown in previous work on intertidal rocky habitats (e.g. Airoldi
2003, Lilley & Schiel 2006). At the same time, these changes in the physical conditions
probably had negative effects on the fitness of encrusting algae and sessile invertebrates. The
development of encrusting coralline algae requires shaded conditions, and probably also
sediment-free substrate (Steneck 1986, Connell 2003). Similarly, several species of sessile
invertebrates, such as sea anemones and colonial bryozoans, prefer shaded microhabitats in the
intertidal at Helgoland (Janke 1986) and perhaps do benefit from the habitat amelioration by
Fucus during periods of high thermal and irradiance stress. On the other hand, canopies can
control the abundance of grazers (Bertness et al. 1999, Jenkins et al. 2004), which in turn can
strongly influence community structure by controlling algae abundances (Aguilera & Navarrete
2007). In our study, nevertheless, the abundance of consumers was not affected by the canopy
removal. Therefore, it is likely that compensatory dynamics were related to changes in abiotic
conditions but unrelated to changes in the strength of consumption. Similar non-trophic habitat
associations have been documented by work on subtidal, intertidal, and terrestrial canopies (e.g.
Irving & Connell 2006, Lilley & Schiel 2006, Felton et al. 2008).
The differing effect of biological habitat amelioration led to compensatory dynamics that
maintained the stability of community abundance, in accordance with analyses of time series
showing high variability of populations but low variability of communities (Ernest & Brown
13
Appendix
2001, Bai et al. 2004, Vasseur & Gaedke 2007). However, the importance of species
compensation might vary with the system studied. For example Houlahan et al. (2007), who
quantified the prevalence of negative covariance in a range of ecosystems, sustain that negative
covariances are rare in comparison to positive covariances. The authors suggest that community
stability is driven by mechanisms causing species to covary positively, such as similar
responses to environmental changes. In addition, parallel effects of bioengineering and/or
asynchronous species fluctuations can lead to other than compensatory patterns (Micheli et al.
1999). Parallel effects of bioengineering can occur in highly stressed habitats, such as high
intertidal zones, where the removal of a habitat-forming species will have a negative effect on
most of the species (Bertness & Callaway 1994). Accordingly, we would anticipate that higher
on the shore species compensation might have a limited effect on stability if canopies are lost.
In systems where species compensation is unimportant, ecological and statistical mechanisms
related to the number of species can influence the stability of communities (e.g. Tilman et al.
2006). Increasing community abundance due to resource complementarity and/or facilitation
(i.e. over-yielding), and decreasing summed variances due to decreasing abundances of
individual species (i.e. portfolio effect) can increase stability as diversity increases (Lehman &
Tilman 2000).
Effects of disturbances on community variability
The limited effect of mechanical disturbances on community variability contrasts with the
results from other experiments in aquatic and terrestrial ecosystems (e.g. Tilman 1996, Bertocci
et al. 2007, Brown 2007). We used in this study punctual destructive events within a 0.09 m2
14
Paper III
area, while others have used disturbances at larger spatial scales. For instance, Tilman (1996)
and Brown (2007) tested the effects of droughts on the variability of grassland and stream
macroinvertebrate communities, respectively; Bertocci et al. (2007) directly manipulated the
aerial exposure of entire intertidal communities. Probably, we disturbed the communities at a
scale too small to provoke a significant effect on recruitment via provision of empty space.
Small patches can quickly develop stable assemblages, because they are re-colonised by both
larval dispersal and lateral expansion of adults (Underwood & Chapman 2006). An alternative
explanation to the lack of disturbance effect is the strong propagule pressure at Helgoland
(Janke 1990). Removal of canopies allowed the massive settlement of opportunistic species
during spring and summer. Intense settlement might have overrode the potential effects of
disturbance-generated patches (Berlow 1997), resulting in a weak interaction between
disturbance and canopy. Moreover, experimental work in Helgoland has shown that even
highly frequent events of destruction have limited effects on the diversity of epibenthic
assemblages (Wollgast et al. 2008). The potential effect of mechanical disturbances on the
stability of this system still needs further attention.
Conclusion
Canopy removal increased the variability of species populations, but compensatory mechanisms
buffered the variability of community abundance. Although most research on community
variability is based on long-term time series (e.g. Ernest & Brown 2001, Bai et al. 2004, Hobbs
et al. 2007, Vasseur & Gaedke 2007), our short-term experiment showed communities
responding quickly to the manipulations of Fucus (after 3 months). This fast response is similar
15
Appendix
to those in studies involving the removal of canopies in other latitudes (e.g. Benedetti-Cecchi et
al. 2001, Lilley & Schiel 2006). On the other hand, the lack of response of the community-level
variability does not mean that Fucus serratus is unimportant in this system. Populations became
more variable due to the loss of canopies, which can increase the risk of extinctions and limit
the long-term persistence of the community (Pimm 1991). We suggest that compensatory
dynamics will have a critical role in maintaining the stability of systems where biological
habitat amelioration has opposing effects on other species.
Acknowledgements. We are grateful to a number of friends and colleagues who enthusiastically
helped during long hours of field work, including S. Domisch, J. Ellrich, A. Engel, M. Honens,
M. Marklewitz, A. Wagner, and H.Y. Yun. Comments by I. Bartsch, A. Perez-Matus and M.
Wahl greatly improved an early version of this manuscript. Part of the experimental data was
generated in the frame of the MarBEF responsive mode project BIOFUSE. Financial support by
the Alfred-Wegener-Institute for Marine and Polar Research to N.V. is acknowledged.
16
Paper III
LITERATURE CITED
Aguilera MA, Navarrete SA (2007) Effects of Chiton granosus (Frembly, 1827) and other
molluscan grazers on algal succession in wave exposed mid-intertidal rocky shores of
central Chile. J Exp Mar Biol Ecol 349:84-98
Airoldi L (2003) The effects of sedimentation on rocky coast assemblages. Oceanogr Mar Biol
41:161-236
Armsworth PR, Roughgarden JE (2003) The economic value of ecological stability. P Natl
Acad Sci USA 100:7147-7151
Austin MP, Cook BG (1974) Ecosystem stability: A result from an abstract simulation. Journal
of Theoretical Biology 45:435-458
Bai YF, Han XG, Wu JG, Chen ZZ, Li LH (2004) Ecosystem stability and compensatory
effects in the Inner Mongolia grassland. Nature 431:181-184
Bartsch I, Tittley I (2004) The rocky intertidal biotopes of Helgoland: present and past. Helg
Mar Res 58:289-302
Benedetti-Cecchi L, Pannacciulli F, Bulleri F, Moschella PS, Airoldi L, Relini G, Cinelli F
(2001) Predicting the consequences of anthropogenic disturbance: large-scale effects of
loss of canopy algae on rocky shores. Mar Ecol Prog Ser 214:137-150
Berlow EL (1997) From canalisation to contingency: historical effects in a successional rocky
intertidal community. Ecol Monogr 67:435-460
Bertness D, Leonard G, Levine J, Schmidt P, Ingraham A (1999) Testing the relative
contribution of positive and negative interactions in rocky intertidal communities.
Ecology 80:2711-2726
Bertness MD, Callaway R (1994) Positive interactions in communities. Trends Ecol Evol
9:191-193
Bertocci I, Vaselli S, Maggi E, Benedetti-Cecchi L (2007) Changes in temporal variance of
rocky shore organism abundances in response to manipulation of mean intensity and
temporal variability of aerial exposure. Mar Ecol Prog Ser 338:11-20
Brown BL (2007) Habitat heterogeneity and disturbance influence patterns of community
temporal variability in a small temperate stream. Hydrobiologia 586:93-106
Christensen NL, Bartuska AM, Brown JH, Carpenter S, D'Antonio C, Francis R, Franklin JF,
MacMahon JA, Noss RF, Parsons DJ, Peterson CH, Turner MG, Woodmansee RG
(1996) The report of the Ecological Society of America committee on the scientific
basis for ecosystem management. Ecol Appl 6:665-691
Connell JH (1961) The influence of interspecific competition and other factors on the
distribution of the barnacle Chathamalus stellatus. Ecology 42:710-723
Connell SD (2003) The monopolization of understorey habitat by subtidal encrusting coralline
algae: a test of the combined effects of canopy-mediated light and sedimentation. Mar
Biol 142:1065-1071
Connolly SR, Muko S (2003) Space preemption, size-dependent competition, and the
coexistence of clonal growth forms. Ecology 84:2979-2298
Costanza R, dArge R, deGroot R, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, ONeill
RV, Paruelo J, Raskin RG, Sutton P, van den Belt M (1997) The value of the world's
ecosystem services and natural capital. Nature 387:253-260
17
Appendix
Ernest SKM, Brown JH (2001) Homeostasis and compensation: the role of species and
resources in ecosystem stability. Ecology 82:2118-2132
Felton A, Wood J, Felton AM, Hennessey B, Lindenmayer DB (2008) Bird community
responses to reduced-impact logging in a certified forestry concession in lowland
Bolivia. Biol Conserv 141:545-555
Hobbs RJ, Yates S, Mooney HA (2007) Long-term data reveal complex dynamics in grassland
in relation to climate and disturbance. Ecol Monogr 77:545-568
Hooper DU, Chapin I, F. S., Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH, Lodge DM,
Loreau M, Naeem S, Schmid B, Setälä H, Symstad AJ, Vandermeer J, Wardle DA
(2005) Effects of biodiversity on ecosystem functioning: a consensus of current
knowledge. Ecol Monogr 75:3-35
Houlahan JE, Currie DJ, Cottenie K, Cumming GS, Ernest SKM, Findlay CS, Fuhlendorf SD,
Gaedke U, Legendre P, Magnuson JJ, McArdle BH, Muldavin EH, Noble D, Russell R,
Stevens RD, Willis TJ, Woiwod IP, Wondzell SM (2007) Compensatory dynamics are
rare in natural ecological communities. P Natl Acad Sci USA 104:3273-3277
Irving A, Connell S (2006) Predicting understorey structure from the presence and composition
of canopies: an assembly rule for marine algae. Oecologia 148:491-502
Ives AR, Gross K, Klug JL (1999) Stability and variability in competitive communities.
Science 286:542-544
Janke K (1986) Die Makrofauna und ihre Verteilung im Nordost-Felswatt von Helgoland. Helg
Mar Res 40:1-55
Janke K (1990) Biological interactions and their role in community structure in the rocky
intertidal of Helgoland (German Bight, North Sea). Helg Mar Res 44:219-263
Jenkins SR, Norton TA, Hawkins SJ (2004) Long term effects of Ascophyllum nodosum canopy
removal on mid shore community structure. J Mar Biol Assoc UK 84:327-329
Kennelly SJ, Underwood AJ (1993) Geographic consistencies of effects of experimental
physical disturbance on understorey species in sublittoral kelp forests in central New
South Wales. J Exp Mar Biol Ecol 168:35-58
Lehman CL, Tilman D (2000) Biodiversity, stability, and productivity in competitive
communities. Am Nat 156:534-552
Lilley S, Schiel D (2006) Community effects following the deletion of a habitat-forming alga
from rocky marine shores. Oecologia 148:672-681
MacArthur RH (1955) Fluctuations of animal populations and a measure of community
stability. Ecology 36:533-536
McCann KS (2000) The diversity–stability debate. Nature 405:228-233
Micheli F, Cottingham KL, Bascompte J, Bjornstad ON, Eckert GL, Fischer JM, Keitt TH,
Kendall BE, Klug JL, Rusak JA (1999) The dual nature of community variability. Oikos
85:161-169
Moore P, Hawkins SJ, Thompson RC (2007) Role of biological habitat amelioration in altering
the relative responses of congeneric species to climate change. Mar Ecol Prog Ser
334:11-19
Morrow K, Carpenter R (2008) Shallow kelp canopies mediate macroalgal composition: effects
on the distribution and abundance of Corynactis californica (Corallimorpharia). Mar
Ecol Prog Ser 361:119-127
Pimm SL (1991) The balance of nature?: ecological issues in the conservation of species and
communities. The University of Chicago Press, Chicago
18
Paper III
Ptacnik R, Solimini AG, Andersen T, Tamminen T, Brettum P, Lepisto L, Willen E,
Rekolainen S (2008) Diversity predicts stability and resource use efficiency in natural
phytoplankton communities. P Natl Acad Sci USA 105:5134-5138
R Development Core Team (2008) R: A language and environment for statistical computing, R
Foundation for Statistical Computing, Vienna
Reichert K, Buchholz F (2006) Changes in the macrozoobenthos of the intertidal zone at
Helgoland (German Bight, North Sea): a survey of 1984 repeated in 2002. Helg Mar
Res 60:213-223
Schluter D (1984) A variance test for detecting species associations, with some example
applications. Ecology 65:998-1005
Shea K, Roxburgh SH, Rauschert SJ (2004) Moving from pattern to process: coexistence
mechanisms under intermediate disturbance regimes. Ecol Lett 7:491-508
Steneck R (1986) The ecology of coralline algal crusts: convergent patterns and adaptative
strategies. Annu Rev Ecol Syst 17:273-303
Ter Braak CJF, Looman CWN (1994) Biplots in reduced-rank regression. Biom J 36:983-1003
Terlizzi A, Benedetti-Cecchi L, Bevilacqua S, Fraschetti S, Guidetti P, Anderson MJ (2005)
Multivariate and univariate asymmetrical analyses in environmental impact assessment:
a case study of Mediterranean subtidal sessile assemblages. Mar Ecol Prog Ser 289:2742
Tilman D (1996) Biodiversity: population versus ecosystem stability. Ecology 77:350–363
Tilman D (1999) The ecological consequences of changes in biodiversity: a search for general
principles. Ecology 80:1455–1474
Tilman D, Downing JA (1994) Biodiversity and stability in grasslands. Nature 367:363-365
Tilman D, Reich PB, Knops JMH (2006) Biodiversity and ecosystem stability in a decade-long
grassland experiment. Nature 441:629-632
Underwood AJ, Chapman MG (2006) Early development of subtidal macrofaunal assemblages:
relationships to period and timing of colonization. J Exp Mar Biol Ecol 330:221-233
Vasseur DA, Gaedke U (2007) Spectral analysis unmasks synchronous and compensatory
dynamics in plankton communities. Ecology 88:2058-2071
Winer BJ, Brown DR, Michels KM (1991) Statistical principles in experimental design.
McGraw-Hill, New York
Wollgast S, Lenz M, Wahl M, Molis M (2008) Effects of regular and irregular temporal
patterns of disturbance on biomass accrual and species composition of a subtidal hardbottom assemblage. Helg Mar Res DOI: 10.1007/s10152-008-0118-7
Worm B, Duffy JE (2003) Biodiversity, productivity and stability in real food webs. Trends
Ecol Evol 18:628-632
Yachi S, Loreau M (1999) Biodiversity and ecosystem productivity in a fluctuating
environment: The insurance hypothesis. P Natl Acad Sci USA 96:1463-1468
19
Appendix
Table 1. ANOVAs on the effects of Fucus serratus canopy (present or removed) and
mechanical disturbance (undisturbed or disturbed) on the variance of total community cover
and the sum of all species variances. Variances were calculated across 7 sample dates on plots
established at two intertidal sites. The F-ratio of each term was calculated using as denominator
the mean square given in the column MSd. Error terms (i.e., S × C, S × D, and S × C × D) that
were non-significant at D = 0.25 were pooled with the residual to increase the power of the
hypothesis tests (Winer et al. 1991)
Source
df
SS
MS
Variance in total community cover
Site, S
1
1.06
1.06
Canopy, C
1
0.05
0.05
Disturbance, D
1
0.80
0.80
S×C
1
0.02
0.02
S×D
1
0.43
0.43
C×D
1
0.16
0.16
S×C×D
1
0.25
0.25
Residual
32
11.13
0.35
Total
39
13.89
Pooled residual
35
11.82
0.34
Cochran’s test
C = 0.643, p = 0.21
Transformation
ln(x)
Summed species variances
Site, S
1
5664695 5664695
Canopy, C
1
4512241 4512241
Disturbance, D
1
69036
69036
S×C
1
209874
209874
S×D
1
3109735 3109735
C×D
1
441520
441520
S×C×D
1
19856
19856
Residual
32 41647729 1301492
Total
39 55674686
Pooled residual
34 41877458 1231690
Cochran’s test
C = 0.634, p = 0.24
Transformation
None
20
F
p
3.15
0.14
2.36
0.05
1.23
0.46
0.73
0.085
0.707
0.133
0.832
0.276
0.500
0.400
4.60
3.66
0.02
0.16
2.52
0.36
0.02
0.039
0.064
0.906
0.691
0.121
0.553
0.902
MSd
Pooled
Pooled
Pooled
Residual
Residual
Pooled
Residual
Pooled
Pooled
S×D
Residual
Residual
Pooled
Residual
Paper III
Table 2. ANOVAs on the effects of Fucus serratus canopy (present or removed) and
mechanical disturbance (undisturbed or disturbed) on the sum of all pair-wise species
covariances and the Bray-Curtis index. Variances were calculated across 7 sample dates on
plots established at two intertidal sites. The F-ratio of each term was calculated using as
denominator the mean square given in the column MSd. Error terms (i.e., S × C, S × D, and S ×
C × D) that were non-significant at D = 0.25 were pooled with the residual to increase the
power of the hypothesis tests (Winer et al. 1991)
Source
df
SS
MS
F
Summed covariance
Site, S
1
49280
49280
Canopy, C
1
1883167 1883167
Disturbance, D
1
366274
366274
S×C
1
15393
15393
S×D
1
254442
254442
C×D
1
30219
30219
S×C×D
1
147437
147437
Residual
32 11361283
355040
Total
39 14107495
Pooled residual
35 11778557
336530
Cochran’s test
C = 0.640, p = 0.22
Transformation
None
Bray-Curtis index
Site, S
Canopy, C
Disturbance, D
S×C
S×D
C×D
S×C×D
Residual
Total
Cochran’s test
Transformation
1
1
1
1
1
1
1
32
39
0.021
0.021
0.109
0.109
0.001
0.001
0.005
0.005
0.016
0.016
0.010
0.010
0.009
0.009
0.100
0.003
0.271
C = 0.218, p = 0.99
None
21
p
MSd
0.15
5.60
1.09
0.04
0.72
0.09
0.42
0.704
0.024
0.304
0.836
0.404
0.766
0.524
Pooled
Pooled
Pooled
Residual
Residual
Pooled
Residual
6.74
21.39
0.04
1.64
5.12
1.12
2.88
0.014
0.136
0.874
0.210
0.031
0.482
0.099
Residual
S×C
S×D
Residual
Residual
S×C×D
Residual
Appendix
FIGURE CAPTIONS
Fig. 1. Effects of canopy removal (Fucus serratus) and mechanical disturbance on (a) variance
of community total cover, (b) sum of all species variances, (c) sum of all pair-wise species
covariances, and (d) averaged Bray-Curtis dissimilarities. The measures of variability were
calculated across 7 sample dates on plot established at two intertidal sites (Nordostwatt or
Westwatt). Values are given as means ± SEM (n = 5).
Fig. 2. nMDS plot ordination plot on the basis of Euclidean distances among centroids of the
interaction between the factors canopy (c+ or c-, present or removed, respectively) and
disturbance (d- or d+, undisturbed or disturbed, respectively) with time sequence given as
numbers on plots (n = 5). Centroids were separately computed for each site using principal
coordinates from Bray-Curtis dissimilarities of untransformed data. Sampling months are 1:
March/06, 2: June/06, 3: September/06, 4: December/06, 5: March/07, 6: June/07, 7:
September/07.
Fig. 3. Redundancy analysis (RDA) bi-plots showing the correlation between functional types
(ECA: encrusting algae, EphA: ephemeral algae, TFA: turf-forming algae, SI: sessile
invertebrates, and MC: mobile consumers) in relation to treatments of canopy (c+ or c-, present
or removed, respectively), disturbance (d- or d+, undisturbed or disturbed, respectively), and
sampling date. The percent cover of sediment (Sed) is included as dependent variable. Time
sequence is given as numbers in the plots as follows. 1: March/06, 2: June/06, 3: September/06,
4: December/06, 5: March/07, 6: June/07, 7: September/07. Response variables are scaled to
zero mean and unit variance.
22
Paper III
Note: cumulative proportions of the variance accounted for by the two axes are as follow.
Nordostwatt: ECA = 0.82, EphA = 0.63, TFA = 0.65, SI = 0.28, MC = 0.54, and Sed = 0.79.
Westwatt: ECA = 0.82, EphA = 0.73, TFA = 0.74, SI = 0.24, MC = 0.43, and Sed = 0.76.
23
Appendix
Nordostwatt
a
Westwatt
Variance
2000
1000
0
Summed variance
b
4000
2000
0
Summed covariance
c
0
-500
-1000
-1500
Undisturbed
Bray-Curtis index
d
Disturbed
0.6
0.4
0.2
0.0
Present
Removed
Canopy
Present
Removed
Canopy
Fig. 1. Valdivia et al
24
Paper III
a. Nordostwatt
2
Stress: 0.07
b. Westwatt
2
Stress: 0.07
2
1
2
2
6
2
6
3
3
4
7
4
1
1
3
4
4
7 7
7
3 7
6
5
6
7
5 5
7
6
5
Fig. 2. Valdivia et al
25
4
5
5
6
5
1
1
2
3
4
6
7
2
4 3
4
5
3
3
1
1
6
1
c+ dc+ d+
c- dc- d+
Appendix
a. Nordostwatt
b. Westwatt
MC
MI
0.5
EphA
EA
MC
MI
CO
EA
EphA
0.5
2
2
d+
6
c-
4
0.0
1
4
Sed
3
c+
5
5
0.0
1 SI
d-
ECA
ECA
SI
3
c-
dd+
Sed
c+
7
6
7
-0.5
-0.5
TFA
TFA
-0.5
0.0
0.5
-0.5
Fig. 3. Valdivia et al.
26
0.0
0.5
Paper IV
Canning-Clode J1, 2 *, Valdivia N3, Molis M3, Thomason JC4, and Wahl M1 (2008) Estimation
of regional richness in marine benthic communities: quantifying the error. Limnology and
Oceanography: Methods. 6: 580-590
1
Leibniz Institute of Marine Sciences at the University of Kiel, Duesternbrooker Weg 20, D-
24105 Kiel, Germany
2
University of Madeira, Centre of Macaronesian Studies, Marine Biology Station of Funchal,
9000-107 Funchal, Madeira Island, Portugal
3
Biologische Anstalt Helgoland, Section Seaweed Biology, Foundation Alfred-Wegener-
Institute for Polar and Marine Research, marine station, Kurpromenade 201, D-27498
Helgoland
4
School of Biology, Newcastle University, Newcastle Upon Tyne, United Kingdom, NE1 7RU
*Corresponding author
E-mail: [email protected]
LIMNOLOGY
and
OCEANOGRAPHY: METHODS
Limnol. Oceanogr.: Methods 6, 2008, 580–590
© 2008, by the American Society of Limnology and Oceanography, Inc.
Estimation of regional richness in marine benthic communities:
quantifying the error
João Canning-Clode1,2*, Nelson Valdivia3, Markus Molis3, Jeremy C. Thomason4, and Martin Wahl1
1
Leibniz Institute of Marine Sciences at the University of Kiel, Duesternbrooker Weg 20, D-24105 Kiel, Germany
University of Madeira, Centre of Macaronesian Studies, Marine Biology Station of Funchal, 9000-107 Funchal, Madeira Island,
Portugal
3
Biologische Anstalt Helgoland, Section Seaweed Biology, Foundation Alfred-Wegener-Institute for Polar and Marine Research,
marine station, Kurpromenade 201, D-27498 Helgoland
4
School of Biology, Newcastle University, Newcastle Upon Tyne, United Kingdom, NE1 7RU
2
Abstract
Species richness is the most widely used measure of biodiversity. It is considered crucial for testing numerous ecological theories. While local species richness is easily determined by sampling, the quantification of regional richness relies on more or less complete species inventories, expert estimation, or mathematical extrapolation from a
number of replicated local samplings. However the accuracy of such extrapolations is rarely known. In this study,
we compare the common estimators MM (Michaelis-Menten), Chao1, Chao2, ACE (Abundance-based Coverage
Estimator), and the first and second order Jackknifes against the asymptote of the species accumulation curve, which
we use as an estimate of true regional richness. Subsequently, we quantified the role of sample size, i.e., number of
replicates, for precision, accuracy, and bias of the estimation. These replicates were sub-sets of three large data sets
of benthic assemblages from the NE Atlantic: (i) soft-bottom sediment communities in the Western Baltic (n = 70);
(ii) hard-bottom communities from emergent rock on the Island of Helgoland, North Sea (n = 52), and (iii) hardbottom assemblages grown on artificial substrata in Madeira Island, Portugal (n = 56). For all community types, Jack2
showed a better performance in terms of bias and accuracy while MM exhibited the highest precision. However, in
virtually all cases and across all sampling efforts, the estimators underestimated the regional species richness, regardless of habitat type, or selected estimator. Generally, the amount of underestimation decreased with sampling effort.
A logarithmic function was applied to quantify the bias caused by low replication using the best estimator, Jack2.
The bias was more obvious in the soft-bottom environment, followed by the natural hard-bottom and the artificial
hard-bottom habitats, respectively. If a weaker estimator in terms of performance is chosen for this quantification,
more replicates are required to obtain a reliable estimation of regional richness.
Introduction
1996) and is crucial for testing ecological models, such as the saturation of local communities colonized from regional species
pools (Cornell 1999). Ecological limitation (i.e., saturation) means
that with increasing number of available species in the regional
pool or with invasion events, local richness does not increase
beyond an intrinsically determined maximum (Srivastava 1999).
Thus, if only regional richness is driving local richness, a positive
linear relationship is often predicted (Cornell and Lawton 1992;
Srivastava 1999). Conversely, while concerns have been expressed
(Loreau 2000; Hillebrand and Blenckner 2002; Ricklefs 2004), it
has been widely accepted that if local assemblages are saturated
with species due to ecological interactions and niche overlap, an
asymptotic relationship is expected (Cornell and Lawton 1992;
Cornell and Karlson 1997; Srivastava 1999).
Several studies that seek to explain and/or test the relationship between local and regional diversity have assessed the
Species richness is the simplest and most commonly accepted
measure of biodiversity (Whittaker 1972; Magurran 1988; Gaston
*Corresponding author: E-mail: [email protected]
Acknowledgments
We appreciate the assistance of the late J. S. Gray and Robert Clarke in
the initial development of these ideas. We would like to thank Heye
Rumohr for providing the soft-bottom data and Mathieu Cusson for his
suggestions and critical review in this manuscript. We further thank Inka
Bartsch and Manfred Kaufmann for providing additional species inventories
for the natural and artificial hard-bottom habitats, respectively. The manuscript was significantly improved by the suggestions of three anonymous
referees. J. Canning-Clode was supported by a Fellowship from the
German Academic Exchange Service (DAAD) and J.C. Thomason by the
Royal Society.
580
Canning-Clode et al.
Estimating regional species richness
using arthropod abundances data and concluded that Chao1
and ACE (Abundance-based Coverage Estimator, Chao and
Lee 1992; Chazdon et al. 1998; Chao et al. 2000) have shown
the best performance among all estimators. For the marine
system, Foggo et al. (2003) performed an evaluation on the
performance of six estimators using simulations. They concluded that the estimator’s performance was affected by sampling effort, and no particular estimator performed best in all
cases. Nevertheless, Foggo et al. (2003) suggested Chao1 as the
most appropriate choice for a limited number of samples,
acknowledging that its performance may vary significantly in
cases of larger spatial scales and species richness. In these circumstances, the frequency of rare species could deteriorate the
performance of Chao1 (Foggo et al. 2003). This was later confirmed by Ugland and Gray (2004) in benthic assemblages of
the Norwegian continental shelf where Chao1 provided a
large underestimation of true richness.
Finally, the third category of assessing inventory completeness is through the extrapolation of SAC. In such curves, the
cumulative number of species is plotted against a cumulative
measure of sampling effort, e.g., the number of individuals
observed, samples or traps (Moreno and Halffter 2000; Gotelli
and Colwell 2001). The species richness can then be estimated
by fitting an equation to the curve and estimating its asymptote. While many functions have been proposed for this task
(see Tjørve 2003 for a review in possible model candidates),
the negative exponential function, the Clench equation, the
Weibull function, and the Morgan-Mercer-Flodin(MMF)
model have been frequently used (Soberon and Llorente 1993;
Colwell and Coddington 1994; Flather 1996; Lambshead and
Boucher 2003; Jimenez-Valverde et al. 2006; Mundo-Ocampo
et al. 2007). In theory, the asymptote’s location represents the
“true richness”, i.e., the total number of species that would be
observed with a hypothetical infinite sampling effort (Soberon
and Llorente 1993; Colwell and Coddington 1994; O’hara
2005; Jimenez-Valverde et al. 2006). The quality of the fitting
of the equation to the curve and, thus, the reliability of the
plateau should relate directly to the number of replicates.
The current study addresses the estimation of regional richness
using a novel approach. First, we extrapolate to the asymptote of the
SAC for three data sets, each with a large number of replicates and
from three different types of marine benthic communities. Second,
using the asymptote’s location as a reference for “true” regional richness, we then compare precision, bias, and accuracy of the regional
richness produced by six different estimators - Michaelis-Menten
(MM), ACE, Chao1, Chao2, Jack1, and Jack2. Finally, we quantify
the estimation error as a function of sampling effort.
regional species pool based on published species lists and by
consulting taxonomic experts (e.g., Hugueny and Paugy 1995;
Lawes et al. 2000; Witman et al. 2004; Harrison et al. 2006).
However, complete inventories of the fauna and flora of a
region are exceptionally hard to obtain and will probably
remain unavailable for most regions for the next few centuries
(Petersen and Meier 2003; Hortal et al. 2006). This problem is
more delicate in the marine environment where there is a large
phyletic diversity in certain groups and limited information
about others, e.g., Porifera (Foggo et al. 2003). Moreover, it is
difficult to appreciate to what degree such inventories are complete or incomplete (Soberon and Llorente 1993), and comparisons between published species lists are frequently unreliable
due to different sampling methods, terminology, or data handling systems (Dennis and Ruggiero 1996). In addition, when
saturation in certain assemblages is to be investigated, the
species capable to recruit into this habitat type (the relevant
richness) are only a subset of the entire regional richness.
To deal with these difficulties, a number of estimation techniques have been developed to extrapolate from the known to
the unknown, i.e., from a reasonable number of properly
inventoried samples to the true number of relevant species in
a certain area (Colwell and Coddington 1994). These techniques can be grouped into three classes: (i) parametric models, (ii) non-parametric estimators, and (iii) extrapolations of
SAC (species accumulation curves) (Magurran 2004). When
species fit a log normal distribution, i.e., a case of a parametric model to estimate species richness, it is possible to estimate
the theoretical number of species in the community by
extrapolating the shape of the curve. Most of the parametric
methods are, however, reported to perform improperly and
have not been used in recent years (Melo and Froehlich 2001).
In contrast, the non-parametric estimators have been suggested to perform better than SAC and parametric methods
(Baltanas 1992; Colwell and Coddington 1994; Walther and
Morand 1998; Walther and Martin 2001; Hortal et al. 2006).
These non-parametric estimators were originally developed to
estimate population size based on capture-recapture data and
adapted to extrapolate total species richness (Williams et al.
2002). With this technique, species richness is estimated from
the prevalence of rare species in each sample but does not
extrapolate beyond the last sample to an asymptote. In its
place, these models predict richness, including species not
found in the sample, from the proportional abundances of
species within the total sample (Soberon and Llorente 1993).
Several evaluations on the performance of different estimators
have been carried out (see review from Walther and Moore
2005). In most cases, the estimators Chao1 (Chao 1984),
Chao2 (Chao 1984, 1987; Colwell 2005), first order Jackknife
(Jack1 - Burnham and Overton 1979; Heltshe and Forrester
1983), and second order Jackknife (Jack2 - Smith and Van Belle
1984) perform better in terms of bias, precision, and accuracy
than other estimators (Walther and Moore 2005). In a recent
study, Hortal et al. (2006) compared 15 species richness estimators
Materials and procedures
For this study, we explored three sets of benthic communities: (i) soft-bottom: In Kiel Bay, Western Baltic, (54°38.3′ N,
10°39.6′ E) 70 replicates of macrofaunal samples were collected
to investigate the performance of species richness estimation
techniques. The 70 samples were collected from the same site
581
Canning-Clode et al.
Estimating regional species richness
species richness curve; c represents the maximum species richness. i.e., asymptote of the curve, as the number of replicates (x)
approaches infinity; b and d are model parameters that describe
the shape of the curve between the two extremes (Morgan et al.
1975). This model was previously used in two studies that performed a regional estimation of deep sea and littoral nematodes
(Lambshead and Boucher 2003; Mundo-Ocampo et al. 2007). In
those studies, estimates were obtained by extrapolation from a
SAC based on number of individuals, rather than number of
samples based on the UGE index as we do here.
Species richness estimations using variable replicate numbers—
We employed the frequently used software ‘EstimateS’ (version
7.5, Colwell 2005) to investigate the effect of sample size
(number of sampling units representing the replicates of ‘local
richness’) in estimating regional richness. This program computes sample-based rarefaction curves for a variety of species
richness estimators, presenting the mean number of random
sample re-orderings. Rarefaction and SAC were computed ten
times (using 10 randomly drawn sub-sets of replicates from
the entire data-set) for the replication levels 2, 4, 8, 16, 32, and
52 for each habitat. Because there were 70, 52, and 56 available replicates for the soft-bottom, natural hard-bottom, and
artificial hard-bottom data sets, respectively, there was a
higher chance of samples overlap when selecting the 32 and
52 samples sets. The rarefaction curve was based on 100 randomizations of the number of replicates sampled. We focused
our investigation on five non-parametric estimators as well as
on the asymptotic Michaelis-Menten (MM) richness estimator
(Raaijmakers 1987) (Table 1). These six estimators were previously used in several evaluations and were reported to perform
well (Walther and Moore 2005, see their table 3). Rosenzweig
et al. (2003) theoretically differentiated these two varieties of
estimators. Non-parametric estimators intend to overcome
sample-size insufficiencies and to report the number of species
present in sampled habitats. They operate only on the results
obtained from a subset of the total data set and do not represent an extrapolation. In contrast, MM extrapolates species
number to the asymptote of the SAC (Rosenzweig et al. 2003).
‘EstimateS’ calculates the MM estimator in two ways: (i) for
each of the 100 randomizations, which is then averaged
(MMRuns), or (ii) the mean accumulation curve is calculated
by averaging over 100 accumulation curves derived from 100
runs (MMMeans). We used the latter because of its less erratic
estimation (Colwell 2005; Walther and Moore 2005).
Estimator performance evaluation—Following Walther and
Moore (2005), we calculated three different quality indicators
that are commonly used to describe the performance of the
chosen estimators: bias, precision, and accuracy. Bias quantifies the mean difference between an estimated richness and
the true species richness. For measuring bias, we used the
scaled mean error:
in the early autumn of 1995 at the Station “Millionenviertel
14” using a 1000 cm2 van Veen grab at a depth of 24 m (covering a total of 7 m2 of sea bed). Samples were preserved in 4%
formaldehyde and later identified to species level (Rumohr
1999; Rumohr et al. 2001). (ii) In spring 2006, in Helgoland
Island, North Sea (54°11.4′ N, 07° 55.2′ E) one of us (NV) sampled sessile hard-bottom communities and identified them to
species level in 52 replicate quadrates of 400 cm2 in intertidal
rocky abrasion platforms. The study site “Nordostwatt” covers
approximately 450 m2 and is located in the northeast part of
the island and was extensively studied and inventoried by
Janke (1986). Janke (1986) described horizontal belts in the
intertidal as the Enteromorpha, Mytilus, Fucus serratus, and Laminaria zones. The data we use in this report are from 7 sub-habitats distributed in the F. serratus habitat. (iii) In early summer
2004, young hard-bottom communities were collected by
immersing 56 replicate polyvinylchloride (PVC) panels (225
cm2) for 5 mo at Madeira Island, Portugal, NE Atlantic (32°38.7′
N, 16° 53.2′ W). The panels were distributed in 12 PVC rings
(60 cm dia, 25 cm height) hung from a buoy at approximately
0.5 m depth. Minimum distance between rings was 5 m. The
original study focused on the influence of disturbance and
nutrient enrichment in hard-bottom assemblages (CanningClode et al. 2008). For the purpose of this analysis, only sessile
species on untreated control panels were taken into consideration. Hereafter, these datasets are referred to as soft-bottom,
natural hard-bottom, and artificial hard-bottom, respectively.
Predicting the asymptote of the SAC—Species accumulation
curves (SAC) were used (PRIMER 6, Clarke and Gorley 2006) to
calculate the total number of species observed (“Sobs Curve”) at
maximum sample size. Here, we used 52 replicates as maximum
sampling size for all habitats because this was the maximum
replicate number found in all habitat samples. Replicates were
permuted randomly 999 times. The analytical form of the mean
value of the accumulation curve over all permutations was
given by the UGE Index (Ugland et al. 2003). Ugland et al.
(2003) developed a total species curve (T-S curve) from SAC
obtained from single subareas. This curve can then be extrapolated to estimate the probable total number of species in a given
area (Ugland et al. 2003). They showed for the Norwegian continental shelf that the conventional SAC gave a large underestimation compared with the T-S curve. To estimate the asymptote
of the SAC (which we treat as ‘true’ regional richness in this
analysis) for all habitats, the nonlinear Morgan-Mercer-Flodin
(MMF) growth model (Morgan et al. 1975) was chosen. The
MMF model was selected by the curve fitting software CurveExpert (Hyams 2005) because of its superior fit regarding coefficient of correlation (r ) and standard error of the estimate (SE )
in all three data sets. The MMF model takes the form:
y = (ab + cxd) / (b + xd)
where y is species richness, and x represents the number of replicates. The parameters a, b, c, and d have the following interpretation: a is the calculated ordinate intercept of the replicates-
Bias = 1
∑
n
(Ej – A),
An j =1
where A is the asymptote of the SAC, Ej is the estimated
582
Canning-Clode et al.
Estimating regional species richness
Table 1. Summary of the species richness estimators used for this analysis
Richness
estimators
Type
Description
References
ACE
NP*
Abundance-based coverage estimator. It is a modification of the Chao & Lee (1992) estimators based on the ratio between rare (less than 10) and common species.
Chao and Lee 1992; Chazdon
et al. 1998; Chao et al. 2000
Chao1
NP*
Abundance-based estimator based on the number of rare species in a sample, i.e., represented by less than 3 individuals.
Chao 1984
Chao2
NP*
Incidence-based estimator. Takes into account the distribution of species amongst samples,
i.e., the number of species that occur in only 1 sample (‘rare species’) and the number of
species that occur in exactly 2 samples.
Chao 1984, 1987
Jack1
NP*
First-order Jackknife. Is based on the species occurring only in a single sample.
Burnham and Overton 1979;
Heltshe and Forrester 1983
Jack2
NP*
Second-order Jackknife. Is based on the species occurring in only 1 sample as well as in the
number that occur in exactly 2 samples.
Smith and Van Belle 1984
MMMean
P†
Michaelis-Menten Mean richness estimator. Computes the mean accumulation curve. Is calcuRaaijmakers 1987
lated by averaging over all accumulation curves derived from the selected runs.
*
NP, non-parametric estimator
P, parametric estimator
†
species richness for the jth replicate, and n is the number of
replicates. Positive and negative bias indicates overestimation
and underestimation, respectively.
Precision measures the variability of estimates among
repeated estimation runs for a given sample. For measuring precision, we used the complement of the coefficient of variation,
_
the latter being the ratio of deviation (SD) and mean ( E ):
when compared with the asymptote of the SAC of a given
habitat, x is the number of replicates, and a and b are model
parameters. Here too, the model was selected by the curve fitting software CurveExpert based on a high value of r and low
estimate SE.
Assessment
_
Predicting the location of the asymptote—In all three habitats,
species richness increased as a function of sampling effort
(Fig. 1). The total number of species observed in maximum
sample size, i.e., 52 replicates, was 55 species in the soft-bottom
habitat, 43 species for the natural hard-bottom assemblages,
and 32 species for the artificial hard-bottom habitat (Fig. 1).
The MMF model was chosen to extrapolate and predict the
location of the asymptote. This model described the data of
the SAC for the three habitats very well, with r ≈ 1 for all
curves (Table 2). Nevertheless, the model performed less well
for the natural hard-bottom assemblages as indicated by a
greater standard error of the estimate. The asymptote of
species richness (parameter c) was located at 103 species for
soft-bottom, 65 for natural hard-bottom, and 38 for the artificial hard-bottom habitat (Table 2).
Estimator’s performance—In general, Jack2 performed better
(with respect to bias and accuracy) at all replicate levels (low
sampling effort: < 8 replicates; intermediate sampling effort: 816 replicates; high sampling effort: > 16 replicates) in the three
habitats (Fig. 2). The estimator MM also had a satisfactory performance at low replication for all habitats, but with increasing sampling effort, its performance in terms of bias and accuracy improved less steeply as for the other estimators. In most
cases, at low and intermediate sampling effort, Chao1, Chao2,
and ACE performed worse. Bias decreased with rising sampling
Precision = 1 – (SD / E )
Accuracy measures the closeness of an estimated value to
the true richness (Brose and Martinez 2004). It is often measured using the mean squared error, combining bias, and precision (Hellmann and Fowler 1999). Here we applied the
scaled mean square error according to the formula:
Accuracy = 1 – (
n
1
(Ej – A)2),
2 ∑ j =1
An
where A is the asymptote of the SAC, Ej is the estimated species
richness for the j th sample, and n is the number of replicates.
Quantifying the relation between estimation error and number of
replicates—The relative estimation error of the six estimators
was expressed using the following formula:
y = (E/A) 100,
where y is the estimation error (in percent), E represents the
estimated species richness given by an elected estimator, and
A is the asymptote of the SAC in a given habitat. The estimation error was then plotted against the number of replicates
using a logarithmic model. This model takes the form:
y = a + b ln(x)
where y represents the underestimation of a given estimator
583
Canning-Clode et al.
Estimating regional species richness
to a large variability of species richness within replicates (Fig.
2H). In this habitat both of the Chao estimators showed weak
precision at low sampling effort. Precision was 1 at maximum
sampling effort for the natural hard-bottom habitat as there
was only one possible combination of the 52 replicates (Fig.
2H). MM showed high precision in almost all levels of replication and all community types. While the shapes of all curves
are comparable for the 3 community types, for a similar quality of estimation fewer replicates are required for the artificial
hard-bottom community than for the soft-bottom community.
In summary, Jack2 seems to be the most appropriate choice
at all levels of sampling effort for all habitats. MM constitutes
an alternative solution at low sampling effort for the natural
and artificial hard-bottom habitats.
For all community types and all estimators, the relative
estimation error and its error decreased with increasing replication (Fig. 3). It should be noted, however, that the decrease
in error, especially at replication levels 32 and 52, might be an
artifact caused by the statistically increased probability of resampling of the same replicates.
In the soft-bottom data-set, underestimation was never
lower than 35%, even at maximum sampling size (Fig. 3A). In
the natural hard-bottom communities, it was always larger
than 20% (Fig. 3B).The underestimation was lowest for the
artificial hard-bottom habitat (Fig. 3C). At low sampling effort,
which probably is the most common case in ecological studies, MM and Jack2 yield a substantially better estimation of
regional richness than the other four estimators for all assemblages. At maximum sampling size for the artificial hard-bottom habitat, average estimation error was below 20% for MM,
ACE, Jack1, Chao1, and Chao2, while Jack2 overestimated the
asymptote of the SAC (Fig. 3C).
To investigate in more detail the estimation error in all habitats, we have selected a logarithmic model and the Jack2 estimator due to its best overall performance. The logarithmic
model properly described the data for all habitats (Fig. 4; softbottom: r = 0.98, SE = 2.67; natural hard-bottom: r = 0.98, SE =
3.39; artificial hard-bottom: r = 0.96, SE = 5.51). The estimation
error decreases with increasing replication. Based on this model,
we quantified the bias caused by low replication for all habitats
(Table 3). With each doubling of replication number the estimation error by Jack2 decreases in average by 6.6% for the softbottom habitat, 8.4% for the natural hard-bottom habitat, and
8.5% for the artificial hard-bottom habitat (Fig. 4, Table 3).
Fig. 1. Species accumulation curves (SAC) for the three community
types. These curves were plotted using the UGE index calculated in
PRIMER 6.0
effort and was consistently negative (i.e., underestimation) for
all estimators in the soft-bottom and natural hard-bottom
habitats (Fig. 2A-B). In the artificial hard-bottom habitat, too,
all six estimators underestimated the asymptote of the SAC
with the single exception that at replicate level 52, Jack2 produced the only overestimation ever observed (Fig. 2C). Generally, the underestimation was more pronounced for the softbottom communities.
Accuracy improved steadily with increasing replication,
with a similar slope in all community types, but generally more
smooth in the soft-bottom communities (Fig. 2D-F). Jack2 was
the most accurate estimator in all habitats when replication
exceeded 2. At low and intermediate sampling effort, MM was
as accurate as Jack2 for the natural and artificial hard-bottom
habitats (Fig. 2E – F). In contrast, MM was the least accurate
estimator for the soft-bottom community (Fig. 2D) and at high
sampling effort for the other two habitats.
Precision of the estimation increased rapidly in the first 10
replicates and more slowly after that (Fig. 2G-I). This pattern
was similar in all communities, probably because it is a statistical property (i.e., it approximates to the standard deviation).
Nevertheless, in the natural hard-bottom assemblages, Jack2
showed a high imprecision at intermediate sampling effort due
Table 2. Coefficients of correlation (r ), standard error of the estimate (SE ), and parameter values of the MMF model used for the
extrapolation of the asymptote of the SAC for all habitats
Parameters
Habitat
Soft-bottom
Hard-bottom
Artificial hard-bottom
a
–8.12
–63.05
–3.62
b
c
d
r
SE
3.41
0.63
1.60
102.67
65.48
37.90
0.38
0.27
0.58
0.999
0.999
0.999
0.074
0.205
0.023
584
Canning-Clode et al.
Estimating regional species richness
Fig. 2. Bias (panels A-C), accuracy (D-F), and precision (G-I) of the selected estimators (MM, Chao1, ACE, Chao2, Jack1, and Jack2) for the three habitats using variable replicate numbers. *In panel H, precision was 1 at replicate level 52 as there was only one set of 52 replicates.
marine realm. In the present study, we have evaluated the potential and limitations of such an approach. For this purpose, we
selected three data sets with a large number of replicates from different temperate shallow water habitats. We compared the performance of six different estimators of regional richness against the
asymptote of the species accumulation curve (SAC) using randomly selected replicates for a range of sample sizes.
The most conspicuous outcome of this analysis is that as a
general rule the estimation of regional species richness based
on local assemblages underestimates the asymptote of the
SAC, regardless of habitat type, number of replicates, or
selected estimator. The only exception was when a single estimator, Jack2, using 52 replicates overestimated the asymptote
of the SAC in the artificial hard-bottom habitat. For all estimators, the amount of underestimation gradually decreased
with increasing sample size.
Overall, we have demonstrated that Jack2 performed best
in all habitats. Using the logarithmic model, we predict that
one would need 1865 samples to reach the asymptote of the
SAC in the soft-bottom habitat (Table 3). For the natural and
artificial hard-bottom habitats, a considerably less sampling
effort would be required to reach the asymptote of the SAC.
Discussion
Studies that are searching for a clear understanding of the local
versus regional diversity pattern in the marine environment have
often defined the number of species in a region by questioning
experts or consulting available species inventories (e.g., Witman et
al. 2004; Harrison et al. 2006). In many poorly studied areas, however, true regional species numbers are unknown. Therefore the
statistical estimation of regional richness, based on a limited
number of replicates, constitutes an important alternative for the
585
Canning-Clode et al.
Estimating regional species richness
The estimation error was greatest in the soft-bottom environment, followed by the natural hard-bottom and the artificial hard-bottom habitats. Nevertheless, the similarity of the
estimation error patterns between the three data sets is surprising in view of the (intentional) differences between the
selected samples regarding community type, community age,
diversity and method of sampling. For instance, the size of a
single sampling unit was 1000, 400, and 225 cm2, in the softbottom, natural, and artificial hard-bottom samples, respectively. Thus, at comparable species density, a single replicate
for the soft-bottom habitat possibly contained a larger proportion of the regional species pool than in the other samples.
Also, the suspended PVC panels used for the Madeira data-set
can be considered island communities on patchy substrata,
with diversity possibly constrained by habitat (panel) size,
whereas the samples from the other two data sets were subareas from much larger contiguous habitats. Sample unit size
and patchiness of habitat may affect the similarity between
replicates and, thus, the initial slope of the curve. Moreover,
the slow accumulation and consequently, the larger number
of replicates required to reach the plateau in the soft bottom
sample may be linked to the number of rare species present, as
well as to the sensitivity of the sampling method.
Despite the extensive differences between the samples chosen with regard to size of sampling area, patchiness of habitat
or age of community, the performance of the estimators
applied to the described data sets was comparable. This may be
indicative of a remarkable robustness of the observed pattern.
The fact that the six estimators underestimated the asymptotic
species richness is consistent with other studies that use the
same and other estimators (e.g., Petersen and Meier 2003;
Brose and Martinez 2004; Cao et al. 2004). Beyond the general
similarity among estimator’s performances, Jack2 was more
accurate and less biased for all habitats in almost all replication levels. In contrast, MM exhibited a high precision in all
habitats. At low sampling effort, MM and Jack2 performed best
in terms of bias, accuracy, and precision for the natural and
artificial hard-bottom communities. For the soft-bottom community, Jack2 was clearly the least biased and the most accurate estimator at all levels of replication. For estimations based
on larger samples, both Chao1 and ACE seem to perform
slightly better than MM but worse than Chao2 and both Jackknifes. These findings are comparable to some previously
reported results. For instance, the study by Walther and Moore
(2005) found that Chao2 (Chao 1987) performed best while
Jack2, Jack1, and Chao1 ranked second, third, and fourth,
respectively. The MMMean and ACE estimators were reported
to perform worse (Walther and Moore 2005). Although they
did not evaluate the performances of ACE, MM, and Jack2,
Foggo et al. (2003) concluded that amongst 6 different techniques to estimate marine benthos species richness, Chao1
represented the best nonparametric alternative for a limited
number of survey units. In contrast, Ugland and Gray (2004)
argue that Chao1 severely underestimates the true richness in
Fig. 3. Percentages of asymptotic species richness estimated by MM,
Chao1, ACE, Chao2, Jack1, and Jack2 using variable replicate numbers for
the (A) soft-bottom, (B) natural hard-bottom, and (C) artificial hard-bottom habitats. Means and 95% confidence intervals are indicated (n = 10).
Dashed line indicates the asymptote of the SAC given by the UGE index.
*In panel B at replicate level 52 confidence intervals are zero as there was
only one set of 52 replicates.
586
Canning-Clode et al.
Estimating regional species richness
Table 3. Quantification of the estimation error by Jack2*
Number of replicates
Habitat
Soft-bottom
Natural hard-bottom
Artificial hard-bottom
2
4
8
16
32
52
y(0)
71.37
62.74
45.36
64.14
54.02
36.07
56.9
45.31
26.71
49.67
36.60
17.39
42.43
27.88
8.07
37.37
21.78
1.54
1865.43
294.13
58.31
*
Based on the logarithmic model, we calculated the approximate estimation error by Jack2 (%) to compensate the bias caused by low replication. With
the same model we also calculated for each habitat, the required sampling effort for the Jack2 estimator to be unbiased (y (0)).
may result. Only one of the previously mentioned studies has
estimated true richness based on the asymptote of the species
accumulation curve (Foggo et al. 2003).
In this study, we estimated true regional richness by extrapolation of the SAC given by the UGE index using the non-linear Morgan-Mercer-Flodin (MMF) growth model (Morgan et al.
1975). The MMF model was previously employed in two surveys on the diversity of deep sea and littoral nematodes (Lambshead and Boucher 2003; Mundo-Ocampo et al. 2007). Lambshead and Boucher (2003) estimated the marine nematode
species richness in 16 locations. They have compared the estimations given by the MMF model with the non-parametric
incidence-based coverage estimator (ICE - Lee and Chao 1994;
Chazdon et al. 1998; Chao et al. 2000). In 88% of cases, the
estimation given by the extrapolation was higher than the estimation provided by ICE. In one instance, both methods provided identical estimates of nematodes species, in another one
ICE produced higher numbers (Lambshead and Boucher 2003).
Mundo-Ocampo et al. (2007) used the same approach to compare nematode biodiversity in two shallow, littoral locations of
the Gulf of California. In both locations, the MMF extrapolation gave a higher estimation of nematode richness than ICE
(Mundo-Ocampo et al. 2007). Both studies did not attempt to
quantify the relationship between estimation error and low
replication, as we do here. In opposition to these investigations
where SAC were plotted as a function of the accumulated number of individuals, our study uses SAC plotted against the accumulated number of samples. Deciding which type of curves to
use depends on the nature of the data available (Gotelli and
Colwell 2001). If sample-based data are available, a SAC based
on samples is preferable, as it can account for natural levels of
sample patchiness (i.e., heterogeneity between replicates) in
the data (Gotelli and Colwell 2001). A further distinction of the
present study from the investigations by Lambshead and
Boucher (2003) and Mundo-Ocampo et al. (2007) is the use of
the T-S curve (given by the UGE index) developed by Ugland et
al. (2003) followed by the MMF model fitting to it. The resulting extrapolation of the asymptotic richness is a more realistic
estimation than the usual SAC (Ugland et al. 2003).
We demonstrate that the minimum sampling effort to
reach a realistic estimation of true regional richness is variable
among communities or sampling methodology. Below this
threshold sampling effort estimation is negatively biased. The
unavoidable estimation error caused by low replication can,
benthic assemblages of the Norwegian continental shelf. In
their study, Chao1 predicted approximately 1100 species from
a data-set with 809 species. Nevertheless, when surveying
larger areas of the shelf than the ones they use in their analysis (see their Table 1), over 2500 species were found (Ugland
and Gray 2004). The large underestimation by Chao1 is
caused by infrequent species (Ugland and Gray 2004). In a
recent evaluation of 15 different estimators using arthropods
abundances, Hortal et al. (2006) concluded that the performance of 10 estimators were highly dependent on the level of
replication. In that study, Chao1 and ACE showed a higher
precision at low replication but the superiority of these two
estimators over the rest decreases with increasing sample size.
Conversely, in a study using Monte Carlo simulations, Brose
and Martinez (2004) showed that ACE, Chao1, Chao2, and
Jack2 were positively biased under high replication. However,
in some of the previously mentioned studies, true richness was
estimated based on inventories, experts, simulated landscape
models, or museum collection data (Brose et al. 2003; Petersen
and Meier 2003; Brose and Martinez 2004; Cao et al. 2004;
Hortal et al. 2006) and not on real and numerous community
sub-units, as done in this study. If incomplete lists suggest a
lower-than-real regional richness, apparent overestimations
Fig. 4. Estimation error by Jack2 using variable replicate numbers for the
soft-bottom, natural hard-bottom, and artificial hard-bottom habitats
using the logarithmic model y = a + b ln(x). Means and 95% confidence
intervals are indicated (n = 10).
587
Canning-Clode et al.
Estimating regional species richness
Burnham, K. P., and W. S. Overton. 1979. Robust estimation of
population-size when capture probabilities vary among animals. Ecology 60:927-936.
Canning-Clode, J., M. Kaufmann, M. Wahl, M. Molis, and M.
Lenz. 2008. Influence of disturbance and nutrient enrichment on early successional fouling communities in an oligotrophic marine system. Mar. Ecol. Evol. Persp. 29:115-124.
Cao, Y., D. P. Larsen, and D. White. 2004. Estimating regional
species richness using a limited number of survey units.
Ecoscience 11:23-35.
Chao, A. 1984. Non-parametric estimation of the number of
classes in a population. Scand. J. Stat. 11:265-270.
———. 1987. Estimating the population size for capture-recapture data with unequal catchability. Biometrics. 43:783-791.
———, W. H. Hwang, Y. C. Chen, and C. Y. Kuo. 2000. Estimating the number of shared species in two communities.
Stat. Sinica 10:227-246.
——— and S. M. Lee. 1992. Estimating the number of classes
via sample coverage. J. Am. Stat. Assoc. 87:210-217.
Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R. Guariguata. 1998. Statistical methods for estimating species richness of woody regeneration in primary and secondary rain
forests of Northeastern Costa Rica, p. 285-309. In F.
Dallmeier and J. A. Comiskey [eds.], Forest biodiversity
research, monitoring and modeling: conceptual background and old world case studies. Parthenon Publishing.
Clarke, K. R., and R. N. Gorley. 2006. PRIMER v6. User manual/tutorial. Plymouth routine in mulitvariate ecological
research. Plymouth Marine Laboratory.
Colwell, R. K. 2005. EstimateS: Statistical estimation of species
richness and shared species from samples. Version 7.5.
User’s guide and application published at: http://purl.oclc.org/
estimates. (accessed October 2007)
——— and J. A. Coddington. 1994. Estimating terrestrial biodiversity through extrapolation. Philos. T. Roy. Soc. B. 345:101-118.
Cornell, H. V. 1999. Unsaturation and regional influences on
species richness in ecological communities: A review of the
evidence. Ecoscience 6:303-315.
——— and R. H. Karlson. 1997. Local and regional processes
as controls of species richness, p. 250-268. In D. Tilman and
P. Kareiva [eds.], Spatial ecology. Princeton Univ. Press.
——— and J. H. Lawton. 1992. Species interactions, local and
regional processes, and limits to the richness of ecological
communities: a theoretical perspective. J. Anim. Ecol. 61:1-12.
Dennis, J. G., and M. A. Ruggiero. 1996. Biodiversity inventory:
building an inventory at scales from local to global, p. 149156. In R. C. Szaro and D. W. Johnston [eds.], Biodiversity in
managed landscapes. Oxford Univ. Press.
Flather, C. H. 1996. Fitting species-accumulation functions
and assessing regional land use impacts on avian diversity.
J. Biogeogr. 23:155-168.
Foggo, A., M. J. Attrill, M. T. Frost, and A. A. Rowden. 2003.
Estimating marine species richness: an evaluation of six
extrapolative techniques. Mar. Ecol. Prog. Ser. 248:15-26.
however, be quantified as shown in this paper. In addition, the
logarithmic function used to quantify the estimation error
also informs at which sampling effort the estimation
approaches the asymptote of the SAC. Consequently, we predict that if a weaker estimator in terms of performance is chosen, the logarithmic function will approach the x-axis far later,
i.e., a greater amount of replicates would be required to equal
the asymptote of the SAC.
The extrapolation of the SAC is a computation and the
shape of the curve is affected by the presence/absence of rare
species in the samples as well as the amount of samples used
in the model. To assess how well the plateau reflects the “real”
regional richness, we compared the plateau values obtained by
our approach to the numbers provided by existing comprehensive inventories in the three systems. (i) A 30-year-long
survey of the soft-bottom macrofauna in the Kiel Bay, Western
Baltic Sea lists 184 species at the Station “Millionenviertel 14”
(Rumohr, pers com). (ii) On Helgoland island, three extensive
studies in the same sub-habitats we used here, reported 53 sessile animal species (Janke 1986; Reichert 2003) and 39 species
of macroalgae (Inka Bartsch, unpublished data). (iii) Finally,
studies on the diversity of hard-bottom communities growing
on artificial substrata conducted during three consecutive
years in the south coast of Madeira Island (Jochimsen 2007;
Canning-Clode et al. 2008, Manfred Kaufmann, unpublished
data) reported a total of 44 species growing on the same type
of substrata, depth, and study site as the artificial hard-bottom
data-set in this analysis. Compared to these values, our extrapolation still underestimates the “real” richness of the investigated habitats by 44% for the soft-bottom, 29% for the natural hard-bottom, and 14% for the artificial hard-bottom
habitats. However, it should be noted that the reference lists
include species from several seasons, years, and successional
stages, which, in contrast to our data set, do not necessarily
co-exist at the local scale. Regional species pools based on such
inventories may include species not susceptible to recruit into
the community considered because they are restricted to different habitats and seasons.
We conclude that regional richness can be estimated from
sub-samples, that the quality of the estimation increases with
sample size, and that the magnitude of the unavoidable estimation error can be quantified and, thus, corrected to some
extent. We encourage further studies to include data from
more locations and then provide more robust correction values to compensate the bias caused by low replication.
References
Baltanas, A. 1992. On the use of some methods for the estimation of species richness. Oikos 65:484-492.
Brose, U., and N. D. Martinez. 2004. Estimating the richness of
species with variable mobility. Oikos 105:292-300.
———, ———, and R. J. Williams. 2003. Estimating species
richness: Sensitivity to sample coverage and insensitivity to
spatial patterns. Ecology 84:2364-2377.
588
Canning-Clode et al.
Estimating regional species richness
individual stones in tropical streams. Freshwater Biol.
46:711-721.
Moreno, C. E., and G. Halffter. 2000. Assessing the completeness of bat biodiversity inventories using species accumulation curves. J. Appl. Ecol. 37:149-158.
Morgan, P. H., L. P. Mercer, and N. W. Flodin. 1975. General
model for nutritional responses of higher organisms. Proc.
Natl. Acad. Sci. USA 72:4327-4331.
Mundo-Ocampo, M., and others. 2007. Biodiversity of littoral
nematodes from two sites in the Gulf of California. Hydrobiologia 586:179-189.
O’hara, R. B. 2005. Species richness estimators: how many
species can dance on the head of a pin? J. Anim. Ecol.
74:375-386.
Petersen, F. T., and R. Meier. 2003. Testing species-richness estimation methods on single-sample collection data using the
Danish Diptera. Biodivers. Conserv. 12:667-686.
Raaijmakers, J. G. W. 1987. Statistical-analysis of the
Michaelis-Menten equation. Biometrics. 43:793-803.
Reichert, K. 2003. Die Makrofauna und ihre räumliche
Verteilung und saisonale Veränderung im Felswatt von Helgoland - Ein Vergleich zur Untersuchung 1984. Universität
Hamburg.
Ricklefs, R. E. 2004. A comprehensive framework for global
patterns in biodiversity. Ecol. Lett. 7:1-15.
Rosenzweig, M. L., W. R. Turner, J. G. Cox, and T. H. Ricketts.
2003. Estimating diversity in unsampled habitats of a biogeographical province. Conserv. Biol. 17:864-874.
Rumohr, H. 1999. Soft bottom macrofauna: collection, treatment and quality assurance of samples. ICES Tech. Mar.
Environ. Sci. 27:1-19.
———, I. Karakassis, and J. N. Jensen. 2001. Estimating species
richness, abundance and diversity with 70 macrobenthic
replicates in the Western Baltic Sea. Mar. Ecol. Prog. Ser.
214:103-110.
Smith, E. P., and G. Van Belle. 1984. Nonparametric-estimation of species richness. Biometrics. 40:119-129.
Soberon, J., and J. Llorente. 1993. The use of species accumulation functions for the prediction of species richness. Conserv. Biol. 7:480-488.
Srivastava, D. S. 1999. Using local-regional richness plots to
test for species saturation: Pitfalls and potential. J. Anim.
Ecol. 68:1-16.
Tjørve, E. 2003. Shapes and functions of species-area curves: a
review of possible models. J. Biogeogr. 30:827-835.
Ugland, K. I., and J. S. Gray. 2004. Estimation of species richness: analysis of the methods developed by Chao and
Karakassis. Mar. Ecol. Progr. Ser. 284:1-8.
———, ———, and K. E. Ellingsen. 2003. The species-accumulation curve and estimation of species richness. J. Anim.
Ecol. 72(5):888-897.
Walther, B. A., and J. L. Martin. 2001. Species richness estimation of bird communities: how to control for sampling
effort? Ibis 143:413-419.
Gaston, K. J. 1996. Species richness: Measures and measurement. In K. J. Gaston [ed.], Biodiversity: a biology of
number and differences. Blackwell Sciences.
Gotelli, N. J., and R. K. Colwell. 2001. Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecol. Lett. 4:379-391.
Harrison, S., H. D. Safford, J. B. Grace, J. H. Viers, and K. F.
Davies. 2006. Regional and local species richness in an
insular environment: Serpentine plants in California. Ecol.
Monogr. 76:41-56.
Hellmann, J. J., and G. W. Fowler. 1999. Bias, precision, and
accuracy of four measures of species richness. Ecol. Appl.
9:824-834.
Heltshe, J. F., and N. E. Forrester. 1983. Estimating species richness using the jackknife procedure. Biometrics. 39:1-11.
Hillebrand, H., and T. Blenckner. 2002. Regional and local
impact on species diversity - from pattern to processes.
Oecologia 132:479-491.
Hortal, J., P. A. V. Borges, and C. Gaspar. 2006. Evaluating the
performance of species richness estimators: sensitivity to
sample grain size. J. Anim. Ecol. 75:274-287.
Hugueny, B., and D. Paugy. 1995. Unsaturated fish communities in African rivers. Am. Nat. 146:162-169.
Hyams, D. G. 2005. CurveExpert 1.3: A comprehensive curve
fitting system for Windows. http://curveexpert.webhop.biz/
(accessed October 2007)
Janke, K. 1986. Die Makrofauna und ihre Verteilung im Nordost-Felswatt von Helgoland. Helgol Meeresunters 40:1–55.
Jimenez-Valverde, A., S. J. Mendoza, J. M. Cano, and M. L.
Munguira. 2006. Comparing relative model fit of several
species-accumulation functions to local Papilionoidea and
Hesperioidea butterfly inventories of Mediterranean habitats. Biodivers. Conserv. 15:177-190.
Jochimsen, M. C. 2007. Role of community structure for invasion dynamics. Master thesis. University of Osnabrück.
Lambshead, P. J. D., and G. Boucher. 2003. Marine nematode
deep-sea biodiversity - hyperdiverse or hype? J. Biogeogr.
30:475-485.
Lawes, M. J., H. A. C. Eeley, and S. E. Piper. 2000. The relationship between local and regional diversity of indigenous
forest fauna in KwaZulu-Natal Province, South Africa. Biodivers. Conserv. 9:683-705.
Lee, S. M., and A. Chao. 1994. Estimating population-size via
sample coverage for closed capture-recapture models. Biometrics 50:88-97.
Loreau, M. 2000. Are communities saturated? On the relationship between alpha, beta and gamma diversity. Ecol. Lett.
3:73-76.
Magurran, A. 1988. Ecological diversity and its measurement.
Princeton Univ. Press.
———. 2004. Measuring biological diversity. Blackwell Publishing.
Melo, A. S., and C. G. Froehlich. 2001. Evaluation of methods
for estimating macroinvertebrate species richness using
589
Canning-Clode et al.
Estimating regional species richness
sis and management of animal populations: modeling, estimation, and decision making. Academic Press.
Witman, J. D., R. J. Etter, and F. Smith. 2004. The relationship
between regional and local species diversity in marine benthic communities: A global perspective. Proc. Natl. Acad.
Sci. USA 111(44):15664-15669.
——— and J. L. Moore. 2005. The concepts of bias, precision
and accuracy, and their use in testing the performance of
species richness estimators, with a literature review of estimator performance. Ecography 28:815-829.
——— and S. Morand. 1998. Comparative performance of species
richness estimation methods. Parasitology 116:395-405.
Whittaker, R. H. 1972. Evolution and measurement of species
diversity. Taxon 21:213-251.
Williams, B. K., J. D. Nichols, and M. J. Conroy. 2002. Analy-
Submitted 17 June 2008
Revised 6 October 2008
Accepted 22 October 2008
590
ANLAGE ZUR DISSERTATION
ANLAGE ZUR DISSERTATION
Name:
Nelson Valdivia
Anschrift:
An der Sapskuhle 516
27498 Helgoland
Helgoland, 11.12.08
Erklärung
Gem.§6 (5) Nr.1-3 PromO
Ich erkläre, dass ich
NELSON VALDIVIA
1. die Arbeit ohne unerlaubte fremde Hilfe angefertigt habe,
2. keine anderen, als die von mir angegebenen Quellen und Hilfsmittel benutzt
habe
3. die den benutzten Werken wörtlich oder inhaltlich entnommenen Stellen als
solche kenntlich gemacht habe.
____________________________________
(Unterschrift)