Download Thesis in pdf

Thesis for the degree of Doctor of Philosophy
Ecological disturbances:
The Good, the Bad and the Ugly
J. Robin Svensson
2010
Department of Marine Ecology - Tjärnö
University of Gothenburg
45296 Strömstad
SWEDEN
© J. Robin Svensson 2010
All rights reserved. No part of this publication may be reproduced or
transmitted, in any form or by any means, without written permission.
ISBN 978-91-628-8200-6
Printed by Geson Hylte Tryck, Göteborg, Sweden 2010
Ecological disturbances
Till Ivar och Kerstin Karlsson,
och Nils-Erik Hjertquist
2
J. Robin Svensson
Svensson, J. Robin 2010.
ECOLOGICAL DISTURBANCES: THE GOOD, THE BAD AND THE UGLY.
Abstract. This thesis focuses on the definitions, characterizations and quantifications of
ecological disturbances, as well as hypotheses on their impacts on biological communities.
The most prominent model on effects of disturbance on diversity is the Intermediate
Disturbance Hypothesis (IDH), which is utilized in management of national reserves, has
received over 3300 citations and has been corroborated by a multitude of studies from
terrestrial and aquatic systems. According to the predictions of the IDH, diversity is high at
intermediate levels of disturbance due to coexistence of competitors and colonizers. At low
levels of disturbance diversity will be low due to competitive exclusion and few species can
persist at high levels of disturbance. In an extension of the IDH, the Dynamic Equilibrium
Model (DEM) predicts that the effects of disturbance depend on the productivity of
communities, because at high growth rates a stronger disturbance is required to counteract
increased rates of competitive exclusion. The IDH and the DEM were tested in a field
experiment on effects of physical disturbance (scraping) and productivity (nutrient
availability) on hard-substratum assemblages in paper I, where the patterns predicted by the
IDH, but not the DEM, were observed. This outcome shows the importance of the nature of
productivity alterations, as the productivity treatment had a general positive effect on growth
rates but only marginal effects on the dominant species, thereby leaving rates of competitive
exclusion unaffected.
In paper II I tested another extension of the IDH, which predicts that smaller,
more frequent disturbances will have different effects on diversity compared to larger, less
frequent disturbances. In this experiment I used two different regimes of disturbance, small
and frequent vs. large and infrequent disturbances, while the overall rate (the product of area
and frequency) was kept equal for both regimes. At the site where the IDH was supported, the
regime with a large proportion of the area disturbed infrequently showed higher richness, due
to a stronger decrease of dominants, compared to the regime with a small proportion disturbed
frequently. In addition to these significant differences in diversity effects between different
disturbance regimes, it may also matter what agent of disturbance that is causing the damage.
In paper III I contrasted the effects of a physical disturbance (wave-action) to that of a
biological disturbance (grazing), as well as their respective interactions with productivity in a
multifactorial design tested on natural epilithic assemblages. The composition of assemblages
and the total species richness was significantly affected by physical disturbance and
interactively by biological disturbance and productivity. The algal richness was significantly
affected by productivity and biological disturbance, whereas the invertebrate richness was
affected by physical disturbance. The results show, for the first time, that biological
disturbance and physical disturbance interact differently with productivity due to differences
in the distribution and selectivity among disturbances.
In paper IV I investigate how the choice of diversity measure may impact the
outcomes of tests of the IDH, which, surprisingly, has not previously been discussed. This
was done by an extensive literature review and meta-analysis on published papers as well as
by two different approaches to mathematical modelling. Both models support the IDH when
biodiversity is measured as species richness, but not evenness. The meta-analysis showed that
two-thirds of the published studies in the survey present different results for different
diversity measures. Hence, the choice of diversity measure is vital for the outcome of tests of
the IDH and related models.
Key words: competitive exclusion; DEM; disturbance; diversity; evenness; IDH; marine
assemblages; productivity; rate of disturbance; regime; species richness; Tjärnö, Sweden.
3
Ecological disturbances
ISBN: 978-91-628-8200-6
Populärvetenskaplig sammanfattning
Som den skamlöst fyndiga titeln syftar till så kan ekologiska störningar se väldigt olika ut och
ha helt olika effekter på den biologiska mångfalden. Men innan vi ger oss i kast med en
djupare tolkning av detta, bör vi bena ut vad en störning egentligen är. Exempel på vanliga
störningar i naturen är skogsbränder, stormar, översvämningar, vågor, trålning, föroreningar,
uttorkning samt istäcken och drivved som skrapar bort arter på hårda bottnar. Lite ibland
räknas även biologiska störningar, d.v.s. djur som tuggar i sig andra djur och växter, eller djur
som i ren illvilja eller okunskap trampar ihjäl levande varelser i sin väg. För att krångla till
detta en smula så får inte allting som kan ge upphov till skada kallas för en störning, utan i
likhet med samhället i stort finns även här vissa som är mer jämlika än andra. Definitioner på
vad som får räknas som en faktisk störning finns det lika många som antalet GAIS supportrar;
ungefär nio. Enligt den mest konkreta och lätthanterliga definitionen ska en störning döda
eller avlägsna organismer i ett samhälle (område med samexisterande arter), och därigenom
underlätta för nya arter att etablera sig. Den till synes harmlösa bisatsen om
etableringsmöjligheter får oanat stor betydelse när man testar ekologiska förklaringsmodeller
om störning och biodiversitet.
Överlag sunda läsare undrar nu förmodligen vad i hela Hisingen en ekologisk
förklaringsmodell är. Dessvärre kan jag inte skryta med att detta är lika komplicerat som det
låter. En förklaringsmodell, eller hypotes, inom ekologi går helt sonika ut på att förklara ett
fenomen eller samspel i naturen. I merparten av mina många experiment (tre) har jag
undersökt om ’the Intermediate Disturbance Hypothesis’ (IDH) verkligen stämmer. Denna
hypotes går i princip ut på att ’Lagom är bäst’ och passar därför väl in i den svenska kulturen.
Anledning till att just lagom störning är bäst är att då finns flest antal arter, eftersom alla arter
dör ut om det blir för mycket störning och att bara en art kommer ta över hela samhället om
det inte finns någon störning alls. Det sistnämnda kallas ’konkurrensuteslutning’ och innebär,
kanske inte helt otippat, att en art kan konkurera så effektivt att den utesluter alla andra arter
ur ett område om ingenting stoppar den. Exempel på när detta sker i naturen är barrskogar och
musselbankar, där en eller ett fåtal arter helt egoistiskt kan ta upp väldigt stora områden. Om
en störning kommer in och dödar ett antal individer i dessa områden kan andra, nya, arter
etablera sig på den nyligen frigjorda ytan eller marken. Antalet arter i området ökar då alltså,
och är man lite fin i kanten kan man istället uttrycka detta som att den biologiska mångfalden
höjts. En annan väldigt rolig hypotes, som bygger på den ovan nämnda IDH, kallas ’the
Dynamic Equilibrium Model’ (DEM). Tillägget i denna hypotes är att mängden störning som
är lagom beror på hur fort arterna i ett samhälle växer. Desto fortare arterna växer, desto
kraftigare störning krävs för att bryta konkurrensuteslutning av någon självupptagen liten
gynnare. Dessa två hypoteser, IDH och DEM, är vad jag, två GAIS:are och ett gäng ohängda
tyskar testar på marina hårdbottensamhällen, bestående av anemoner, havsborstmaskar,
havstulpaner, hydroider, musslor, mossdjur, svampdjur, sjöpungar samt grön-, brun- och
rödalger, i den första artikeln i avhandlingen.
De andra nagelbitarna till artiklar handlar även de om hypoteserna IDH och DEM, om än lite
mer indirekt och med större fokus på själva störningsmekanismerna. Den näst första artikeln
handlar om störningar som är lika stora i total omfattning, men där en störning som sker
dubbelt så ofta då påverkar en hälften så stor yta. Skillnaden vi hittade här var att störning
med stor yta som skedde mer sällan gav upphov till fler arter, eftersom detta mer effektivt
4
J. Robin Svensson
kunde bryta de slemmiga sjöpungarnas konkurrensuteslutning. I det tredje experimentet
slängde vi ett getöga på skillnaderna mellan samhällen på stenar som skrapar mot varandra i
vågrörelser (fysisk störning), jämfört med samhällen på stenar som blir mumsade på av
promiskuösa strandsnäckor (biologisk störning), samt vilken effekt dessa olika störningar får i
samspel med hur fort samhällen tillväxer (produktivitet). Förutom att de olika typerna av
störning interagerade på olika sätt med tillväxthastigheten, hade de även olika stor effekt
djuren och växterna (algerna) i samhällena. Den fjärde och sista artikeln är mer lik en
debattartikel, fast med stöd av matematisk modellering och en litteraturundersökning, där jag
väldigt ödmjukt påstår att alla andra som jobbar med ekologiska störningar och biodiversitet
gör fel, medan jag själv tvivelsutan gör allt rätt. Anledningen till felaktigheterna är att en del
testar hypoteser om förändring i antal arter med ett mått på hur jämt arter är fördelade istället
för hur många de är. Detta är lite som när Kurt Olsson frågade Patrik Sjöberg hur brett han
har hoppat, eller som att räkna antalet äpplen i päronträd, makrillar i änglaklacken eller
marxister i vita huset.
Summan av kardemumman, efter ett halvt decennium på skattepengar och ett ointagligt
rekord i spindelharpan, är alltså att effekterna av störning hänger på vilken slags störning som
sker, hur man väljer att mäta den, samt vilka arter som finns i samhället där störningen
inträffar. Vill man testa hypoteser om biodiversitet och störning lite grann, så spelar det även
roll hur stark konkurrensen mellan arter och nyetableringen av arter är, samt vilket mått på
biologisk mångfald som används i studien.
5
Ecological disturbances
LIST OT PAPERS
This thesis is a summary of the following papers:
Paper I
Svensson, J. R., M. Lindegarth, M. Siccha, M. Lenz, M. Molis, M. Wahl, and H.
Pavia. 2007. Maximum species richness at intermediate frequencies of
disturbance: Consistency among levels of productivity. Ecology 88:830-838.
Paper II
Svensson, J. R., M. Lindegarth, and H. Pavia. 2009. Equal rates of disturbance
cause different patterns of diversity. Ecology 90:496-505.
Paper III
Svensson, J. R., M. Lindegarth, and H. Pavia. 2010b. Physical and biological
disturbances interact differently with productivity: effects on floral and faunal
richness. Ecology 91:3069-3080.
Paper IV
Svensson, J. R., M. Lindegarth, P. R. Jonsson, and H. Pavia. The Intermediate
Disturbance Hypothesis predicts different effects on species richness and
evenness. Manuscript.
Papers I, II and III was reprinted with the kind permission of from the Ecological Society of
America.
6
J. Robin Svensson
What is ecological disturbance, really? .................................................................................. 8
Definitions of disturbance ...................................................................................................... 8
Agents of disturbance............................................................................................................. 9
Components and quantities of disturbance........................................................................... 11
Differences between Disturbance, Perturbation and Stress ................................................. 13
Ecological Theories on Disturbance ..................................................................................... 15
The Intermediate Disturbance Hypothesis (IDH) ................................................................ 15
The Dynamic Equilibrium Model (DEM)............................................................................ 17
Additional related models .................................................................................................... 19
Prerequisites for the IDH and the DEM .............................................................................. 21
Aspects of Colonization ....................................................................................................... 21
Aspects of Competition........................................................................................................ 22
Considerations of diversity.................................................................................................... 24
Conclusions ............................................................................................................................. 25
References ............................................................................................................................... 26
Acknowledgements................................................................................................................. 32
7
Ecological disturbances
What is ecological disturbance, really?
Since this thesis is entirely devoted to ecological disturbances, we might as well start at the
beginning. That is, to elucidate the concept of ‘disturbance’. There are quite a few definitions
of disturbance that I will explain and discuss in the first section, whereafter I move on to
agents of disturbance, followed by measures and components of disturbance. An agent of
disturbance is the instrument that causes the damage, such as an animal, waves or fire. The
components of disturbance are the properties of the damaging force of the disturbance agent,
i.e. the heat of the fire, the strength of the waves and the extent of borrowing by an animal.
The issues regarding agents and components of disturbance are discussed in paper I and
specifically tested in papers II and III. Should I not have failed entirely in my attempt at
illuminating the audience on the topic of disturbance in these earlier sections, she or he will
have an appropriate background for the following sections on ecological theories on
disturbance. More specifically, I will sort out the most prominent hypotheses and models on
the effects of disturbance on biodiversity, i.e. the Intermediate Disturbance Hypothesis (IDH)
and the Dynamic Equilibrium Model (DEM), as well as a few related models on colonization
and the specific components of disturbance. The IDH predicts maximum diversity at
intermediate levels of disturbance, whereas the DEM predicts that the level of disturbance
required to maximize diversity depends on the level of productivity. The IDH is tested by
manipulative experimentation in papers I-III and theoretically evaluated in paper IV, and
tests of the DEM is incorporated in the experiments in papers I and III. Furthermore, I will
present and discuss a number of possible prerequisites, or assumptions, which these models
may rely on. In conclusion, readers that have the stamina to go through the entire thesis will
be handsomely rewarded by superior knowledge about definitions, agents and components of
disturbance as well as of theories on disturbance and their associated predicaments. Hence,
they will know what ecological disturbance really is.
Definitions of disturbance
There are quite a few definitions of disturbance, which may or may not help the reader
depending on their complexity and explicitness. The most straightforward definition is that by
Grime (1977), who defines disturbance as partial or total destruction of biomass. Although
simplicity is something to strive for, especially to increase the operationalization of a
definition for manipulative experiments, a too simple definition can include processes and
mechanism that may in fact only have a marginal effect on species assemblages. The
definition by Pickett and White (1985) where disturbance is “…any relative discrete event in
time that disrupts ecosystems, community, or population structure and changes resources,
substrate availability, or the physical environment”, is also very broad. Although this
definition is undoubtedly more explicit, it still encompasses many events that occur naturally
and frequently without necessarily have any measurable effects on either diversity or density
of species. An extension to this definition was added by Pickett et al. (1989), in which
“Disturbance is a change in the minimal structure caused by a factor external to the level of
interest”. A benefit with this hierarchical view of disturbance is that one must consider the
scale at which a certain disturbance operates. For instance, an herbivorous insect can be a
disturbance to the leaves of a single tree, whereas if the study site is an entire forest it may be
more relevant to consider wind-throws by hurricanes or large scale forest fires. However, this
hierarchal view does not compensate for the drawbacks of the broadness of the original
definition.
Notable distinctions in definitions comes from of Pain and Levin (1981) and Reynolds et al.
(1993), who argue that disturbance should be defined exclusively based on its measurable
8
J. Robin Svensson
effect on ecological communities. In contrast to descriptions encompassing a range of
different processes (c.f. Pickett and White 1985). According to Pain and Levin (1981), “Patch
birth rate, and mean and maximum size at birth” can be used as “adequate indices of
disturbance.” The definition of a ‘patch’ here is the primary substratum, i.e. space, that is
affected by the disturbance. Similarly, Reynolds et al. (1993) defines disturbances as
”primarily non-biotic, stochastic events that results in distinct and abrupt changes in the
composition and which interfere with internally-driven progress towards self-organisation and
ecological equilibrium; such events are understood to operate through the medium of (e.g.)
weather and at the frequency scale of algal generation times”. As indicated by the subordinate
clause in this definition, it is explicitly intended for studies on phytoplankton, and the
definition by Pain and Levin (1981) only holds for communities where primary space is the
limiting resource. Hence, while both definitions are useful within their own fields of study,
they will not hold for ecological studies on disturbance and diversity in general.
The more operational definitions of disturbance include the alterations of resources as a
consequence of a disturbing force. For instance, Shea et al. (2004) define disturbance as an
event which “alters the niche opportunities available to the species in a system” by removing
biomass and “freeing up resources for other organisms to use” or in any other way cause “a
direct shift in available nutrients”. Similarly, Mackey and Currie (2000) define disturbance as
“a force often abrupt and unpredictable, with a duration shorter than the time between
disturbance events, that kills or badly damages organisms and alters the availability of
resources”. The inclusion of freeing of resources is important because this is the characteristic
of a disturbance which may ultimately lead to a positive effect on diversity, if the availability
of resources enables, or maintains, coexistence in a community. According to Sousa (1984),
disturbance is defined as “…a discrete, punctuated killing, displacement, or damaging of one
or more individuals (or colonies) that directly or indirectly creates an opportunity for new
individuals (or colonies) to become established.” Hence, instead of considering availability or
resources, which may or may not affect recruitment, this definition goes straight to the core of
the potential for a disturbance to mediate coexistence. That is, opportunities for recruitment
created, directly or indirectly, by disturbance, because without new species recruiting into the
space freed by disturbance diversity cannot increase (Osman 1977, Collins et al. 1995,
Huxham et al. 2000). Thus, like many other researchers, I find this definition of disturbance to
be the most practical and operational for investigations of patterns between diversity and
disturbance. Consequently, the definition of disturbance by Sousa (1984) will be used
throughout this thesis, with the addition that the disturbance should be ecologically relevant
for the system under study. Similar to the arguments by Pickett et al. (1989), a disturbance
should be considered in relation to scale, but also to relevance of agents and components of
disturbance for the specific system and/or the phenomena the model or hypothesis is intended
to explain.
Agents of disturbance
The mechanisms and processes that are inflicting damage upon species assemblages are called
agents of disturbance. Commonly, researchers on disturbance distinguish between biological
and physical agents of disturbance (McGuinness 1987, Wootton 1998, Sousa 2001), while
some authors use more explicit subdivision (Menge and Sutherland 1987). In order to give a
clear picture of what these agents are, I will describe some of the more common agents of
disturbance used in previous studies. Examples of agents of physical disturbance include
anoxia (Diaz and Rosenberg 1995), boat traffic (Willby et al. 2001), desiccation (Lenz et al.
2004), deposition (Miyake and Nakano 2002), drifting logs (Dayton 1971), erosion (Fox
9
Ecological disturbances
1981), fire (Eggeling 1947), floods (Lake et al. 1989), ice-scouring (Gutt and Piepenburg
2003), pesticides (Szentkiralyi and Kozar 1991), pollution (Benedetti-Cecchi et al. 2001),
sediment movement (Cowie et al. 2000), temperature (Flöder and Sommer 1999), tilling
(Wilson and Tilman 2002), trawling (Tuck et al. 1998), tree poisoning (Sheil 2001), tree
lopping (Vetaas 1997), wind (Molino and Sabatier 2001), wave action (McGuinness 1987),
and even warfare (Rapport et al. 1985). Biological disturbances are mainly predation (Talbot
et al. 1978) and grazing (Collins 1987), although some authors add algal whiplash (Dayton
1975), burrowing (Guo 1996), disease (Ayling 1981), parasites (Mouritsen and Poulin 2005)
and trampling (Eggeling 1947).
Due to the differences among these agents of
disturbance, agents are commonly divided
into groups based on their functional or
mechanical characterizations. Menge and
Sutherland (1987) divide the agents of
disturbance into four different groups:
physical
disturbance,
physiological
disturbance, biological disturbance and
predation/grazing. Physical disturbance is
produced by mechanical forces (e.g.
movement of air, water, and sediment),
whereas physiological disturbance is the
lethal effects produced by biochemical
reactions (influenced by e.g. temperature,
light or salinity). Biological disturbance is
the lethal effects of the activities of mobile
animals (e.g. trampling, burrowing, and
digging), and predation and grazing is
defined as mortality resulting from
consumption by animals. In a similar fashion,
Wootton (1998) suggests that the effects of
consumers should be considered separate to
the effects from physical disturbance,
because “the biota of the community is less
likely to directly control the dynamics of the
latter”. That is, agents of biological
disturbance may be density dependent to a
much higher degree than agents of physical
disturbance.
Fig. 1 Disturbance treatment in papers I and II.
Physical scraping of settling panels removing all
organisms from a given percentage (i.e. 20 or 40 %)
of the panel at each disturbance event.
10
An even more important distinction between
agents, than those given above, is based on
their possibility for selectiveness in the
damage they exert. Grazing and predation
have been argued to be unsuitable agents of
disturbance in studies on disturbancediversity patterns, because consumers, unlike
physical agents, may have preferences in
prey species (e.g. McGuinness 1987, Sousa
2001). Due to this predicament, Sousa (2001)
J. Robin Svensson
reserves the term disturbance to include “damage, displacement or mortality caused by
physical agents or incidentally by biotic agents”, thus, excluding consumption by grazers and
predators. Since this possible high degree of selectivity has no comparison in physical
disturbances, outcomes of studies on disturbance using biological agents may be confounded
and, therefore, not generally applicable. For instance, if a consumer prefers prey species that
are inferior competitors, this biological disturbance will increase the rate of competitive
exclusion instead of breaking the dominance of competitive superiors. This degree of
selectivity may be even more complex in disturbance-diversity models that include
productivity, i.e. the DEM, because grazers have been shown to prefer plants with higher
nutrient content in both terrestrial (Onuf et al. 1977) and marine systems (Cruz-Rivera and
Hay 2000). Accordingly, in paper III I show that a biological disturbance (grazing by
periwinkles) and productivity interactively affected the number of macroalgal species,
whereas the physical disturbance (wave-action) only affected the number of invertebrate
species in natural marine epilithic assemblages. These patterns were, in part, explained by
differences in the degree of selectivity between disturbances. Accordingly, the non-selective
physical disturbance (scraping) in papers I and II (Fig. 1) affected all groups of species in the
hard-substratum assemblages; annelids, barnacles, bryozoans, hydroids, mussels, seaanemones, sponges and tunicates, as well as green, brown and red macroalgae. Thus, in
contrast to the plain distinction between biological and physical agents of disturbance, a more
operationally beneficial distinction may be that between selective and non-selective agents of
disturbance.
Components and quantities of disturbance
In relation to agents of disturbance, i.e. ‘what is disturbing’, there are also components of
disturbance, i.e. ‘how is it disturbing’. These components, also called attributes (Shea et al.
2004), commonly differ in the way they are characterized and measured. According to Osman
and Whitlach (1978), “a disturbance agent will have two components, frequency and
magnitude”, where frequency is how often a patch is disturbed and the magnitude refers to the
number of disturbed patches. Wootton (1998) identifies three components of disturbance
“increasing average mortality, increasing temporal variability, and increasing spatial
heterogeneity”. There are, however, many more components of disturbance. These may be
divided into conceptual and operational terms of disturbance. The conceptual terms; level,
intensity, severity, magnitude, regime, timing, and shape, are intended to verbally explain or
describe aspects of disturbance, whereas the operational; frequency, extent, duration, time,
size, rate and predictability, can be measured using their defined quantities (Table 1).
The drawback with the inexplicitly defined conceptual terms is that they are not easily
generalized among studies. For example, ‘intensity’ has been used to describe a variety of
experimental manipulations and variables, such as penetration depth per bite by limpets
(Steneck et al. 1991), type of mechanical scrubbing (McCabe and Gotelli 2000) and degree of
oscillation in sediment (Garstecki and Wickham 2003). Similarly, ‘magnitude’ can be a
general description, occasionally used synonymously to level, intensity and severity.
However, magnitude can also be used for more specific measures, such as the number of
patches affected by disturbance (Osman and Whitlatch 1978) and the percentage of biomass
removed by floods (Kimmerer and Allen 1982). The fact that the units and meaning of
disturbance can be unclear, and differ among studies (Pickett and White 1985, Sousa 2001,
Shea et al. 2004), may be a consequence of the unclear formulations of the hypotheses the
studies aim to test. This is because the most prominent models on patterns between
disturbance and diversity (see section: ecological theories on disturbance) are conceptual
11
Ecological disturbances
models based on relatively scaled variables (Schoener 1972, Peters 1991). However, in order
to evaluate general ecological theories, it is important that concepts are commensurable
among studies.
Table 1 Conceptual and operational terms of disturbance commonly used in ecological studies.
Term
Meaning
Quantity
Conceptual
Generic term for the types and components of
disturbance currently acting in a given area
-
‘level’
General description of overall amount of disturbance
-
‘severity’
General description used synonymously to intensity and
magnitude, and/or specific for damage caused
-
‘intensity’
General description used synonymously to severity and
magnitude, and/or specific for disturbing force
-
‘magnitude’
General description, but also used synonymously to
severity and intensity
-
‘regime’
‘timing’
‘shape’
When a disturbance occurs and influence of the current
conditions at that time
Specific shape (i.e. oval, rectangular, square) of two- or
three-dimensional space disturbed
-
Operational
-1
‘frequency’
Number of disturbance events per unit time
time
‘time’
Period of time since last disturbance event
time
‘duration’
The amount of time a disturbance event lasts
time
‘phasing’
‘predictability’
Temporal pattern of disturbance
Variance in mean time between disturbances
"S", i.e. time
variance
‘size’
Size of an individual disturbance events
‘extent’
‘rate’
Total two- or three-dimensional space disturbed
Product of area and frequency
area
area or volume
-1
area x time
One effort to increase the commensurability among studies on disturbance is the proposal of
the term ‘rate’ of disturbance by Miller (1982), where rate is the sum of the size of all
disturbance events in a given area per unit time, i.e. the product of area and frequency of
disturbance. This is comparable to the argument of Osman and Whitlach (1978), who
suggested that disturbance is composed of the two components frequency and magnitude,
although they did not suggest a general joint measure. Similarly, Petraitis et al. (1989) defines
‘intensity’ as the product of area and frequency (not be confused with the common definition
of the term intensity; Connell 1978, Sousa 1984, Shea et al. 2004). Taking into account the
combined effects of area and frequency is important, because information about one of these
components makes little sense without the context of the other. For instance, specifying an
experimental manipulation where a community is disturbed once a week is completely
12
J. Robin Svensson
uninformative if we do not know the extent of the damage. Without doubt, the differences in
effects on diversity will differ massively if the area disturbed each week is 1% of the total
area compared to if it is 99%. However, disturbances composed of area and frequency are not
the only ones that would benefit from a measure that combines the quantities of components.
For example, in experiments on forest fires the temperature is vital for the effects on
communities (e.g. Gignoux et al. 1997), and this can be combined with both the extent and the
duration for increased commensurability among studies. Although the combined effects of
disturbance components are always implicit in experimental studies, it is necessary to
transform the measure of disturbance into a joint measure, i.e. rate, in order to put any
experimental result into a wider context, and to allow for direct and meaningful comparisons
among studies.
The main benefit of careful specifications of the components of disturbance is that they give
information of the manner in which a particular disturbance is exerted. Even for joint
measures, such as rate, it is important to specify each component clearly. This is important
because disturbances that are equal in extent can nonetheless have significantly different
effects on diversity, depending on how the disturbance is distributed (Bertocci et al. 2005,
papers II and III). In paper II I show that equal rates of disturbance may still give different
patterns in diversity depending on the specific combination of area and frequency, i.e. the
regime of disturbance. In accordance with the predictions by Miller (1982), the regime with
small, frequent disturbances favoured colonizing species, whereas large, less frequent
disturbances favoured competitive dominants. On a similar note, Bender et al. (1984)
identified two different types of disturbance, pulse and press, defined as instantaneous
alteration of species number (pulse) and the sustained alteration of species densities (press).
The distinction between two clearly different mechanisms of disturbance, which may
nonetheless be equal in total extent, can be useful for predictions of patterns of diversity. In
paper III, the biological, continuous small-scale, disturbance (i.e. press) differed in effects on
diversity from the physical disturbance, instantaneous removal or damage of individuals (i.e.
pulse). This shows that clear specification of components of disturbance is important, because
the way the damage of a given disturbance is exerted can be vital for the outcome of studies
on disturbance-diversity patterns.
Differences between Disturbance, Perturbation and Stress
In ecological studies, the two concepts ‘perturbation’ and ‘stress’ are often used
synonymously to disturbance (e.g. Connell 1978, Bender et al. 1984, Rapport et al. 1985).
Processes and mechanisms that are generally described as disturbance may instead be
classified as either perturbation (Webster and Patten 1979, Lane 1986) or stress (e.g.
McGuinness 1987), and the terms perturbation and stress are often used interchangeably with
disturbance without explicitly definitions of any of the terms (e.g. Caswell and Real 1987,
Davies et al. 1999). Similarly, the term perturbation can be used to refer to the effects of stress
on a system (Petraitis et al. 1989) and the term stress can be used to describe a perturbation
(Odum et al. 1979). That these three terms are used haphazardly can be problematic, because
definitions of ecological phenomena may be vital for experimental design in tests of
hypotheses. Especially, since the concept of disturbance is in itself a quagmire, confounding it
with stress or perturbation would be severely suboptimal.
The most clear distinction among these three terms is that between disturbance and stress,
where disturbance is generally considered to cause more severe damage (Grime 1977,
Pickett et al. 1989, Wootton 1998). Among the most common mechanisms and processes
13
Ecological disturbances
described as stress are desiccation (Dayton 1971), pollutant discharges (Rapport et al.
1985) and fluctuations in temperature (Jackson 1977), nutrients (Menge and Sutherland
1987) and light (Grime 1977). According to Grime (1977) stress in plant communities is
defined as “the external constraints which limit the rate of dry-matter production of all or
part of the vegetation”, which is clearly distinct from disturbance events that “limit the
plant biomass by causing its destruction”. Wootton (1998) makes a similar distinction
between stress and disturbance, where the upper limit of what can be defined as stress is
mortality. Stress is here defined by “causing changes in performance as opposed to
mortality”, and he states that stress can also “reduce conversion efficiency or increase
metabolic costs”. This view is also shared by Sousa (2001) who states that the difference
between disturbance and stress, although possibly caused by the same agent, is that
disturbance only occurs when “an organisms tolerance is exceeded, resulting in its death or
sufficient loss of biomass that the recruitment or survival of other individuals is affected”.
Pickett et al. (1989) defines stress as a “change in the interaction maintaining a minimal
structure”, caused “directly or indirectly by an external factor”. For example, an
herbivorous insect can be a disturbance to a leaf by disrupting its physiological integrity,
but a stress to the plant because leaf damage may affect the performance and reproduction
of the plant. Thus, the same mechanism will be classified as either disturbance or stress
depending on the level of interest (Pickett et al. 1989). Rapport et al. (1985) defines stress
as “an external force or factor, or stimulus that causes changes in the ecosystem, or causes
the ecosystem to respond, or entrains ecosystemic dysfunctions that may exhibit
symptoms”. This definition is not among the more operational, since it is only applicable at
the ecosystem level and it is not intuitive what a symptom of an ecosystemic dysfunction
may be. Another thought-provoking definition of stress is that by Rykiel in which stress is
“a physiological or functional effect; the physiological response of an individual, or the
functional response of a system caused by disturbance or other ecological process; relative
to a specified reference condition; characterized by direction, magnitude, and persistence; a
type of perturbation”. Thus, according to this definition, stress is a type of perturbation that
is the effect of disturbance. Here, I much prefer the views of Grime (1977), Wootton
(1998) and Sousa (2001), where stress is generally distinguished from disturbance as nonlethal effects and responses.
Agents of perturbation are commonly similar to those of disturbance and stress, such as flood
scouring (Webster and Patten 1979), environmental variation (Lane 1986), alteration of
species densities (Bender et al. 1984). Furthermore, this concept is also used for processes and
mechanisms that are not easily defined, such as departure from a normal state (Pickett and
White 1985), divergence in spatial organization of badger populations due to bovine
tuberculosis (Tuyttens et al. 2000) and the falling of leaves on spider webs (Leclerc 1991).
Moreover, the term unperturbed is used by Padisak (1993) to describe systems unaffected by
either disturbance or stress. Although definitions of perturbation are scarce in the literature,
there are a few notable exceptions. Rykiel (1985) defines perturbation as “the response of an
ecological component or system to disturbance or other ecological process as indicated by
deviations in the values describing the properties of the component or system; relative to a
specified reference condition; characterized by direction, magnitude, and persistence”. Hence,
according to Rykiel (1985) disturbance is the agent causing damage whereas perturbation, as
well as stress, is the effects of a disturbance. Distinguishing between the cause and effect of
disturbances is not unimportant, for instance, if a process defined as disturbance does not
invoke any measurable response in the recipient community it is questionable whether a
disturbance has really occurred. However, this interpretation of the terms has not been widely
accepted, which is likely due to the rather counter-intuitive terminology of stress- and
14
J. Robin Svensson
perturbation-causing disturbances. Another exception is the definition by Picket and White
(1985), where perturbation is “a departure (explicitly defined) from a normal state, behaviour,
or trajectory (also explicitly defined)”. Although this definition is rather unclear and
exceptionally broad, it may in this case be both appropriate and useful. In the sense that
Padisak (1993) uses the term, but in contrast to Rykiel (1985), it may be beneficial to reserve
a word that describes process and mechanisms that can be either disturbance or stress, or in
fact neither.
Ecological Theories on Disturbance
Disturbance has been recognized as a structuring force in ecological communities since the
beginning of the last century (Cooper 1913). However, it was not until the 1970ies that
disturbance was regarded as a key process in general ecological theory (Dayton 1971, Grime
1973, Levin and Paine 1974). Since then, a number of hypotheses have been proposed to
address the involvement of disturbance in ecological phenomena. These hypotheses mainly
concern succession and biodiversity (Connell 1978, Miller 1982, Dial and Roughgarden
1998), but also on evolutionary processes (Benmayor et al. 2008), biological invasions (Davis
et al. 2000) and ecosystem functions (Cardinale and Palmer 2002). More recently, the
productivity in natural communities, another key process in ecology (Connell and Orias 1964,
Tilman 1980, Abrams 1995), has been suggested to act in concert with disturbance, which
may explain more complex patterns in species diversity (Huston 1979, Kondoh 2001, Worm
et al. 2002). The following sections will focus on the most common hypotheses and models
on effects of disturbance on biological diversity, the interactive effects of disturbance and
productivity, as well as possible assumptions or prerequisites that these models may rely on.
The Intermediate Disturbance Hypothesis (IDH)
The most prominent theory on disturbance, and possibly ecology in general, is the
Intermediate Disturbance Hypothesis (IDH; Connell 1978) (Fig. 2). The original paper by
Connell (1978) has been cited over 3300 times and the IDH also represents one of few well
established ecological theories with an impact on management of marine and terrestrial
national reserves and parks, e.g. Yellowstone National Park, USA (Wootton 1998). The origin
of the IDH is, however, debated (Wilkinson 1999). Even though J. H. Connell is commonly
credited as the originator of the IDH, his main argumentation relies on the much earlier work
of Eggeling (1947) on patterns of diversity in African rain forests (see: Fig. 1 in Connell
1978). In his article, Wilkinson (1999) also identifies three well-known authors who all, prior
to the work of Connell, discussed relatively higher diversity at some form of intermediate
level of disturbance; E. P. Odum (1963), J. P. Grime (1973), and H. S. Horn (1975).
Similarly, Osman (1977) identified “an optimal frequency of disturbance at which diversity is
maximized” in his study on marine epifaunal communities, which he argues is caused by
reductions at high and low levels of disturbance “because of a decrease in the number of
species present or an increase in dominance”. Surprisingly, neither of Odum (1963), Grime
(1973), Horn (1975) or Osman (1977) is cited in the review article by Connell (1978).
The IDH predicts that diversity will reach its maximum at intermediate levels of disturbance,
while remaining low at high and low levels of disturbance (Fig. 2). The rationale for this is
that at low levels of disturbance strong competitors exclude competitively inferior species and
communities are dominated by a few species. Intermediate levels of disturbance, however,
disrupt competitive hierarchies by increasing levels of mortality and thus making free space
15
Ecological disturbances
available for recruitment of competitively inferior species. At successively higher levels of
disturbance, recruitment cannot balance the high levels of mortality and slow recruiting
High
Diversity
B
A
Low
A
C
High
Disturbance
B
C
Fig. 2 The hump-shaped pattern between disturbance and diversity as predicted by the Intermediate
Disturbance Hypothesis (IDH). The mechanisms of the IDH are illustrated by settling panels (A, B and C)
used in papers I and II. At point A diversity is low due to competitive exclusion, at point B coexistence is
enabled by freeing space for new species, and at point C few species survive due to high level of disturbance.
species disappear from the community. The drawback of this straightforward logic, and hence
its conceptual appeal, is that it has received criticism from both empirical and theoretical
studies for being too simplistic (Pacala and Rees 1998, Huxham et al. 2000, Shea et al. 2004).
Furthermore, a literature review revealed that only 20 % of the studies on effects of
disturbance on diversity showed the unimodal pattern predicted by the IDH (Mackey and
Currie 2001). Nevertheless, the IDH has been supported in field experiments in terrestrial
(e.g. Armesto and Pickett 1985, Collins 1987, Molino and Sabatier 2001), freshwater (e.g.
Padisak 1993, Reynolds 1995, Flöder and Sommer 1999) and marine communities (e.g.
Osman 1977, Sousa 1979a, Valdivia et al. 2005), as well as in laboratory experiments (e.g.
Widdicombe and Austen 1999, Buckling et al. 2000, Cowie et al. 2000) and model
evaluations (Petraitis et al. 1989, Dial and Roughgarden 1998, Li et al. 2004). In accordance
with these studies, the characteristic hump-shape pattern between disturbance and diversity
was observed in papers I, II and IV.
The apparent simplicity of the IDH may, however, be slightly deceiving. There are, in fact,
many aspects of the IDH and the way that disturbance may determine levels of diversity.
Although I will spare the reader yet another section on components of disturbance, there are
16
J. Robin Svensson
some fundamental differences among the mechanisms of disturbance in relation to the
hypothesis that should be noted. For instance, how often a disturbance occurs (i.e. frequency),
how large the disturbance is (i.e. area or extent) and time since the last disturbance (i.e. time).
Even though they are all interrelated, through the main rationale of disrupting competitive
exclusion, the underlying mechanisms may be different. In the case of frequency, high levels
of diversity can be maintained if the disturbance events occur often enough to prevent any one
species from achieving dominance, while not occurring so often that only few species can
persist. When the extent of disturbance is considered, areas that are too large will eliminate all
species, areas that are too small will have little or no impact, whereas intermediate areas may
disrupt competitive exclusion and allow establishment of new species in the disturbed
patches. In comparison, the time aspect states that high diversity will be observed at some
point in time after recolonization of the disturbed area, but before the community returns to its
successional climax (i.e. dominance by few species). The main difference here is commonly
referred to as the ‘between patch’ vs. ‘within patch’ mechanisms (e.g. Wilson 1990), or
sometimes as the resetting of a patch successional clock vs. the creation of a successional
mosaic (e.g. Chesson and Huntly 1997). This distinction is articulated in a straightforward
way by Wilson (1994): “A single patch does not have a frequency of disturbance, only a time
since last disturbance”. Albeit a bit drastic, it has been suggested that the within patch aspect
is not a mechanisms of coexistence, as much as a mere observation of succession (Wilson
1990, Wilson 1994, Chesson and Huntly 1997). In contrast, the successional mosaic, or
between patch, explanation relies on disturbances occurring in a greater area, where disturbed
patches are all in different stages of succession and may, thus, together compose a high
regional diversity (Levin and Paine 1974, Chesson and Huntly 1997, Sheil and Burslem
2003).
One way to resolve the discussion about the differences between the within-patch and the
between-patch mechanisms of the IDH, could be to consider the different components of
disturbance, i.e. how the damage from the disturbance is exerted. Bender et al. (1984)
distinguishes between ‘pulse disturbance’, i.e. instantaneous alteration of species number,
and ‘press disturbance’, i.e. the sustained alteration of species densities (see also section
‘Components and quantities of disturbance’). A press disturbance could unceasingly prevent
competitive exclusion of a dominant species, which yields higher within-patch diversity. In
contrast, a pulse disturbance would provide patches of different successional stages and ages
(younger more r-selected and older more K-selected species), giving rise to the higher
between-patch diversity. Hence, this subdivision of disturbance could perhaps be a missing
link in the so far unresolved issue (see Sheil and Burslem 2003) of differentiating the withinpatch from the between-patch mechanisms of the IDH.
The Dynamic Equilibrium Model (DEM)
The Dynamic Equilibrium Model (DEM; Huston 1979, Kondoh 2001) relies on the same
general coexistence mechanisms as the IDH (Fig. 3). At low levels of disturbance one, or few,
species will dominate and exclude all other species, and at high levels of disturbance very few
species can persist, while coexistence is possible at intermediate levels. The addition in the
multifactorial model DEM is that the relationship between disturbance and diversity is
modified by the level of productivity. Huston (1979) suggested that increased productivity,
and thus growth rates of individuals and populations, means that a more severe disturbance is
required to prevent competitive exclusion. Consequently, at low productivity, and slow
growth rates, maximum diversity is observed already at low levels of disturbance because
competitive exclusion occurs at a lower rate. Thus, the shape of the relationship between
17
Ecological disturbances
disturbance and diversity is predicted to be of three general types: monotonically decreasing
(at low productivity), unimodal (when productivity is intermediate) and monotonically
increasing (when productivity is high). Although the DEM has not been experimentally
evaluated nearly as much as the IDH, there are corroborating manipulative studies from
aquatic as well as terrestrial systems (e.g.Turkington et al. 1993, Worm et al. 2002, Jara et al.
2006). However, in paper I, there was no effect on diversity of the manipulated increase in
productivity, whereas maximum species richness was observed at intermediate levels of
physical disturbance, in accordance with the IDH. This is likely explained by the productivity
treatment, which, despite a general effect on growth rates of algae, did not affect the
competitive dominants in the hard substratum assemblages. Thus, the rate of competitive
exclusion was not measurably affected and more frequent disturbance was consequently not
required to prevent exclusion of inferior competitors at high levels of productivity.
Diversity
High
Low
P-high
Disturban
ce
P-intermediate
High
P-low
Fig. 3 The patterns predicted by the Dynamic Equilibrium Model (DEM). At low levels of
productivity, maximum diversity is observed already at low levels of disturbance due to low rates
of competitive exclusion. At intermediate levels of productivity intermediate levels of
disturbance is required, and high levels of productivity high levels of disturbance is required, in
order to disrupt competitive exclusion by dominants and free resources for colonizing species.
Similar to the IDH, agents and components of disturbance may influence the outcome of tests
on the DEM. For instance, biological and physical agents may differ in selectivity
(McGuinness 1987, Wootton 1998, Sousa 2001) and consumers often prefer prey with higher
nutrient content (Emlen 1966, Onuf et al. 1977, Pavia and Brock 2000). One indication of a
discrepancy between agents of disturbance is that interactive effects between biological
disturbance and productivity has been observed in many studies from various environments
(see Proulx and Mazumder 1998 and references therein), whereas tests of the DEM using
physical disturbance have more variable outcomes (e.g. Turkington et al. 1993, Death and
Winterbourn 1995, Death 2002, Jara et al. 2006). In paper III, in order to test for possible
differences among agents, I contrasted the effects of a biological to that of a physical
disturbance in an experiment on the DEM. Using natural sessile assemblages on boulders (i.e.
epilithic communities) composed of invertebrates and macroalgae, I tested for interactive
18
J. Robin Svensson
effects between productivity (high vs. ambient), physical disturbance (simulated wave-action
at five distinct frequencies) and biological disturbance (grazing by periwinkles manipulated as
absent or present). The number of algal species was interactively affected by productivity and
biological disturbance, whereas the invertebrate richness was affected by physical disturbance
only. This may in part be explained by difference in degree of selectivity between agents, but,
more interestingly, also in the way the damage is exerted. When biomass is slowly reduced, as
exerted by the biological, continuous small-scale disturbance (i.e. press disturbance; Bender et
al. 1984), this effect can more easily be counteracted by increased growth of the affected
organisms (Huston 1979, Kondoh 2001). In contrast, increased individual growth rate cannot
easily compensate for instantaneous loss of individuals, as exerted by the physical disturbance
(i.e. pulse disturbance; Bender et al. 1984). In accordance with these arguments and our
results, Kneitel and Chase (2004), the only previous study that has tested for interactions of
all three factors, also found that biological disturbance (predation), but not physical
disturbance (drying), and productivity interactively affected species richness. Thus, agents
and components of disturbance may not only influence disturbance-diversity patterns, but also
the specific interactive effects between disturbance and productivity on biological diversity of
natural communities.
Additional related models
The only model on effects of disturbance on diversity that specifically considers the different
components of disturbance is that by Miller (1982). In his article, he introduces the term ‘rate’
of disturbance, i.e. the product of area and frequency, which, thus, takes into account the total
amount of disturbance inflicted upon a community (see also section ‘Components and
quantities of disturbance’). According to Miller (1982), small, frequent disturbances favour
species with rapid vegetative growth (i.e. ‘competitors’), whereas large, less frequent
disturbances favour species with high capacity for dispersal (i.e. ‘colonizers’) due to the
differences in perimeter to area ratios among patches. Although Miller (1982) predominantly
focuses on the area of disturbance, the other component of the rate, frequency, is equally
important. Similar to variations in area, differences in frequency and timing of disturbance
will influence the abundance and composition of natural communities (Sousa 2002). This is
because species are likely to increase in abundance when the disturbance regime matches their
preferred recruitment time (Underwood and Anderson 1994, Crawley 2004). Furthermore,
because of the natural large variation in temporal distribution of propagules among species
(Roughgarden et al. 1988, Underwood and Anderson 1994) a single large disturbance can
only be colonized by the propagules that are available at the specific time when a limiting
resource, i.e. space, is made free. In paper II I tested the model by Miller (1982), or more
specifically if the specific combination if area and frequency matters even if the rate is kept
constant. In accordance with the predictions by Miller (1982), the regime with small, frequent
disturbances favoured colonizing species, whereas large, less frequent disturbances favoured
competitive dominants. Thus, as is claimed in the title, equal rates of disturbance did cause
different patterns in diversity.
In a model on the importance of the timing of disturbance, Abugov (1982) introduces the
concept of disturbance ‘phasing’. Abugov (1982) distinguishes between disturbances that are
phased compared to those that are unphased. A phased disturbance means that all patches are
cleared simultaneously, and the patches are termed to be ‘in phase’. Conversely, during
unphased disturbance, the probability of a patch being cleared by disturbance is independent
of the disturbance of other patches. Phased disturbances are considered to be more large scale
disturbance events such as storms or forest fires, whereas constant predation is given as an
19
Ecological disturbances
example of unphased disturbance. The outcome of Abugov’s model showed that highest
diversity always occurred at intermediate levels of disturbance, regardless of the degree of
phasing, but also that the diversity at any given level of disturbance depend on the degree of
phasing. Furthermore, similar to the multifactorial model DEM, high levels of diversity was
observed at intermediate degree of phasing at intermediate levels of disturbance. The idea of
phasing is similar to that of temporal variability in disturbance, which has been shown to
affect the community structure of benthic assemblages on rocky shores (i.e. Bertocci et al.
2005, but see: Sugden et al. 2007). It is also similar to the concepts of ‘Nonadditivity’
(Chesson 2000), ‘Storage Effect’ (Chesson and Huntly 1997) and ‘Spatiotemporal Niche
Creation’ (Pacala and Rees 1998). The key argumentation in these concepts is that
coexistence is enabled because different species utilize different spatiotemporal niches. The
spatiotemporal niches may differ, depending on environmental fluctuations or disturbance, in
the amount of available resources, the free space for settling and in their current stage of
succession (Amarasekare et al. 2004, Roxburgh et al. 2004, Shea et al. 2004). Due to the
suggestions of coexistence mechanisms that are consider to be alternative, the IDH and the
DEM have been argued to give “inadequate, inconsistent, or improbable explanations” of
species coexistence (see: Chesson and Huntly 1997). However, the main mechanism of
coexistence in all these concepts, including phasing and temporal variability, is that different
patches are at different successional stages and/or differ in availability of resources. Hence, it
could be argued that they are all describing the ‘between-patch’, or ‘successional mosaic’,
aspect of the IDH, where coexistence is maintained, or enabled, by disturbance, because
patches at different stages in succession differ in species composition.
In their investigation of the theoretical validity of the IDH, Dial and Roughgarden (1998)
found what they call ‘the intermediate area hypothesis’ and ‘the intermediate recruitment
hypothesis’. In contrast to most other models on disturbance (Petraitis et al. 1989, Chesson
and Huntly 1997, Kondoh 2001), their mathematical model incorporates the dynamics of
pelagic larvae and benthic adults, as well as hierarchal competition for the limiting resource
space. The larval-benthic dynamics was purposely considered because the pattern predicted
by the IDH is often observed in communities where species have long-lived propagules and
space-limited adults, such as marine invertebrates, macroalgae and seed plants (Sousa 1979a,
Sousa 1979b, Molino and Sabatier 2001, Jara et al. 2006). More specifically, in these systems
the disturbance only affects the sessile adults, while leaving the propagule mortality
unaffected (Dial and Roughgarden 1998). The two key points of the outcome of the model
was that the IDH is a moderate to high settlement phenomenon, and that a subordinate species
must have an adaptation allowing it to survive and/or colonize at levels of disturbance that are
lethal to the dominant, if disturbance, area, or settlement is to allow coexistence. According to
Dial and Roughgarden (1998), these two key points show that the IDH is not a universal
phenomenon, which also leads to the additional outcome of the model, the intermediate area
and recruitment hypotheses. If the level of disturbance, at an intermediate value, is kept
constant, intermediate levels of recruitment lead to coexistence among species. This is
explained by the exclusion of the subordinate species of a dominant superior at high
recruitment, and at low recruitment the dominant cannot exist. However, in their model, area
is equivalent to settlement, thus, yielding a similar intermediate area effect, where smaller
habitats can favour subordinate species’ coexistence with a dominant species. Although it
could be argued that the proposed hypotheses are in fact already inherent functions of the
IDH, since diversity cannot increase if no new species settle (Osman 1977, Huxham et al.
2000, papers II and III), it may still be noteworthy to point out that disturbance is not the only
way exclusion can be prevented and coexistence maintained. Furthermore, it gives important
20
J. Robin Svensson
insights in the underlying mechanisms of coexistence for the IDH, as well as the possible
prerequisites for observing the pattern predicted by the IDH discussed in the next section.
Prerequisites for the IDH and the DEM
In response to the inconsistencies in the outcome of manipulative tests of the IDH (reviewed
by Mackey and Currie 2001), several authors have suggested that the predictions of the IDH
relies on a number of prerequisites. The most common prerequisites, or assumption, are
competitive exclusion (Fuentes and Jaksic 1988), large regional species pool (Osman 1977),
multiple stages in succession (Collins and Glenn 1997), nonlinear resource use (Chesson and
Huntly 1997), availability of spatiotemporal niches (Pacala and Rees 1998) and trade-offs
between competition and tolerance (Petraitis et al. 1989) and between competition and
colonisation (Dial and Roughgarden 1998). Furthermore, Menge and Sutherland (1987)
argued that the effects of disturbance depends on the amount of environmental stress in the
system. However, the constructive criticism in the suggestions of the prerequisites primarily
concerns aspects of two key processes; competition and colonization.
Aspects of Colonization
According to Dial and Roughgarden (1998), the IDH is a ”moderate-to-high settlement
phenomenon”, and Collins et al. (1995) pointed out that it is settlement by propagules that
may allow for increases in diversity, not disturbance per se. That colonization is important in
order for disturbance to have a positive effect on diversity is intuitive and logic. Diversity
cannot increase if there are no available propagules to occupy the space, or any other limiting
resource, which is freed by disturbance (Sousa 2001). Another suggested prerequisite, that is
equally straightforward, but maybe less intuitive, is the importance of a large regional species
pool (Osman 1977). This is because diversity cannot increase if the propagules that establish
in the cleared space, are the same species that originally inhabit the assemblage. This was
clearly shown in a manipulative experiment by Huxham et al. (2000), where the species pool
in the intertidal macrofaunal communities was too small to allow for settlement of new
species in the assemblages subjected to disturbance. Low rate of colonization is also
something that may explain the lack of positive effects of disturbance on diversity in paper III
and at one of three sites in paper II. In the experiment on the effects of physical and
biological disturbance and productivity on natural epilithic assemblages (paper III), the
recruitment of new species occurred at a rate that was not sufficient to counteract the negative
effects of disturbance. Similarly, in paper II, the physical disturbance did not have a
significant effect on the richness of the hard substratum assemblages at one site, where
richness was generally low and new species did not settle in disturbed patches. In contrast to
paper III, this experiment was setup in the waters of the Tjärnö archipelago, where the
regional species pool and availability of propagules per definition was natural. However, it
has previously been shown that local hydrodynamics in areas near this site may hamper the
settling of invertebrate larvae (Berntsson et al. 2004, Jonsson et al. 2004), which also could
explain the surprisingly low total cover in the controls assemblages at this site. Thus, local
hydrodynamics may be of equal importance to the availability of propagules and the size of
the regional species pool, for the outcome of manipulative experiments on the effects of
disturbance on diversity.
21
Ecological disturbances
Aspects of Competition
The other key process in the suggested prerequisites, competition, was mentioned already by
Connell (1978), who considered competitive exclusion to be an assumption for the
coexistence facilitating mechanism of disturbance. Similar to the arguments for colonization,
disturbance cannot increase diversity if there is no exclusion process to interrupt by removing
the dominant(s) and allow new species to establish in a community (Huston 1979, Sousa
1984, 2001). This is also linked to the suggested trade-off between competition and
colonization. If the inferior species cannot out-compete the dominant at colonizing newly
freed substrata, competitive exclusion may not be prevented and diversity will not increase in
response to disturbance (Dial and Roughgarden 1998). Similarly, for the trade-off between
competition and disturbance tolerance, the inferior species must be better adapted to cope
with destructive events, either by physiological tolerance or other means such as fast growth
and re-colonization (Petraitis et al. 1989). Thus, in order for a disturbance to facilitate
coexistence, the dominant species must be comparatively more susceptible to the damage
exerted. Furthermore, the dominant species must also be able to maintain their competitive
advantage in the absence of disturbance (Connell 1978). The importance of competition for
the outcome of experiments on disturbance is clearly shown in paper II, where the three
different responses to disturbance at the three different sites clearly corresponded to the
differences in species composition (fig. 4). Competitive exclusion was evident at the site
where support for the IDH was found, as also observed in paper I, whereas increasing levels
of disturbance only decreased diversity at the site lacking clear dominants in the undisturbed
controls. Although assemblages at the third site also lacked dominants, there was no effect of
disturbance because the initial diversity was so low that even the limited colonization in this
area could counteract the effects of disturbance. Consequently, the same disturbance can give
widely different patterns in diversity depending on the composition of species, and the level
of competition, in communities.
In order to disrupt the competitive advantage of dominants, the destructive event of a
disturbance must potentially affect all species in a similar manner, or, conversely, fall heavier
on the competitive dominants. The problem with possible selectivity of agents has been
discussed for manipulations of disturbance, but not for manipulations of productivity. This
lack of considerations of selectivity in agents may severely confound tests of the DEM. The
DEM predicts that competitive exclusion will increase with productivity, thus requiring a
stronger disturbance to be disrupted, but if the inferior competitors are more strongly affected
by the productivity treatment this could instead slow down the rate of exclusion. This would
cause diversity to peak at lower, rather than the predicted higher, intensities of disturbance.
The issue of the selectivity of agents of productivity was clearly shown in paper I, where the
IDH was supported, but the DEM was not. The most likely explanation for this outcome is
that the dominant species exerting competitive exclusion, the tunicate Ciona intestinalis, was
unlikely to benefit from the manipulation of nutrient availability. Hence, even though the
productivity treatment had a general, positive, effect on growth rates in the assemblages, the
rate of competitive exclusion did not increase, and higher levels of disturbance was
consequently not required to maximize diversity. Even in studies that recognize the issue of
selectivity, there is a practical difficulty of designing a non-selective agent of productivity in
manipulative experiments. Experimental manipulation of productivity in tests of the DEM is
commonly done indirectly, i.e. by adding nutrients or organic matter (Turkington et al. 1993,
Widdicombe and Austen 2001, Worm et al. 2002, Kneitel and Chase 2004, Jara et al. 2006,
Canning-Clode et al. 2008, Sugden et al. 2008). In such manipulations it is necessary to test
independently whether the actual experimental treatment (the adding of nutrients or organic
matter) has an effect on productivity. Without evidence for an actual increase in productivity
22
J. Robin Svensson
a
10
Species Richness
b
Site 1
Site 2
Site 3
9
8
7
6
5
4
0
5
10
15
20
25
Rate of Disturbance
30
(cm2
35
40
45
/ week)
Fig 4 Three significantly different communities at sites 1, 2 and 3 which showed three different
responses to disturbance in paper II. (a) Species composition, as well as pictures, of the control
assemblages at sites 1, 2 and 3 and (b) responses to rates of physical disturbance, significant quadratic
and linear quadratic components, respectively, at sites 1 and 2, and no significant pattern at site 3.
23
Ecological disturbances
experiments cannot perform an adequate test of the DEM, and without information on the
selectivity of the agent of productivity the outcome of tests cannot be adequately interpreted.
Unfortunately, this issue is generally overlooked (e.g. Widdicombe and Austen 2001, Scholes
et al. 2005, Jara et al. 2006). Nevertheless, if predictions about effects of productivity and
disturbance on diversity are to be tested in field experiments, indirect manipulations, such as
adding nutrients or organic matter, may be the only conceivable solution.
Considerations of diversity
Something that is conspicuously absent in the literature is a discussion on the potentially large
variation in outcomes among studies depending on the measure of diversity that is used in
tests of the IDH. As discussed in the earlier sections, nearly every aspect of disturbance has
been considered, e.g. the definitions, the agents, the components, the quantities, how the
damage from disturbance is exerted and a multitude of prerequisites have been suggested to
explain inconsistencies in outcomes of the IDH. In addition, many other aspects of the IDH
have been discussed, such as alternative mechanisms underlying coexistence (Pacala and Rees
1998), influence of the characteristics of communities (Fuentes and Jaksic 1988), interactive
effects of disturbances (Collins 1987), importance of the specific traits of individual species
(Haddad et al. 2008) and the context dependence of intermediacy (Shea et al. 2004). Yet,
despite over 3300 citations of Connell (1978) and ample attention in the scientific literature,
no one has considered the response variable for the conceptual model IDH, i.e. the aspect of
diversity.
Consequently, in paper IV I investigated how the measure of diversity may affect the
outcome of studies on effects of disturbance on diversity. This was done by scrutinizing the
original formulations of the models, conducting a meta-analysis of previously published
studies and through two different approaches to mathematical modelling. In the formulation
of the IDH, Connell (1978) uses the word diversity without any further definition, while
Huston (DEM; 1979) rejects all various indices and considers diversity to be solely richness
and evenness. In the model presented by Miller (1982) diversity is defined as a measure that
includes both “species abundance and number”. However, neither Huston nor Miller makes
an effort to explain what kind of effects disturbance would have on species abundances in
contrast to the number of species. In the meta-analysis I investigated if all measures of
diversity show the same response in studies that use two or measures of diversity within the
same experiment. The mathematical modelling was performed using one already established
spatially implicit model (Kondoh 2001) and one spatially explicit automation model, in order
to specifically contrast the responses to disturbance of the two major components of diversity:
richness and evenness. Both models support the IDH when biodiversity is measured as
species richness, but, in contrast, predict that evenness increases monotonically with
increasing levels of disturbance. The meta-analysis showed that two-thirds of the published
studies in the survey present different results for different diversity measures, and the
comparisons between richness and evenness showed an even higher degree of dissimilarity. In
addition, when the analyses from papers I and II were rerun to include evenness as response
variable (these results were not included in any of the papers), the same patterns as in the
models emerges. Hence, in accordance with the predictions of the two model, species richness
was maximized at intermediate levels of disturbance, and evenness showed linear increases
with increasing rates of disturbance (Evenness: linear component MS=0.95, F=28.7, p<0.01;
MS=1.75, F=81.8, p<0.01, respectively, quadratic component MS=0.010, F=0.30, p=0.58;
MS=0.0052, F=0.24, p=0.63, respectively, Fig. 5). Thus, the meta-analysis, as well as the
24
J. Robin Svensson
mathematical two models and the re-analysis of previous field experiments clearly show that
the measure of diversity is vital for outcomes of tests of the IDH.
0,4
16
0,35
14
0,25
10
0,2
8
6
Richness
4
Evenness
0,15
Evenness
0,3
12
0,1
7
0.7
0.6
6
Evenness
5
0.4
0
0,0
0,2
0,4
0,6
0,8
4
1,0
20
0.00
d
1,2
0.3
0.20
0.40
0.60
0.80
9
0.8
14
0,8
Richness
12
Evenness
10
0,6
8
0,4
6
4
0,2
Evenness
1
Species Richness
18
16
1.00
0.7
8
0.6
7
0.5
Richness
6
Evenness
5
Evenness
0
Species Richness
0.5
Richness
0,05
2
b
c
Evenness
18
Species Richness
Species Richness
a
0.4
0.3
2
0
0
0,0
0,2
0,4
0,6
Magnitude of Disturbance
0,8
1,0
4
0.00
0.2
0.20
0.40
0.60
0.80
1.00
Frequency of Disturbance
Fig. 5 Hump-shaped patterns between species richness and disturbance, but linear increases in
evenness, in the two models from paper IV (a and b) as well as the re-analyzed results from the field
experiments in papers I (c) and II (d).
Conclusions
In this thesis I have clearly (i.e. hopefully) shown that the definition of disturbance can
influence the outcome of studies, depending on which characteristics of disturbances a
particular definition encompasses. The type of agent that is causing the disturbance is crucial,
because selectivity can differ among disturbance agents and biological agents may choose
prey depending on nutritional value. Different components of disturbance can affect
communities in different ways, and even the specific proportions of area and frequency within
the same rate of disturbance can cause different patterns in diversity. The effects of
disturbance will also to a large extent depend on the species composition of the community
upon which it is inflicted. In tests of hypotheses on disturbance-diversity pattern, outcomes
are generally influenced by the rate of competition, the availability of propagules, the regional
species pool and interactions with the abiotic environment. Experimental tests of models that
include productivity should also include explicit investigations of whether the manipulative
treatment significantly affects the overall productivity, as well as the recognition of the
possible selectivity of productivity agents. Furthermore, the measure of diversity used as
response variable is vital for the outcome of tests of hypotheses on effects of disturbance on
diversity. Clearly, there are many aspects to consider in experimental design and
interpretation of results in disturbance-diversity studies. Consequently, in order to increase the
generality and commensurability among studies, it will be of great benefit if experimenters (i)
define the type of disturbance used in the study, (ii) assign ecologically relevant agents of
disturbance and productivity with quantifiable components, (iii) recognize the characteristics
of the community the disturbance is inflicted upon, and (iv) specify, and justify, the measure
of diversity to be used in tests of hypotheses on effects of disturbance on diversity.
25
Ecological disturbances
References
Abrams, P. A. 1995. Monotonic or Unimodal Diversity Productivity Gradients - What Does
Competition Theory Predict. Ecology 76:2019-2027.
Abugov, R. 1982. Species diversity and phasing of disturbance. Ecology 63:289 – 293
Amarasekare, P., M. F. Hoopes, N. Mouquet, and M. Holyoak. 2004. Mechanisms of
coexistence in competitive metacommunities. American Naturalist 164:310-326.
Armesto, J. J. and S. T. A. Pickett. 1985. Experiments on disturbance in old-field plant
communities: impact on species richness and abundance. Ecology 66:230-240.
Ayling, A. M. 1981. The Role of Biological Disturbance in Temperate Subtidal Encrusting
Communities. Ecology 62:830-847.
Bender, E. A., T. J. Case, and M. E. Gilpin. 1984. Perturbation Experiments in Community
Ecology: Theory and Practice. Ecology 65:1-13.
Benedetti-Cecchi, L., F. Pannacciulli, F. Bulleri, P. S. Moschella, L. Airoldi, G. Relini, and F.
Cinelli. 2001. Predicting the consequences of anthropogenic disturbance: large-scale
effects of loss of canopy algae on rocky shores. Marine Ecology-Progress Series
214:137-150.
Benmayor, R., A. Buckling, M. B. Bonsall, M. A. Brockhurst, and D. J. Hodgson. 2008. The
interactive effects of parasitesf disturbance, and productivity on experimental adaptive
radiations. Evolution 62:467-477.
Berntsson, K. M., P. R. Jonsson, A. I. Larsson, and S. Holdt. 2004. Rejection of unsuitable
substrata as a potential driver of aggregated settlement in the barnacle Balanus
improvisus. Marine Ecology-Progress Series 275:199-210.
Bertocci, I., E. Maggi, S. Vaselli, and L. Benedetti-Cecchi. 2005. Contrasting effects of mean
intensity and temporal variation of disturbance on a rocky seashore. Ecology 86:20612067.
Buckling, A., R. Kassen, G. Bell, and P. B. Rainey. 2000. Disturbance and diversity in
experimental microcosms. Nature 408:961-964.
Canning-Clode, J., M. Kaufmann, M. Molis, M. Wahl, and M. Lenz. 2008. Influence of
disturbance and nutrient enrichment on early successional fouling communities in an
oligotrophic marine system. Marine Ecology-an Evolutionary Perspective 29:115-124.
Cardinale, B. J. and M. A. Palmer. 2002. Disturbance moderates biodiversity-ecosystem
function relationships: Experimental evidence from caddisflies in stream mesocosms.
Ecology 83:1915-1927.
Caswell, H. and L. A. Real. 1987. An Approach to the Perturbation Analysis of Optimal Life
Histories. Ecology 68:1045-1050.
Chesson, P. 2000. Mechanisms of Maintenance of Species Diversity. Annual Review of
Ecology and Systematics 31:343-366.
Chesson, P. and N. Huntly. 1997. The roles of harsh and fluctuating conditions in the
dynamics of ecological communities. American Naturalist 150:519-553.
Collins, S. L. 1987. Interaction of Disturbances in Tallgrass Prairie - a Field Experiment.
Ecology 68:1243-1250.
Collins, S. L. and S. M. Glenn. 1997. Intermediate disturbance and its relationship to withinand between-patch dynamics. New Zealand Journal of Ecology 21:103-110.
Collins, S. L., S. M. Glenn, and D. J. Gibson. 1995. Experimental-Analysis of Intermediate
Disturbance and Initial Floristic Composition - Decoupling Cause and Effect. Ecology
76:486-492.
Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302 –
1310.
26
J. Robin Svensson
Connell, J. H. and E. Orias. 1964. The ecological regulation of species diversity. American
Naturalist 98: 399 – 414.
Cooper, W. S. 1913. The climax forest of Isle Royale, Lake Superior and its development.
Botanical Gazette 55:189-235.
Cowie, P. R., S. Widdicombe, and M. C. Austen. 2000. Effects of physical disturbance on an
estuarine intertidal community: field and mesocosm results compared. Marine Biology
136:485-495.
Crawley, M. J. 2004. Timing of disturbance and coexistence in a species-rich ruderal plant
community. Ecology 85:3277-3288.
Cruz-Rivera, E. and M. E. Hay. 2000. The effects of diet mixing on consumer fitness:
macroalgae, epiphytes, and animal matter as food for marine amphipods. Oecologia
123:252-264.
Davies, R. G., P. Eggleton, L. Dibog, J. H. Lawton, D. E. Bignell, A. Brauman, C. Hartmann,
L. Nunes, J. Holt, and C. Rouland. 1999. Successional response of a tropical forest
termite assemblage to experimental habitat perturbation. Journal of Applied Ecology
36:946-962.
Davis, M. A., J. P. Grime, and K. Thompson. 2000. Fluctuating resources in plant
communities: a general theory of invasibility. Journal of Ecology 88:528-534.
Dayton, P. K. 1971. Competition, disturbance, and community organization: the provision
and subsequent utilization of space in a rocky intertidal community. Ecological
Monographs 41:351 – 389.
Dayton, P. K. 1975. Experimental evaluation of ecological dominance in a rocky intertidal
algal community Ecological Monographs 45:137-159.
Death, R. G. 2002. Predicting invertebrate diversity from disturbance regimes in forest
streams. Oikos 97:18-30.
Death, R. G. and M. J. Winterbourn. 1995. Diversity patterns in stream benthic invertebrate
communities: the influence of habitat stability. Ecology 76:1446-1460.
Dial, R. and J. Roughgarden. 1998. Theory of marine communities: The intermediate
disturbance hypothesis. Ecology 79:1412-1424.
Diaz, R. J. and R. Rosenberg. 1995. Marine benthic hypoxia: A review of its ecological
effects and the behavioural responses of benthic macrofauna. Pages 245-303
Oceanography and Marine Biology - an Annual Review, Vol 33.
Eggeling, W. J. 1947. Observations on the ecology of the Budongo rain forrest, Uganda.
Journal of ecology 34:20 – 87.
Emlen, J. M. 1966. The Role of Time and Energy in Food Preference. The American
Naturalist 100:611-617.
Flöder, S. and U. Sommer. 1999. Diversity in planktonic communities: An experimental test
of the intermediate disturbance hypothesis. Limnology and Oceanography 44:11141119.
Fox, J. F. 1981. Intermediate levels of soil disturbance maximize alpine plant diversity.
Nature 293:564-565.
Fuentes, E. R. and F. M. Jaksic. 1988. The Hump-Backed Species Diversity Curve: Why Has
It Not Been Found among Land Animals? Oikos 53:139-143.
Garstecki, T. and S. A. Wickham. 2003. The response of benthic rhizopods to sediment
disturbance does not support the intermediate disturbance hypothesis. Oikos 103:528536.
Gignoux, J., J. Clobert, and J. C. Menaut. 1997. Alternative fire resistance strategies in
savanna trees. Oecologia 110:576-583.
Grime, J. P. 1973. Competitive exclusion in herbaceous vegetation. Nature 242:344-347.
27
Ecological disturbances
Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its
relevance to ecological and evolutionary theory. American Naturalist 111:1169 –
1194.
Guo, Q. F. 1996. Effects of bannertail kangaroo rat mounds on small-scale plant community
structure. Oecologia 106:247-256.
Gutt, J. and D. Piepenburg. 2003. Scale-dependent impact on diversity of Antarctic benthos
caused by grounding of icebergs. Marine Ecology-Progress Series 253:77-83.
Haddad, N. M., M. Holyoak, T. M. Mata, K. F. Davies, B. A. Melbourne, and K. Preston.
2008. Species' traits predict the effects of disturbance and productivity on diversity.
Ecology Letters 11:348-356.
Horn, H. S. 1975. Markovian properties of forest succession. Pages 196-211 in M. L. Cody
and J. M. Diamond, editors. Ecology and evolution of communities. Belknap Press of
Harvard University Press, Cambridge.
Huston, M. A. 1979. A general hypothesis of species diversity. American Naturalist 113:81 –
101.
Huxham, M., I. Roberts, and J. Bremner. 2000. A field test of the intermediate disturbance
hypothesis in the soft-bottom intertidal. International Review of Hydrobiology 85:379394.
Jackson, J. B. C. 1977. Competition on Marine Hard Substrata: the adaptive significance of
Solitary and colonial Strategies. American Naturalist 111:743 - 767.
Jara, V. C., J. H. S. Miyamoto, B. A. P. da Gama, M. Molis, M. Wahl, and R. C. Pereira.
2006. Limited evidence of interactive disturbance and nutrient effects on the diversity
of macrobenthic assemblages. Marine Ecology-Progress Series 308:37-48.
Jonsson, P. R., K. M. Berntsson, and A. I. Larsson. 2004. Linking larval supply to
recruitment: Flow-mediated control of initial adhesion of barnacle larvae. Ecology
85:2850-2859.
Kimmerer, R. W. and T. F. H. Allen. 1982. The Role of Disturbance in the Pattern of a
Riparian Bryophyte Community. American Midland Naturalist 107:370-383.
Kneitel, J. M. and J. M. Chase. 2004. Disturbance, predator, and resource interactions alter
container community composition. Ecology 85:2088-2093.
Kondoh, M. 2001. Unifying the relationships of species richness to productivity and
disturbance. Proceedings of the Royal Society of London Series B-Biological Sciences
268:269-271.
Lake, P. S., T. J. Doeg, and R. Marchant. 1989. Effects of Multiple Disturbance on
Macroinvertebrate Communities in the Acheron River, Victoria. Australian Journal of
Ecology 14:507-514.
Lane, P. A. 1986. Symmetry, change, perturbation, and observation mode in natural
communities. Ecology 67:223-239.
Leclerc, J. 1991. Optimal foraging strategy of the sheet-web spider Lepthyphantes flavipes
under perturbation. Ecology 72:1267-1272.
Lenz, M., M. Molis, and M. Wahl. 2004. Testing the intermediate disturbance hypothesis:
response of fouling communities to various levels of emersion intensity. Marine
Ecology-Progress Series 278:53-65.
Levin, S. A. and R. T. Paine. 1974. Disturbance, patch formation, and community structure.
Proc. Nat. Acad. Sci. Usa. 71:2744 – 2747.
Li, J., W. A. Loneragan, J. A. Duggin, and C. D. Grant. 2004. Issues affecting the
measurement of disturbance response patterns in herbaceous vegetation - A test of the
intermediate disturbance hypothesis. Plant Ecology 172:11-26.
Mackey, R. L. and D. J. Currie. 2000. A re-examination of the expected effects of disturbance
on diversity. Oikos 88:483-493.
28
J. Robin Svensson
Mackey, R. L. and D. J. Currie. 2001. The diversity-disturbance relationship: Is it generally
strong and peaked? Ecology 82:3479-3492.
McCabe, D. J. and N. J. Gotelli. 2000. Effects of disturbance frequency, intensity, and area on
assemblages of stream macroinvertebrates. Oecologia 124:270-279.
McGuinness, K. A. 1987. Disturbance And Organisms On Boulders .2. Causes Of Patterns In
Diversity And Abundance. Oecologia 71:420-430.
Menge, B. A. and J. P. Sutherland. 1987. Community Regulation - Variation in Disturbance,
Competition, and Predation in Relation to Environmental-Stress and Recruitment.
American Naturalist 130:730-757.
Miller, T. E. 1982. Community diversity and interactions between the size and frequency of
disturbance. American Naturalist 120:533 – 536.
Miyake, Y. and S. Nakano. 2002. Effects of substratum stability on diversity of stream
invertebrates during baseflow at two spatial scales. Freshwater Biology 47:219-230.
Molino, J. F. and D. Sabatier. 2001. Tree diversity in tropical rain forests: A validation of the
intermediate disturbance hypothesis. Science 294:1702-1704.
Mouritsen, K. N. and R. Poulin. 2005. Parasites boosts biodiversity and changes animal
community structure by trait-mediated indirect effects. Oikos 108:344-350.
Odum, E. P. 1963. Ecology. Holt, Rinehart and Winston, New York
Odum, E. P., J. T. Finn, and E. H. Franz. 1979. Perturbation theory and the subsidy-stress
gradient. Bioscience 29:349-352.
Onuf, C. P., J. M. Teal, and I. Valiela. 1977. Interactions of nutrients, plant growth and
herbivory in a mangrove ecosystem. Ecology 58:514-526.
Osman, R. W. 1977. The establishment and development of a marine epifaunal community.
Ecological Monographs 47:37 – 63.
Osman, R. W. and R. B. Whitlatch. 1978. SocietyPatterns of Species Diversity: Fact or
Artifact? Paleobiology 4:41-54.
Pacala, S. W. and M. Rees. 1998. Models suggesting field experiments to test two hypotheses
explaining successional diversity. American Naturalist 152:729-737.
Padisak, J. 1993. The Influence of Different Disturbance Frequencies on the Species
Richness, Diversity and Equitability of Phytoplankton in Shallow Lakes.
Hydrobiologia 249:135-156.
Paine, R. T. and S. A. Levin. 1981. Intertidal landscapes: disturbance and the dynamics of
pattern. Ecological Monographs 51:145 – 178.
Pavia, H. and E. Brock. 2000. Extrinsic factors influencing phlorotannin production in the
brown alga Ascophyllum nodosum. Marine Ecology-Progress Series 193:285-294.
Peters, R. H. 1991. A critique for ecology. Cambridge University Press, Cambridge.
Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The maintenance of species
diversity by disturbance. Quarterly Review Of Biology 64:393-418.
Pickett, S. T. A., J. Kolasa, J. J. Armesto, and S. L. Collins. 1989. The ecological concept of
disturbance and its expression at various hierarchical levels. Oikos 54:129-136.
Pickett, S. T. A. and P. S. White. 1985. The ecology of natural disturbance and patch
dynamics. Academic press, London.
Proulx, M. and A. Mazumder. 1998. Reversal of grazing impact on plant species richness in
nutrient-poor vs. nutrient-rich ecosystems. Ecology 79:2581-2592.
Rapport, D. J., H. A. Regier, and T. C. Hutchinson. 1985. Ecosystem Behavior Under Stress.
American Naturalist 125:617 – 640.
Reynolds, C. S. 1995. The Intermediate Disturbance Hypothesis and its applicability to
planktonic communities: Comments on the views of Padisak and Wilson. New
Zealand Journal of Ecology 19:219-225.
29
Ecological disturbances
Reynolds, C. S., J. Padisak, and U. Sommer. 1993. Intermediate Disturbance in the Ecology
of Phytoplankton and the Maintenance of Species-Diversity - a Synthesis.
Hydrobiologia 249:183-188.
Roughgarden, J., S. Gaines, and H. Possingham. 1988. Recruitment Dynamics in Complex
Life-Cycles. Science 241:1460-1466.
Roxburgh, S. H., K. Shea, and J. B. Wilson. 2004. The intermediate disturbance hypothesis:
Patch dynamics and mechanisms of species coexistence. Ecology 85:359-371.
Rykiel, E. J. 1985. Towards a definition of ecological disturbance. Australian Journal of
Ecology 10:361-365.
Schoener, T. W. 1972. Mathematical ecology and its place among the sciences. I. The
biological domain. Science 178:389-391.
Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The combined effects of energy and
disturbance on species richness in protist microcosms. Ecology Letters 8:730-738.
Shea, K., S. H. Roxburgh, and E. S. J. Rauschert. 2004. Moving from pattern to process:
coexistence mechanisms under intermediate disturbance regimes. Ecology Letters
7:491-508.
Sheil, D. 2001. Long-term observations of rain forest succession, tree diversity and responses
to disturbance. Plant Ecology 155:183-199.
Sheil, D. and D. Burslem. 2003. Disturbing hypotheses in tropical forests. Trends in Ecology
& Evolution 18:18-26.
Sousa, W. P. 1979a. Disturbance in marine intertidal boulder fields: the nonequilibrium
maintenance of species diversity. Ecology 60:1225 – 1239.
Sousa, W. P. 1979b. Experimental investigations of disturbance and ecological succession in
a rocky intertidal algal community. Ecological Monographs 49:227-254.
Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of
Ecology and Systematics 15:353 – 391.
Sousa, W. P. 2001. Natural disturbance and the dynamics of marine benthic communities.
Sinauer, Sunderland, Massachusetts, USA.
Steneck, R. S., S. D. Hacker, and M. N. Dethier. 1991. Mechanisms of Competitive
Dominance between Crustose Coralline Algae - an Herbivore-Mediated Competitive
Reversal. Ecology 72:938-950.
Sugden, H., M. Lenz, M. Molis, M. Wahl, and J. C. Thomason. 2008. The interaction between
nutrient availability and disturbance frequency on the diversity of benthic marine
communities on the north-east coast of England. Journal of Animal Ecology 77:24-31.
Sugden, H., R. Panusch, M. Lenz, M. Wahl, and J. C. Thomason. 2007. Temporal variability
of disturbances: is this important for diversity and structure of marine fouling
assemblages? Marine Ecology-an Evolutionary Perspective 28:368-376.
Szentkiralyi, F. and F. Kozar. 1991. How many species are there in apple insect communities
– testing the resource diversity and intermediate disturbance hypotheses. Ecological
Entomology 16:491-503.
Talbot, F. H., B. C. Russel, and G. R. V. Anderson. 1978. Coral reef fish communities:
unstable, high diversity systems? . Ecological Monographs 48:425 – 440.
Tilman, D. 1980. Resources, a graphical-mechanistic to competition and predation. American
Naturalist 116:362-393.
Tuck, I. D., S. J. Hall, M. R. Robertson, E. Armstrong, and D. J. Basford. 1998. Effects of
physical trawling disturbance in a previously unfished sheltered Scottish sea loch.
Marine Ecology-Progress Series 162:227-242.
Turkington, R., E. Klein, and C. P. Chanway. 1993. Interactive Effects of Nutrients and
Disturbance - an Experimental Test of Plant Strategy Theory. Ecology 74:863-878.
30
J. Robin Svensson
Tuyttens, F. A. M., R. J. Delahay, D. W. MacDonald, C. L. Cheeseman, B. Long, and C. A.
Donnelly. 2000. Spatial perturbation caused by a badger (Meles meles) culling
operation: implications for the function of territoriality and the control of bovine
tuberculosis (Mycobacterium bovis). Journal of Animal Ecology 69:815-828.
Underwood, A. J. and M. J. Anderson. 1994. Seasonal and Temporal Aspects of Recruitment
and Succession in an Intertidal Estuarine Fouling Assemblage. Journal of the Marine
Biological Association of the United Kingdom 74:563-584.
Valdivia, N., A. Heidemann, M. Thiel, M. Molis, and M. Wahl. 2005. Effects of disturbance
on the diversity of hard-bottom macrobenthic communities on the coast of Chile.
Marine Ecology-Progress Series 299:45-54.
Webster, J. R. and B. C. Patten. 1979. Effects of Watershed Perturbation on Stream Potassium
and Calcium Dynamics. Ecological Monographs 49:51-72.
Vetaas, O. R. 1997. The effect of canopy disturbance on species richness in a central
Himalayan oak forest. Plant Ecology 132:29-38.
Widdicombe, S. and M. C. Austen. 1999. Mesocosm investigation into the effects of
bioturbation on the diversity and structure of a subtidal macrobenthic community.
Marine Ecology-Progress Series 189:181-193.
Widdicombe, S. and M. C. Austen. 2001. The interaction between physical disturbance and
organic enrichment: An important element in structuring benthic communities.
Limnology And Oceanography 46:1720-1733.
Wilkinson, D. M. 1999. The disturbing history of intermediate disturbance. Oikos 84:145147.
Willby, N. J., J. R. Pygott, and J. W. Eaton. 2001. Inter-relationships between standing crop,
biodiversity and trait attributes of hydrophytic vegetation in artificial waterways.
Freshwater Biology 46:883-902.
Wilson, J. B. 1990. Mechanisms for species co-existence: twelve explanations for
Hutchinton’s ‘paradox of the plankton’: evidence from New Zealand plant
communities. N.Z. J. Ecol. 13 17 – 42.
Wilson, J. B. 1994. The Intermediate Disturbance Hypothesis of Species Coexistence Is
Based on Patch Dynamics. New Zealand Journal of Ecology 18:176-181.
Wilson, S. D. and D. Tilman. 2002. Quadratic variation in old-field species richness along
gradients of disturbance and nitrogen. Ecology 83:492-504.
Wootton, J. T. 1998. Effects of disturbance on species diversity: A multitrophic perspective.
American Naturalist 152:803-825.
Worm, B., H. K. Lotze, H. Hillebrand, and U. Sommer. 2002. Consumer versus resource
control of species diversity and ecosystem functioning. Nature 417:848-851.
31
Ecological disturbances
Acknowledgements
Egentligen skulle man kanske, nu natten innan tryckning, köra en Kjell Bergquist här; få allt
snabbt överstökat genom att tacka sig själv och sen helt sonika dra, men detta skulle givetvis
vara särdeles osant. Något som dessutom tar lite emot att erkänna är väl att två stycken
ohängda GAIS:are haft en stor del i att allting inte blivit helt värdelöst. Detta trots initierandet
av ett mått på en tillsynes kvantifierbar tidsrymd kallad ”nästa vecka”, där saker som utlovas
hända inom denna tidsrymd, namnet till trots, aldrig inträffar veckan efter det yttrades. Mina
kunskaper inom kvantfysik, parallella universa och strängteori är väldigt begränsade, men jag
antar att det är någon ytterligare dimension här som jag ännu inte fått grepp om. Ett annat bra
knep, som utövas i den icke-omänskliga tidsrymden, har beskrivits av förbipasserande
finlandssvenskar som ”Vad fan gör ni på era möten egentligen? Ni sitter helt tysta och stirrar
förvirrat upp i taket varje gång man går förbi!”. Förutom att detta är ett tydligt tecken på total
inkompetens från alla inblandade, har det även lett till storslagna vetenskapliga genombrott
som av internationellt erkända forskare officiellt benämnts som ”a fundamental flaw in the
authors logic” och ”the conclusions of the work are not supported by the data”. Trots att 50 %
av handledarna föreslog kollektivt självmord som respons på kritiken, röstades detta förslag
(uppenbarligen) ned av hela 67 % av de mer livslustsfyllda medförfattarna. Mer konkreta
framsteg är kunskapsöverflyttning från Nycklebybor till Majornabor inom statistikens
svårbegripliga värld, där ofattliga akronymer (CAP, SNK, MDS, anova, permanova) lärts ut
med ett leende på GAIS-läpparna, citerandes en mer filosofisk approach som i stora drag går
ut på att ’tortera datan med olika analyser tills den erkänner’. Det hemliga epitetet ’HH’, som
visat sig vara helt oförknippat med 30-talets Centraleuropa, har även lömskt, genom att tvinga
fram oändligt många versioner av varje manus, lärt mej tekniken för gränslös episkhet
oberoende av data. Lite som när Daniel Larusso vaxade Kesuke Miyagi’s bilar, svärandes
samt ovetandes om tragglandets potentiella storhet. Framåt slutet i en tillsynes oändlig GAISdimma har det även funnits en Åtvidabergiansk fyr, i vars ljus planer smitts, modeller
uppkommit, hypoteser utbenats och elektronisk post besvarats, smått chockerande, redan
samma dag. Slutklämligen, utan er hade detta aldrig gått.
Flertalet kemiekologiska G-människor har funnits vid min sida där stöttande aktiviteter inte
bara inneburit balkongvistande rusdrycksinmundigande, utan även nakenbadande,
trolldegsskapande, kemisk analys-assistans, vågmaskinssnickrande, hälsovådliga
undervattensaktiviteter, nikotin-snikande, tröstlöst trålletande, fältmässigt Jägermeister
shottande och klippstrands överblickande butterkaksätande. Till detta gäng hör även min
kontorssambo som inte bara guidat mej genom hela Honshu och Hokkaido, utan även
förklarar för mej obegripliga saker som projektuppföljning i datalagret och elektronisk
fakturahantering. I Tjärnös begynnelse fanns smålänningar och dalmasar, men även nollåttor,
Disney-karaktärer och finnar av båda kön. Skåningar, som det finns alldeles för många av, har
lyckligtvis befunnits på behörigt avstånd, men som trots detta, och gärna i kombination med
en viss smålänning, ständigt lyckas leta upp en och håna ens fiskekunskaper. There has also
been a sensei in the underappreciated art of sandwich making, who relentlessly remind me
that things are rarely as good as they seem, and his beloved wife who initially adopted me as a
second boyfriend, only to leave me heartbroken for Sverker’s southern regions. Eftersom det
finns ett oändligt antal oidentifierbara kräk och slemmiga växter i havsdjupen har jag varit
helt beroende de barmhärtiga samariterna Elisabet, Anneli, Fredrik P och Hans-G. I stark
kontrast till mentala prövningar finns en slagsmålsklubb som äger rum på Tjärnö skola varje
onsdag, där revben brutits, blåa ögon mottagits och utdelats av Per B, Erik B, Swantje, Greg,
Fidde, Micke, Lars, Finn, Petri, Tuuli, Ankan, Erkan, Göran, Gunnar, Henke, Mats, Martin G,
Martin S, Malin, Piff, Puff, Erika, Andreas, Anders, Geno, Josefin, Carl-Johan, Johanna,
32
J. Robin Svensson
Johan E, Johan H, Johan R, Johan W, Hanna H, Hanna S, Eva-Lotta, Mia, Filip, Stina,
Rickard Lasse Pereyrasson och många, många fler.
I slutet på flertalet välbehövliga flyktförsök från en öde ö fanns ett stadsdelsstort högkvarter
idogt bevakat av MB, Åsmund, DAF, Pejlert, Mr. Däjvid, Greken samt även Ryssen och
Dödskristian, där i rusdryckernas glada brödraskap energi återskapats, smärta och glädje
delats och livslånga minnen bildats. Nästgårds har jag även funnit landets vänaste prinsessa
som, förutom av mej påtvingad korrekturläsning, håller mej på topp genom överraskningsbrottning likt Kato i Clouseaus kylskåp, och som dragit upp mej ur träsket och visat mej en
värld fylld av ’hummer och rosa champagne’. Till sist min familj som funnits där, inte bara
under doktorand-perioden, utan alltid.
33
Paper I
Paper II
PAPER
Paper
III I
Paper IV
Paper V
“I have opinions of my own, strong opinions, but I don't
always agree with them.”
- George H. W. Bush
Paper VI
Ecology, 88(4), 2007, pp. 830–838
Ó 2007 by the Ecological Society of America
MAXIMUM SPECIES RICHNESS AT INTERMEDIATE FREQUENCIES OF
DISTURBANCE: CONSISTENCY AMONG LEVELS OF PRODUCTIVITY
J. ROBIN SVENSSON,1,5 MATS LINDEGARTH,1 MICHAEL SICCHA,2 MARK LENZ,3 MARKUS MOLIS,4 MARTIN WAHL,3
1
AND HENRIK PAVIA
1
Department of Marine Ecology, Göteborg University, Tjärnö Marine Biological Laboratory, 452 96 Strömstad, Sweden
2
Institute for Geological Science, Eberhard Karls University Tübingen Sigwartstr.10, 72076 Tübingen, Germany
3
Leibniz-Institute for Marine Science, Düsternbrooker Weg 20, 24105 Kiel, Germany
4
Biologische Anstalt Helgoland, Alfred Wegener Institute for Polar and Marine Research, Marine Station,
Kurpromenade 201, 27498 Helgoland, Germany
Abstract. Development of a mechanistic understanding and predictions of patterns of
biodiversity is a central theme in ecology. One of the most influential theories, the intermediate
disturbance hypothesis (IDH), predicts maximum diversity at intermediate levels of
disturbance frequency. The dynamic equilibrium model (DEM), an extension of the IDH,
predicts that the level of productivity determines at what frequency of disturbance maximum
diversity occurs. To test, and contrast, the predictions of these two models, a field experiment
on marine hard-substratum assemblages was conducted with seven levels of disturbance
frequency and three levels of nutrient availability. Consistent with the IDH, maximum
diversity, measured as species richness, was observed at an intermediate frequency of
disturbance. Despite documented effects on productivity, the relationship between disturbance
and diversity was not altered by the nutrient treatments. Thus, in this system the DEM did not
improve the understanding of patterns of diversity compared to the IDH. Furthermore, it is
suggested that careful consideration of measurements and practical definitions of productivity
in natural assemblages is necessary for a rigorous test of the DEM.
Key words: competitive exclusion; disturbance; productivity; species richness.
INTRODUCTION
Spatial and temporal patterns of diversity in natural
communities are central themes in classical natural
history as well as in contemporary theoretical ecology
(e.g., Huston 1994, Hubbell 2001). Throughout history
the magnitude of existing biological diversity and its
heterogeneous distribution have continuously challenged ecologists to develop and test models to explain
patterns at a multitude of temporal and spatial scales,
using increasingly more complex models (e.g., Connell
1978, Huston 1994, Hubbell 2001). Some of these
models have been based on biological interactions
(e.g., Miller 1958, Fischer 1960, Paine 1966, Paine and
Vadas 1969, Menge and Sutherland 1987), while others
have primarily focused on abiotic processes (e.g.,
Hutchinson 1961, Levin and Paine 1974, Connell 1978,
Paine and Levin 1981).
Many of these ideas rely on disturbances to disrupt
the effects of biological interactions, such as competitive
exclusion, on diversity. A variety of abiotic (e.g., fire,
wind, wave action, and drifting logs) and biotic factors
(e.g., grazing, predation, and trampling) may act as
agents of disturbance, depending on the specific
Manuscript received 8 June 2006; revised 25 September 2006;
accepted 29 September 2006. Corresponding Editor: S. G.
Morgan.
5 E-mail: [email protected]
830
properties of the particular ecological system. There is
also a range of definitions of what constitutes an actual
disturbance. Grime (1977) defined disturbance as partial
or total destruction of biomass. Sousa (1984) extended
this definition by adding that disturbance also creates
opportunities for new individuals to become established.
Pickett and White (1985) have a broader definition
where disturbance is ‘‘. . . any relative discrete event in
time that disrupts ecosystems, community, or population structure and changes resources, substrate availability, or the physical environment.’’ Thus, despite
some ambiguity in the definition of the concept of
disturbance, it has direct effects on vital rates and
population dynamics and it is therefore a potentially
useful generalization.
One important conceptual formulation of the effects
of natural disturbances on diversity is the intermediate
disturbance hypothesis, IDH (Connell 1978). The IDH
predicts that diversity will be large at intermediate rates
of disturbance and smaller at higher and lower rates of
disturbance. The rationale for this idea is that at low
rates of disturbance strong competitors exclude competitively inferior species and communities are dominated by a few species. Intermediate rates of disturbance,
however, disrupt competitive hierarchies by increasing
rates of mortality and thus making free space available
for recruitment of competitively inferior species. At
successively higher rates of disturbance, recruitment
April 2007
DISTURBANCE, PRODUCTIVITY, AND DIVERSITY
cannot balance the high rates of mortality, and slowrecruiting species disappear from the community.
Findings consistent with the predictions of the IDH
have been made in manipulative studies in both
terrestrial (e.g., Molino and Sabatier 2001, Anderson
et al. 2005) and marine (e.g., Osman 1977, Sousa 1979,
Valdivia et al. 2005, Patricio et al. 2006) ecosystems.
However, contradictory observations have also been
made (Lake et al. 1989, Collins et al. 1995, Gutt and
Piepenburg 2003), and due to difficulties of incorporating all components of natural environments, laboratory
studies are often relatively less supportive (Cowie et al.
2000). In summary, the IDH has been an influential
concept in research and also as a tool in management of
nature reserves (Wootton 1998).
In response to observations that did not appear
consistent with the IDH, Huston (1979) suggested that
the relationship between disturbance and diversity is
modified by the level of productivity. Using a dynamic
equilibrium model (DEM), Huston (1979, later elaborated by Kondoh 2001) suggested that increased
productivity, and thus growth rates of individuals and
populations, means that a more severe disturbance is
required to prevent competitive exclusion. As a consequence, maximum diversity is observed at lower
intensities of disturbance when productivity is low,
compared to when productivity is high. The shape of the
relationship between disturbance and diversity may
therefore be of three general types: monotonically
decreasing (at low productivity), unimodal (when
productivity is intermediate), and monotonically increasing (when productivity is high). These three types
of relationships have been observed in various habitats
(e.g., Mackey and Currie 2001), but evidence from
explicit manipulative studies demonstrating the interactive effects of disturbance and productivity is scarce
(Rashit and Bazin 1987, Widdicombe and Austen 2001).
One pioneering test in marine rocky environments is the
study by Worm et al. (2002), who observed interactive
effects of nutrient enrichment and disturbance (grazing
by mesoherbivores) on algal diversity, which they found
consistent with those predicted by the DEM.
The development from a simple general model
involving only one factor, into a more complex and
detailed model involving multiple factors, may represent
important conceptual progress within a field of research
(e.g., Hilborn and Mangel 1997, Underwood 1997). The
benefit of a more complex model is that it may be used
to accurately predict a more diverse set of conditions
with little discrepancy due to approximation (Zucchini
2000). There are, however, no guarantees that a complex
model is more powerful than a simple one (e.g., Zucchini
2000, Ginzburg and Jensen 2004). This is because a
complex model has a greater uncertainty, as it requires
more parameters to be estimated. Thus, in terms of
predictive power, the utility of a complex model relies on
whether the reduction of error due to approximation is
larger than the increase in error due to estimation.
831
Indeed, from observational data it appears that the great
range of observed responses of diversity to disturbance
(Mackey and Currie 2001) can potentially be more
accurately represented if productivity is included (Huston 1979). Whether this really is the case in a wide range
of ecological systems remains to be tested in manipulative experiments.
In this study we contrast predictions from the IDH to
those of the DEM in a marine hard-substratum
community. Physical disturbance and nutrient availability were manipulated in subtidal communities in the
field, with seven distinct frequencies of disturbance and
three levels of nutrient availability. Manipulative studies
on epibenthic assemblages have made important contributions to the development and testing of general
ecological models (e.g., Paine 1966, Dayton 1971,
Lubchenco and Menge 1978, Sousa 1979). Due to their
potential for quick recovery, epibenthic assemblages
have proven particularly useful for investigating disturbance–diversity patterns over ecologically relevant time
scales in manipulative studies (e.g., Worm et al. 2002,
Bertocci et al. 2005, Jara et al. 2006).
MATERIALS
AND
METHODS
Study site
The field experiment was conducted in the vicinity of
Tjärnö Marine Biological Laboratory on the west coast
of Sweden. The experimental sites were two bays located
;1 km apart (58852.92 0 N, 1188.31 0 E and 58852.17 0 N;
1188.82 0 E for sites 1 and 2, respectively). Site 1 has an
average depth of 8 m and is surrounded by muddy and
rocky shores. The surrounding cliffs were covered with
red, green, and brown macroalgae as well as mussels and
tunicates. Site 2 has an average depth of 6 m and is
surrounded by sandy beaches and boulder fields. Site 2
also has an extensive Zostera meadow and the boulders
were commonly overgrown by fucoids, barnacles, and
mussels. The grazers in this system are exclusively socalled mesoherbivores, such as amphipods, isopods, and
littorinid gastropods (Pavia et al. 1999, Wikstrom et al.
2006). Gastropods were effectively excluded from
reaching the panels due to the positioning and construction of the experimental units (see Experimental design),
and because of the low abundance of crustacean
mesoherbivores in the vicinity of the experimental units,
possible effects of grazing are not likely to have affected
the results of this study. The waters off the Swedish west
coast are generally low in nutrients during the summer
months (Nilsson 1991), and nutrients therefore become
a limiting resource in this system (Soderstrom 1996).
Experimental design
Mooring units, made from 2100 3 250 3 4 mm
polyvinyl chloride (PVC) strips bent into a ring, were
hung from a buoy ;0.5 m below the water surface. In
this way, benthic consumers were excluded from the
setup. The rings were deployed on 1 March to allow
settling and establishment of communities before the
832
J. ROBIN SVENSSON ET AL.
experimental manipulation started on 12 May. The
experimental manipulation had a duration of 24 weeks
and was terminated on 27 October 2004.
On each ring 10 PVC panels (150 3 150 3 3 mm),
roughened with emery paper, were attached with cable
ties. The panels were randomly allocated to combinations of seven disturbance levels and three nutrient
levels. Disturbance treatments consisted of a manual
removal of biomass from two randomly selected
nonoverlapping areas, each covering 10% of the panel
area, at each disturbance event. The scraping not only
kills or damages individuals, but also facilitates recruitment by the freed substratum, and the disturbance is
therefore coherent with the definition by Sousa (1984).
This disturbance was applied at six different frequencies:
every second, fourth, sixth, eighth, 10th, and 12th week
(treatments D1–D6), or left undisturbed (treatment D0).
Treatments D0–D6 were present in all rings, with two
replicates of D0 on each ring, and the remaining two
panels were randomly assigned disturbance treatments
to allow additional replication within rings.
One of three different levels of nutrient enrichment
was applied to each ring by attaching 10 fertilizer bags
(1-mm mesh) among the panels. For the highest level of
enrichment (Nþþ), bags were filled with 100 g of
fertilizer; for the moderately enriched level (Nþ), bags
were filled with gravel and 50 g fertilizer, and bags with
ambient nutrient concentration (N0) were filled only
with gravel. The slow-release Plantacote Depot 6-M,
(5.7% NO3, 8.3% NH4, 9% P2O5, and 15% K2O;
Aglukon, Düsseldorf, Germany) was used as fertilizer
due to its steady release rate in relation to mass, where a
doubling in mass leads to twice the amount of nutrients
being released (Worm et al. 2000). Each level of nutrient
availability was replicated on four randomly assigned
rings. All bags were placed inside the rings at the start of
the experiment and changed every fourth week in order
to have constant nutrient release throughout the
experiment.
Sampling
Sampling of abundance of each species and composition of the experimental communities was done before
the start of the manipulation and thereafter every eighth
week until the termination of the experiment. Data on
undisturbed communities obtained from the sampling
after eight weeks were used for testing effects of nutrient
availability on algal cover. The time of sampling was
selected to be early in the growth season to minimize
confounding influences of competition. Data from the
last sampling after 24 weeks were used for the main
analyses, i.e., the tests of the IDH and the DEM, and
data on undisturbed communities from all sampling
events were used for studying changes in the communities over time. Panels were detached and brought into
the laboratory submerged in seawater, kept under
running seawater in the laboratory during the entire
sampling procedure, and brought back into the field
Ecology, Vol. 88, No. 4
within 16 hours of each sampling event. Before
sampling, the back side and edges of all panels were
scraped clean and their wet mass was measured. The
percentage cover of bare space and sessile species was
then estimated in 5% intervals using a 15 3 15 cm plastic
grid (mesh size 5 cm2). A 1-cm margin to all edges of the
panels was not assessed, and the percentage cover of
species with a small holdfast and wide thallus was
estimated from the two-dimensional projection of the
organism on the panel. Sessile epibionts were also
accounted for. Thus, total cover was allowed to exceed
100%.
Statistical analyses
The data on species richness were analyzed with
analysis of variance (ANOVA) using Statistica 6.0
(Statsoft Incorporated, Tulsa, Oklahoma, USA). The
models were tested, with species richness as a measure of
diversity, following the elaboration of the DEM by
Kondoh (2001). Hypotheses about effects of main
factors and interactions were tested using the following
general linear model:
Xijklm ¼ l þ Si þ Nj þ SNij þ Dk þ SDik þ NDjk þ SNDijk
þ RðSNÞlðijÞ þ DRðSNÞklðijÞ þ eijklm
where l is the overall mean, site (Si) is a random factor
with two levels, nutrient enrichment (Nj) and disturbance frequency (Dk) are fixed factors with three and
seven levels respectively, ring (R[SN]l(ij)) is a nested
random factor with four levels, and eijklm is a random
deviation. Due to loss of one ring and lack of complete
replication of all levels of disturbance on each ring, type
III sums of squares was used for estimation (Henderson
1953). The residual was estimated from the variability
between undisturbed panels and from the additional
replicated treatments within each ring. To optimize
statistical power of tests, post hoc elimination and
pooling of negligible variance components (i.e., if P .
0.25) were performed (Winer et al. 1991, Underwood
1997).
Support for either of the two models, IDH or DEM, is
provided by two different terms in the linear model. The
IDH is supported if there is a significant effect of
disturbance and if the relationship between richness and
disturbance is unimodal with an optimum at intermediate levels of disturbance. This is equivalent to the
presence of a significant quadratic component in a
polynomial regression. In contrast the DEM is supported by a significant interaction between disturbance and
nutrient enrichment. The predictions of the DEM then
need to be further evaluated using polynomial regression
within individual levels of nutrient enrichment.
A fundamental premise for any experimental support
for the DEM is that the nutrient treatments actually
cause an increased primary productivity. In order to
detect effects on productivity as a consequence of the
nutrient treatment, differences in cover of macroalgae
April 2007
DISTURBANCE, PRODUCTIVITY, AND DIVERSITY
833
TABLE 1. Abundance (mean percent cover 6 SE) of sessile invertebrate and algal species present in the experimental communities
from both sites after 24 weeks, averaged over nutrient treatment for all levels of disturbance (D0–D6).
Taxon
D0
D1
D2
D3
D4
D5
D6
0
0
0.17 6 0.14
0
Chlorophyceae
Ulva intestinalis
Ulva lactuca
0.09 6 0.04 0.10 6 0.06 0.10 6 0.06 0.39 6 0.19 0.06 6 0.04
0
0
0
0.04 6 0.04
0
Phaeophyceae
Ectocarpus siliculosus
0.45 6 0.26 0.03 6 0.03 0.59 6 0.29 0.11 6 0.06 0.11 6 0.05 0.18 6 0.16 0.20 6 0.15
Rhodophyceae
Bonnemaisonia hamifera
Ceramium rubrum
Ceramium strictum
Dasya baillouviana
Osmundea truncata
Polysiphonia fucoides
Polysiphonia urceolata
Spermothamnion repens
0
1.98 6 0.57 2.62
0.07 6 0.04 0.07
0.04 6 0.03 0.03
0
0.03
0.83 6 0.49 0.17
0
0.20 6 0.12 0.07
Porifera
Leucosolenia botryoides
0.78 6 0.28 1.28 6 0.50 1.10 6 0.43 1.21 6 0.53 1.40 6 0.49 1.36 6 0.43 0.74 6 0.24
Cnidaria
Clytia hemispherica
Laomedea flexuosa
Metridium senile
Sargatiogeton undatus
0
12.8 6 2.30 18.9
0.11 6 0.05 0.14
0.07 6 0.04 0.14
Annelida
Pomatoceros triqueter
0.11 6 0.05 0.21 6 0.08 0.38 6 0.09 0.54 6 0.10 0.51 6 0.09 0.61 6 0.08 0.49 6 0.09
Crustacea
Balanus crenatus
0
Mollusca
Mytilus edulis
Podesmus sp.
0
6
6
6
6
6
0
6
0
4.21 6 1.06 3.39
0.03 6 0.03 0.07
0.03 6 0.03 0.04
0
0.38 6 0.18 0.61
0.03 6 0.03 0.21
0.05 0.03 6 0.03 0.25
0.53
0.05
0.03
0.03
0.07
0
6
6
6
0
6
6
6
0
0
0.04 6
6 3.09 19.7 6 3.07 23.9 6
6 0.06 0.31 6 0.18 0.39 6
6 0.06 0.03 6 0.03 0.18 6
0.03 6 0.03
0
0
0
1.07 2.23 6 0.55
0.05 0.03 6 0.03
0.04 0.03 6 0.03
0.06 6 0.04
0.36 1.46 6 1.15
0.18
0
0.18 0.23 6 0.15
0.04
0.64
0.14
0.04
6
6
6
6
0
0.21 6
0
0.38 6
0.03
0
0.17 2.09 6 1.07
0.06 0.03 6 0.03
0.03 0.17 6 0.14
0.03 6 0.03
0.07 0.40 6 0.29
0
0.17 0.09 6 0.05
0.04
0
0
0
4.30 19.9 6 3.04 30.7 6 3.44 24.4 6 2.44
0.36 0.14 6 0.06 0.14 6 0.06 0.37 6 0.20
0.07 0.11 6 0.05 0.14 6 0.06 0.06 6 0.04
0.06 6 0.04
0
0
0.35 6 0.12 0.52 6 0.18 0.59 6 0.24 0.64 6 0.25 0.63 6 0.24 1.00 6 0.37 0.57 6 0.29
0
0.17 6 0.17
0
0
0
0.04 6 0.03
0
Bryozoa
Cryptosula pallasiana
0.39 6 0.19 0.28 6 0.18 0.03 6 0.03 0.04 6 0.04 0.23 6 0.15 0.29 6 0.16 0.06 6 0.04
Electra pilosa
0.91 6 0.29 1.34 6 0.46 0.76 6 0.28 0.93 6 0.33 0.46 6 0.20 1.36 6 0.34 0.49 6 0.20
Membranipora membranacea 0.33 6 0.19 0.03 6 0.03
0
0.71 6 0.42 0.14 6 0.14 0.18 6 0.16
0
Hemichordata
Ascidiella aspersa
Botryllus schlosseri
Botrylloides leachi
Ciona intestinalis
11.9 6 2.38 12.3
0
0
84.0 6 3.50 75.3
6 2.74 12.5 6 2.27 8.32 6 2.13 8.94 6 1.81 8.82 6 1.60 5.03 6 1.17
0
0
0.04 6 0.04
0
0
0
0
0.17 6 0.17
0
0
0
0
6 5.15 64.3 6 5.17 71.4 6 4.82 68.0 6 4.69 53.6 6 4.04 17.3 6 2.25
among levels of enrichment were tested using undisturbed panels (D0) after eight weeks. Data were
analyzed using ANOVA:
Xijk ¼ l þ Si þ Nj þ SNij þ RðSNÞkðijÞ þ eijk
RESULTS
General observations
During the experiment a total of 15 species of algae
and 17 species of sessile invertebrates were observed.
The most abundant organisms, occupying large areas of
the panels, were the tunicates Ciona intestinalis and
Ascidiella aspersa and the hydroid Laomedea flexuosa.
At the end of the experiment, ephemeral algae,
bryozoans, and sea anemones were frequent in the
communities, although usually low in cover (Table 1).
Studies of the development of undisturbed communities
showed that richness was highest after 8 weeks at site 1
and after 16 weeks at site 2 (Fig. 1A). The decrease in
richness at later stages suggests that some species were
excluded as a result of competition. This is consistent
with the observation of an earlier peak in richness at site
1, following the establishment of a dense cover of C.
intestinalis at this site (Fig. 1B). The ascidians occupied
.95% of the space on control panels after 24 weeks at
site 1, suggesting that C. intestinalis is a competitive
dominant in this system, capable of excluding both other
invertebrates as well as most species of macroalgae (Fig.
1B).
Assessment of productivity
The analysis of algal cover in undisturbed communities after 8 weeks showed that there was a statistically
significant response to increased nutrient availability
(F2,44 ¼ 10.74, P , 0.001). Inspection of means (mean
[6SE] cover of algae for N0, Nþ, and Nþþ were 54.5 6
834
J. ROBIN SVENSSON ET AL.
Ecology, Vol. 88, No. 4
Testing predictions from the IDH and DEM
FIG. 1. Temporal patterns of (A) species richness and (B)
percent cover of C. intestinalis in fouling communities at sites 1
and 2. Data are presented as mean 6 SE.
5.13%, 82.5 6 4.41%, and 71.1 6 3.55%, respectively)
and the SNK test revealed that there were significant
differences between unfertilized panels (N0) and those
fertilized (Nþ and Nþþ). Furthermore, there was no
significant interaction term between the two factors, site
(S) and nutrient enrichment (N) (F2,42 ¼ 1.45, P ¼ 0.25),
suggesting that nutrient availability had a general effect
on productivity and that useful tests of the DEM were in
fact possible. However, no significant difference in algal
cover was observed between Nþ and Nþþ, which could be
due to a saturation of nutrients already at the Nþ level.
Analysis of species richness at the end of the
experiment showed that there was a significant effect
of disturbance, but no interactive effect of disturbance
and nutrients (Table 2a). In all levels of nutrients, there
was a tendency for maximum richness at intermediate
levels of disturbance (Fig. 2). Initially it might appear
that maximum diversity occurred at different levels of
disturbance, but the variability among and within levels
of disturbance was large and the predicted shift toward
more frequent disturbances was not observed (maximum
richness was observed at D5, D5, and D2 for N0, Nþ,
and Nþþ, respectively). Considering the fact that the
hypothesis about simple effects of disturbance and that
of interactive effects involving disturbance and nutrients
were both tested using the same pooled mean square as
the error term (with 189 df ), conclusions about lack of
interactive effects appear robust and not caused by a
lack of statistical power. This view is supported by
calculation of effect-sizes from estimated mean squares,
which reveal that the effect of disturbance was ;20
times larger than that of the interaction (kD2 ¼ 1.82 and
k2N3D ¼ 0.10). There was no significant interaction
involving disturbance and any of the spatial scales, i.e.,
sites and rings (Table 2a). This indicates that the effect
of disturbance was consistent among places. Nevertheless, significant variability among rings indicates that
there was small-scale variability in richness within sites.
Further analysis showed that, not only were there
differences among levels of disturbance, but there was
also a significant quadratic component in the polynomial regression (Table 2b), i.e., maximum richness at
intermediate disturbances (Fig. 3A). Consistent with the
IDH, these results suggest that sessile species are
removed at low and high frequencies of disturbance.
Inspection of the mean cover of the most abundant taxa
suggests that they differ in their responses to disturbance
TABLE 2. (a) ANOVA on species richness at the end of the experiment and (b) regression analysis.
Source
df
MS
F
P
Error term
a) ANOVA on species richness
Site, S
Nutrients, N
Disturbance, D
S3N
S3D
N3D
Ring, R(S 3 N)
S3N3D
D 3 R(S 3 N)
Residual
Pooled
1
2
6
2
6
12
17
12
102
69
189
7.47
2.66
7.98
3.88
3.01
2.58
7.90
0.90
2.75
2.44
2.53
0.94
0.69
3.16
0.49
1.10
1.03
3.24
0.33
1.13
0.34
0.59
0.01
0.62
0.37
0.43
0.00
0.98
0.30
R(S 3 N)
S3N
pooled
R(S 3 N)
D 3 R(S 3 N)
pooled
residual
D 3 R(S 3 N)
residual
2
4
0.44
0.04
10.04
0.03
b) Regression analysis Regression
Residual
R2
0.83
Notes: (a) Hypotheses about effects of disturbance (consistent with predictions from IDH) and interactions between disturbance
and nutrients (consistent with predictions from DEM) were tested using a pooled error term following nonsignificant tests (P .
0.25) of D 3 R(S 3 N), S 3 N 3 D, and S 3 D. (b) Regression analysis for effects of disturbance on species richness.
Coefficients for the regression analysis are as follows. For the intercept, b ¼ 5.25, t ¼ 29.12, P ¼ 0.00; for D (disturbance), b ¼
3.73, t ¼ 3.96, P ¼ 0.02; for D2, b ¼ 3.77, t ¼ 4.39, P ¼ 0.01.
April 2007
DISTURBANCE, PRODUCTIVITY, AND DIVERSITY
FIG. 2. Effects of disturbance on species richness at
different nutrient levels (see Materials and Methods: Experimental design). Data are presented as mean 6 SE.
(Fig. 3B). Thus there is a strong negative effect on the
cover of the dominant tunicates Ciona intestinalis and
Ascidiella aspersa, while a rapid colonizer such as the
hydroid Laomedea flexuosa is positively affected by
disturbance.
DISCUSSION
In this study we found empirical evidence supporting
the IDH, but not the DEM. Species richness was highest
at an intermediate frequency of disturbance, and this
pattern was not significantly affected by different levels
of nutrient enrichment. This was in spite of the fact that
the nutrient treatment had a significant effect increasing
percentage cover of macroalgae, which is closely linked
to productivity (Death 2002). In contrast to the IDH,
the empirical support for the DEM is scarce. So far
support has come from observational studies of flooding
in riparian wetlands (Pollock et al. 1998), a mesocosm
study of sediment movement and organic enrichment in
deep-sea benthos (Widdicombe and Austen 2001),
laboratory experiments of energy availability and
mortality in microcosms (Rashit and Bazin 1987), and
in the only two experiments that have manipulated
disturbance and productivity simultaneously in the field
(Worm et al. 2002, Jara et al. 2006). The conclusions
from our experiment differ from the few previous studies
testing the DEM. Because productivity was manipulated
using the same procedures as in Worm et al. (2002) and
Jara et al. (2006), the nutrient treatment cannot explain
the different results. Instead, it is more likely that the
divergent outcomes were caused by differences in (1) the
composition of the experimental communities, and/or
(2) the way the communities were disturbed.
The communities in this study were not only rich in
species, but also in terms of higher taxa and functional
groups. During the experiment .30 different species
were observed in the communities, 15 species of macroalgae and 17 species from such different taxonomic
groups as tunicates, mussels, hydroids, bryozoans,
barnacles, annelids, and sea anemones. Other experi-
835
ments on the DEM have used more restricted taxon
sampling and studied communities composed mainly of
algae (Worm et al. 2002), polychaetes (Widdicombe and
Austen 2001), protist bacterivores (Scholes et al. 2005),
and bacteria, flagellates, and ciliates (Rashit and Bazin
1987). Experiments conducted in more diverse systems
can be advantageous due to the possibility of recognizing patterns among more distantly related taxa. In this
experiment tunicates occupied most of the space on
control panels, and were thus capable of excluding a
variety of both invertebrate and macroalgal species. Had
the hypotheses been tested in assemblages of solely
macroalgae or invertebrates, this dominance of one
taxon over several taxa from distant groups might not
have been revealed, and patterns among, for instance,
only macroalgae (cf. Worm et al. 2002) might have been
different and not representative for the natural diversity
of hard-substratum assemblages of temperate marine
waters. Because the DEM and the IDH are general
ecological models intended to explain gradients of
diversity in nature, their generality and explanatory
power should be assessed using natural communities.
The diversity and composition of communities can
influence the outcome of an experiment, because
different species and functional groups respond differently to experimental treatments. It is therefore impor-
FIG. 3. (A) Species richness on the experimental panels, and
(B) percent cover of Ascidiella aspersa, Laomedea flexuosa, and
Ciona intestinalis, as functions of relative disturbance frequency
(see Materials and Methods: Experimental design). Data are
presented as mean 6 SE.
836
J. ROBIN SVENSSON ET AL.
tant also to consider the composition of communities,
and not only the design of experimental treatments,
when comparing results and conclusions from experiments on effects of disturbance and productivity on
diversity.
Another potential explanation of the difference in
results and conclusions between this study and previous
studies on the DEM is based on the application and the
definition of disturbance. In this study we used
controlled levels of mechanical scraping. This type of
disturbance shares important properties with natural
disturbances, such as ice-scouring (Åberg 1992), drifting
logs (Dayton 1971), and wave action (Dudgeon et al.
1999), in the sense that it makes free space available for
settling (i.e., the limiting resource). This is a central
component in definitions of disturbance (Sousa 1984,
2001), which is not always considered in experimental
manipulations (e.g., Rashit and Bazin 1987, Scholes et
al. 2005). Another potentially complicating issue is the
selectivity of agents of disturbance in manipulative
experiments. Worm et al. (2002) used mesoherbivores
as agents of disturbance in communities composed
largely of macroalgae. In this case it is possible that
interactions, not predicted by the DEM, occurred
between grazing and nutrient enrichment of algae.
Grazers have been shown to prefer plants (Onuf et al.
1977) and macroalgae (Cruz-Rivera and Hay 2000) with
higher nutrient content, whereas physical disturbance
has no such selectivity. Grazing has previously been
argued as an unsuitable agent of disturbance in studies
on the IDH (e.g., McGuinness 1987, Sousa 2001). Due
to selective preference for nutrient-rich individuals,
grazing might be an even less appropriate agent of
disturbance in studies on the DEM.
Despite its conceptual appeal, the scarcity of manipulative studies suggests that empirical testing of the
DEM may not be straightforward. One important issue
that has to be considered in experimental tests of the
DEM is that the extensive discussion about agents and
definitions of disturbance (Grime 1977, Pickett and
White 1985, Sousa 2001) has no equivalence for
productivity. Experimental manipulation of productivity is often done indirectly, i.e., by adding nutrients. This
has two fundamental implications for the interpretation
of manipulative experiments. First, it becomes necessary
to test not only for effects of the experimental treatment
on diversity, but also to test independently whether the
actual experimental treatment (the adding of nutrients)
has an effect on productivity. Without evidence for an
actual increase in productivity, it is not clear whether the
experiment is testing the DEM or not. Unfortunately,
this is not always made clear (e.g., Widdicombe and
Austen 2001, Scholes et al. 2005, Jara et al. 2006).
Another problematic issue is the fact that productivity
of an assemblage is determined both by external factors
(i.e., light, temperature, energy transport, and nutrients)
and internal processes (i.e., differences in usage of
resources, resource capture ability, and energy conver-
Ecology, Vol. 88, No. 4
sion ability within and among species [Tilman 1980,
Tilman and Pacala 1993]). In a field experiment on
natural assemblages, energy conversion ability is usually
not amenable to manipulation. One consequence is that
there may be a lack of independence between the
response variable and the levels of the experimental
factor. This is because the productivity of an assemblage
may influence diversity (e.g., Connell and Orias 1964,
Abrams 1995) at the same time as the diversity
influences the productivity (e.g., Tilman et al. 1996).
Therefore, in an experiment where productivity is
manipulated indirectly, the response variable (i.e., some
measure of diversity) may modify the effect of the
experimental treatment. This relationship may lead to
confusion about cause and effect in otherwise carefully
planned experiments. Nevertheless, if predictions about
effects of productivity on diversity are to be tested in
field experiments, indirect manipulations may be the
only conceivable solution. In this system, addition of
nutrients, which are often a limiting resource, is
probably the most effective way to increase productivity
in a field experiment (e.g., Widdicombe and Austen
2001, Worm et al. 2002, Jara et al. 2006).
In a manner similar to manipulations of disturbance,
the experimental manipulations of productivity in
natural communities are often selective. The matter of
selectivity is probably of greater concern in experimental
manipulations of productivity, because designing a
nonselective agent of productivity is more complicated
then designing a nonselective agent of disturbance. If all
organisms are affected equally by the productivity
treatment, or if the dominant organisms are affected
relatively more strongly, it would require a stronger
disturbance to prevent competitive exclusion, as predicted by the DEM. However, if the inferior competitors
are more strongly affected by the productivity treatment, this could instead slow down the process of
competitive exclusion, which would cause diversity to
peak at lower intensities of disturbance, rather than at
the predicted higher intensities. In this experiment, the
dominant tunicates, unlike the ephemeral macroalgae,
did not noticeably increase their growth rates in
response to the nutrient treatment. This result could
explain why an interaction between disturbance and
productivity was not found. Jara et al. (2006) also
discussed the nutrient treatment as a possible cause for
their weak support for the DEM, because the nutrients
may only have affected the autotrophic part of the
community. Studies that have found the predicted
interaction between disturbance and productivity have
predominantly been made in plant communities (Pollock et al. 1998, Death 2002) or algae (Worm et al.
2002). In such experiments, the species in the communities would be more equally affected, even though
individual species of plants and algae differ in their
ability to utilize available resources.
In this study, we found maximum richness at
intermediate frequencies of disturbance, which is in
April 2007
DISTURBANCE, PRODUCTIVITY, AND DIVERSITY
accordance with the IDH. A literature review showed
that this is not a universal pattern in experimental tests
of effects of disturbance on diversity (Mackey and
Currie 2001). Less than 20% of the published studies on
disturbance–diversity relations supported the IDH. As
an extended theory, the DEM may explain some of the
results that are inconsistent with the IDH; and it has
therefore been suggested that it is preferable to the IDH
(Stallins 2003). In their review, Mackey and Currie
(2001) found that .50% of all experiments on disturbance showed either monotonically positive or negative
patterns with increasing disturbance. These patterns
could in principle be explained by the DEM, if it could
be shown that productivity was high in cases where
diversity increased with disturbance, and low when
diversity decreased with disturbance. The explanatory
power of the DEM is therefore potentially large.
Nevertheless, many alternative explanations may be
proposed for results that are inconsistent with the IDH.
Several authors have suggested that the predictions of
the IDH rely on a number of prerequisites, such as
competitive exclusion (Connell 1978), large regional
species pool (Osman 1977), multiple stages in succession
(Collins and Glenn 1997), and trade-off between
competition and tolerance (Dial and Roughgarden
1998) and between competition and colonization (Petraitis et al. 1989). Menge and Sutherland (1987) argued
that the effects of disturbance depend on the amount of
environmental stress in the system. Accordingly, experiments in systems where these prerequisites are not
fulfilled seldom find support for the IDH. For instance
Cowie et al. (2000) and Huxham et al. (2000) did not
observe maximum diversity at intermediate levels of
disturbance, because settling propagules and a small
regional species pool were lacking. Studies testing the
DEM have also explained lack of support for the model
with the failure of fulfilment of these requirements.
Scholes et al. (2005) suggested that the absence of
recolonization of bacteria and ciliates could explain lack
of support in the closed microcosms, while Death (2002)
found that the DEM could not predict patterns of
diversity in forest streams because such systems are not
driven by competition. These results imply that models
incorporating productivity is only one of several
possibilities for improving our understanding of mechanisms behind patterns of diversity. Furthermore, the
predictive power and general applicability of the DEM
needs to be assessed by further experiments in natural
assemblages, where the definition and the ecological
relevance of disturbance and productivity treatments are
explicitly considered.
ACKNOWLEDGMENTS
This work was funded by Formas through contract
21.0/2004-0550 to H. Pavia and by Stiftung Mercator through
a grant to M. Wahl. Fertilizer was provided by Aglukon,
Düsseldorf, Germany. We thank Anneli Lindgren and Elisabet
Brock for assistance with identifying species of macroalgae;
without their help it would not have been possible to do this
837
project. We also thank R. T. Paine and one anonymous
reviewer, whose comments greatly improved an earlier version
of the manuscript.
LITERATURE CITED
Åberg, P. 1992. A demographic study of two populations of the
seaweed Ascophyllum nodosum. Ecology 73:1473–1487.
Abrams, P. A. 1995. Monotonic or unimodal diversity–
productivity gradients. What does competition theory
predict? Ecology 76:2019–2027.
Anderson, R. C., D. Nelson, M. R. Anderson, and M. A.
Rickey. 2005. White-tailed deer (Odocoileus virginianus
Zimmermann) browsing effects on tallgrass prairie forbs:
diversity and species abundances. Natural Areas Journal 25:
19–25.
Bertocci, I., E. Maggi, S. Vaselli, and L. Benedetti-Cecchi. 2005.
Contrasting effects of mean intensity and temporal variation
of disturbance on a rocky seashore. Ecology 86:2061–2067.
Collins, S. L., and S. M. Glenn. 1997. Intermediate disturbance
and its relationship to within- and between-patch dynamics.
New Zealand Journal of Ecology 21:103–110.
Collins, S. L., S. M. Glenn, and D. J. Gibson. 1995.
Experimental analysis of intermediate disturbance and initial
floristic composition—decoupling cause and effect. Ecology
76:486–492.
Connell, J. H. 1978. Diversity in tropical rain forests and coral
reefs. Science 199:1302–1310.
Connell, J. H., and E. Orias. 1964. The ecological regulation of
species diversity. American Naturalist 98:399–414.
Cowie, P. R., S. Widdicombe, and M. C. Austen. 2000. Effects
of physical disturbance on an estuarine intertidal community:
field and mesocosm results compared. Marine Biology 136:
485–495.
Cruz-Rivera, E., and M. E. Hay. 2000. The effects of diet
mixing on consumer fitness: macroalgae, epiphytes, and
animal matter as food for marine amphipods. Oecologia 123:
252–264.
Dayton, P. K. 1971. Competition, disturbance, and community
organization: the provision and subsequent utilization of
space in a rocky intertidal community. Ecological Monographs 41:351–389.
Death, R. G. 2002. Predicting invertebrate diversity from
disturbance regimes in forest streams. Oikos 97:18–30.
Dial, R., and J. Roughgarden. 1998. Theory of marine
communities: the intermediate disturbance hypothesis. Ecology 79:1412–1424.
Dudgeon, S. R., R. S. Steneck, I. R. Davison, and R. L. Vadas.
1999. Coexistence of similar species in a space-limited
intertidal zone. Ecological Monographs 69:331–352.
Fischer, A. G. 1960. Latitudinal variations in organic diversity.
Evolution 14:64–81.
Ginzburg, L. R., and C. X. J. Jensen. 2004. Rules of thumb for
judging ecological theories. Trends in Ecology and Evolution
19:121–126.
Grime, J. P. 1977. Evidence for the existence of three primary
strategies in plants and its relevance to ecological and
evolutionary theory. American Naturalist 111:1169–1194.
Gutt, J., and D. Piepenburg. 2003. Scale-dependent impact on
diversity of Antarctic benthos caused by grounding of
icebergs. Marine Ecology Progress Series 253:77–83.
Henderson, C. R. 1953. Estimation of variance and covariance
components. Biometrics 9:226–252.
Hilborn, R., and M. Mangel. 1997. The ecological detective.
Princeton University Press, Princeton, New Jersey, USA.
Hubbell, S. P. 2001. The unified theory of biodiversity and
biogeography. Princeton University Press, Princeton, New
Jersey, USA.
Huston, M. A. 1979. A general hypothesis of species diversity.
American Naturalist 113:81–101.
Huston, M. A. 1994. Biological diversity: the coexistence of
species. Cambridge University Press, Cambridge, UK.
838
J. ROBIN SVENSSON ET AL.
Hutchinson, G. E. 1961. The paradox of the plankton.
American Naturalist 95:137–145.
Huxham, M., I. Roberts, and J. Bremner. 2000. A field test of
the intermediate disturbance hypothesis in the soft-bottom
intertidal. International Review of Hydrobiology 85:379–394.
Jara, V. C., J. H. S. Miyamoto, B. A. P. da Gama, M. Molis,
M. Wahl, and R. C. Pereira. 2006. Limited evidence of
interactive disturbance and nutrient effects on the diversity of
macrobenthic assemblages. Marine Ecology Progress Series
308:37–48.
Kondoh, M. 2001. Unifying the relationships of species richness
to productivity and disturbance. Proceedings of the Royal
Society B: Biological Sciences 268:269–271.
Lake, P. S., T. J. Doeg, and R. Marchant. 1989. Effects of
multiple disturbance on macroinvertebrate communities in
the Acheron River, Victoria. Australian Journal of Ecology
14:507–514.
Levin, S. A., and R. T. Paine. 1974. Disturbance, patch
formation, and community structure. Proceedings of the
National Academy of Sciences (USA) 71:2744–2747.
Lubchenco, J., and B. A. Menge. 1978. Community development and persistence in a low rocky intertidal zone.
Ecological Monographs 48:67–94.
Mackey, R. L., and D. J. Currie. 2001. The diversity–
disturbance relationship: is it generally strong and peaked?
Ecology 82:3479–3492.
McGuinness, K. A. 1987. Disturbance and organisms on
boulders 2. Causes of patterns in diversity and abundance.
Oecologia 71:420–430.
Menge, B. A., and J. P. Sutherland. 1987. Community
regulation: variation in disturbance, competition, and predation in relation to environmental stress and recruitment.
American Naturalist 130:730–757.
Miller, A. H. 1958. Ecologic factors that accelerate formation
of races and species of terrestrial vertebrates. Evolution 10:
262–277.
Molino, J. F., and D. Sabatier. 2001. Tree diversity in tropical
rain forests: a validation of the intermediate disturbance
hypothesis. Science 294:1702–1704.
Nilsson, P. 1991. Surveillance program in the coastal area
between Sweden and Norway. Swedish Evironmental Protection Agency Report 3910, Stockholm, Sweden.
Onuf, C. P., J. M. Teal, and I. Valiela. 1977. Interactions of
nutrients, plant growth and herbivory in a mangrove
ecosystem. Ecology 58:514–526.
Osman, R. W. 1977. The establishment and development of a
marine epifaunal community. Ecological Monographs 47:37–
63.
Paine, R. T. 1966. Food web complexity and species diversity.
American Naturalist 100:65–75.
Paine, R. T., and S. A. Levin. 1981. Intertidal landscapes:
disturbance and the dynamics of pattern. Ecological Monographs 51:145–178.
Paine, R. T., and R. L. Vadas. 1969. The effects of grazing by
sea urchins, Strongylocentrotus spp., on benthic algal
populations. Limnology and Oceanography 14:710–719.
Patricio, J., F. Salas, M. A. Pardal, S. E. Jorgensen, and J. C.
Marques. 2006. Ecological indicators performance during a
re-colonisation field experiment and its compliance with
ecosystem theories. Ecological Indicators 6:43–57.
Pavia, H., H. Carr, and P. Aberg. 1999. Habitat and feeding
preferences of crustacean mesoherbivores inhabiting the
brown seaweed Ascophyllum nodosum (L.) Le Jol. and its
epiphytic macroalgae. Journal of Experimental Marine
Biology and Ecology 236:15–32.
Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989.
The maintenance of species diversity by disturbance.
Quarterly Review of Biology 64:393–418.
Ecology, Vol. 88, No. 4
Pickett, S. T. A., and P. S. White. 1985. The ecology of natural
disturbance and patch dynamics. Academic Press, London,
UK.
Pollock, M. M., R. J. Naiman, and T. A. Hanley. 1998. Plant
species richness in riparian wetlands: a test of biodiversity
theory. Ecology 79:94–105.
Rashit, E., and M. Bazin. 1987. Environmental fluctuations,
productivity, and species diversity: an experimental study.
Microbial Ecology 14:101–112.
Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The
combined effects of energy and disturbance on species
richness in protist microcosms. Ecology Letters 8:730–738.
Soderstrom, J. 1996. The significance of observed nutrient
concentrations in the discussion about nitrogen and phosphorus as limiting nutrients for the primary carbon flux in
coastal water ecosystems. Sarsia 81:81–96.
Sousa, W. P. 1979. Disturbance in marine intertidal boulder
fields: the nonequilibrium maintenance of species diversity.
Ecology 60:1225–1239.
Sousa, W. P. 1984. The role of disturbance in natural
communities. Annual Review of Ecology and Systematics
15:353–391.
Sousa, W. P. 2001. Natural disturbance and the dynamics of
marine benthic communities. Sinauer, Sunderland, Massachusetts, USA.
Stallins, J. A. 2003. Dune plant species diversity and function in
two barrier island biogeomorphic systems. Plant Ecology
165:183–196.
Tilman, D. 1980. Resources: a graphical–mechanistic approach
to competition and predation. American Naturalist 116:363–
393.
Tilman, D., and S. Pacala. 1993. The maintenance of species
richness in plant communities. University of Chicago Press,
Chicago, Illinois, USA.
Tilman, D., D. Wedin, and J. Knops. 1996. Productivity and
sustainability influenced by biodiversity in grassland ecosystems. Nature 379:718–720.
Underwood, A. J. 1997. Experiments in ecology: their logical
design and interpretation using analysis of variance. Cambridge University Press, Cambridge, UK.
Valdivia, N., A. Heidemann, M. Thiel, M. Molis, and M. Wahl.
2005. Effects of disturbance on the diversity of hard-bottom
macrobenthic communities on the coast of Chile. Marine
Ecology Progress Series 299:45–54.
Widdicombe, S., and M. C. Austen. 2001. The interaction
between physical disturbance and organic enrichment: an
important element in structuring benthic communities.
Limnology and Oceanography 46:1720–1733.
Wikstrom, S. A., M. B. Steinarsdottir, L. Kautsky, and H.
Pavia. 2006. Increased chemical resistance explains low
herbivore colonization of introduced seaweed. Oecologia
148:593–601.
Winer, B. J., D. R. Brown, and K. M. Michels. 1991. Statistical
principles in experimental design. Third edition. McGrawHill, New York, New York, USA.
Wootton, J. T. 1998. Effects of disturbance on species diversity:
a multitrophic perspective. American Naturalist 152:803–
825.
Worm, B., H. K. Lotze, H. Hillebrand, and U. Sommer. 2002.
Consumer versus resource control of species diversity and
ecosystem functioning. Nature 417:848–851.
Worm, B., T. B. H. Reusch, and H. K. Lotze. 2000. In situ
nutrient enrichment: methods for marine benthic ecology.
International Review of Hydrobiology 85:359–375.
Zucchini, W. 2000. An introduction to model selection. Journal
of Mathematical Psychology 44:41–61.
Paper I
Paper II
PAPER
Paper
III II
Paper IV
Paper V
“You tried your best and you failed miserably. The lesson is
'never try'.“
-Homer J. Simpson
Paper VI
Ecology, 90(2), 2009, pp. 496–505
Ó 2009 by the Ecological Society of America
Equal rates of disturbance cause different patterns of diversity
J. ROBIN SVENSSON,1 MATS LINDEGARTH,
AND
HENRIK PAVIA
Department of Marine Ecology–Tjärnö, University of Gothenburg, 452 96 Strömstad, Sweden
Abstract. Empirical evidence suggests that disturbance has profound effects on the species
diversity of aquatic and terrestrial assemblages. Conceptual ecological theories, such as the
intermediate disturbance hypothesis (IDH), predict maximum diversity at intermediate levels
of disturbance. Tests of the predictive power and generality of these models are, however,
hampered by the fact that the meaning and units of ‘‘disturbance’’ are not clearly defined. For
example, it is seldom recognized that the rate of disturbance is the product of both frequency
and extent (e.g., area or volume) of disturbance events. This has important consequences for
the design and interpretation of experiments on disturbance. Here we present, for the first
time, an experimental design that allows for unconfounded testing of combinations of area
and frequency (i.e., regimes) for a given rate of disturbance. We tested the prediction that
species richness responds differently to equal rates of disturbance, depending on the specific
combination of frequency and area, on marine hard-substratum assemblages. Five different
rates of disturbance and two regimes (small frequent or large infrequent disturbances) were
applied at three sites. The results showed that the effect of a certain rate of disturbance (1)
varies strongly among assemblages and (2) also depends on the specific combination of
frequency and area of disturbance events. Maximum species richness was observed at
intermediate rates of disturbance at site 1 (i.e., support for the IDH), whereas there was a
monotonic decline at site 2 and there was no evident pattern at site 3. The variable responses
among sites were explained by differences in degree of competitive exclusion and rates of
recruitment. At the site where the IDH was supported, the regime with a large proportion of
the area disturbed infrequently showed higher richness, compared to the regime with a small
proportion disturbed frequently. This was likely due to a stronger decrease of dominants,
which allowed for the recruitment of new colonizing species. In summary, we conclude that
tests and general syntheses of models of disturbance–diversity patterns would benefit from
more explicit definitions of the components of disturbance, as well as a stronger focus on the
importance of variation in inherent properties of natural assemblages.
Key words: competitive exclusion; diversity; marine assemblages; rate of disturbance; regime; species
richness.
INTRODUCTION
Disturbance is an important factor explaining patterns of biodiversity in many terrestrial (e.g., Eggeling
1947, Grime 1977, Molino and Sabatier 2001) and
aquatic environments (e.g., Sousa 1979, Patricio et al.
2006). One of the most influential formulations of the
effects of disturbance on biological diversity is the
intermediate disturbance hypothesis (IDH; Connell
1978). The IDH predicts low diversity at low levels of
disturbance due to competitive exclusion and also at
high levels as a result of local extinction. At intermediate
levels of disturbance, diversity is higher due to coexistence of rapid colonizers and competitive dominants.
The IDH has been supported in laboratory experiments
(e.g., Widdicombe and Austen 1999, Buckling et al.
2000, Cowie et al. 2000) as well as field experiments in
terrestrial (e.g., Armesto and Pickett 1985, Collins 1987,
Manuscript received 4 October 2007; revised 4 June 2008;
accepted 12 June 2008. Corresponding Editor: H. Hillebrand.
1 E-mail: [email protected]
496
Molino and Sabatier 2001), freshwater (e.g., Padisak
1993, Reynolds 1995, Flöder and Sommer 1999) and
marine communities (e.g., Osman 1977, Sousa 1979,
Valdivia et al. 2005). Nevertheless, a literature review
revealed that only 20% of the studies on effects of
disturbance on diversity showed the unimodal pattern
predicted by the IDH (Mackey and Currie 2001). Thus,
despite its conceptual appeal, it appears that the
predictive value of the IDH is often limited. This has
led to the development of more complex models, e.g., by
incorporating effects of productivity (the dynamic
equilibrium model [DEM]; Huston 1979, 1994), which
could potentially account for a wider range of patterns.
One important focus for research in general is to
understand the limitations of a theory and to unravel
causes of variable predictive power. A fundamental
difficulty with the IDH and the DEM is that they are
largely conceptual and verbal models based on relatively
scaled variables (Schoener 1972, Peters 1991). Therefore
the units and meaning of disturbance is sometimes
unclear and often different among studies (Pickett and
White 1985, Sousa 2001; Table 1). Conceptual terms, such
February 2009
DISTURBANCE RATE AND COMMUNITY DYNAMICS
497
TABLE 1. Conceptual and operational terms used to define the magnitude of disturbance in
ecological literature.
Term
Conceptual
Level
Severity
Intensity
Regime
Operational
Frequency
Time
Extent
Size
Rate
Meaning
Quantity
general description of overall magnitude of disturbance
general description often used as ‘‘strength’’ of the
disturbing force
general description sometimes used synonymously to severity
generic term for the types and components of disturbance
currently acting in a given area
number of disturbance events per unit time
period of time since last disturbance event
total two- or three-dimensional space disturbed
size of an individual disturbance event
product of area and frequency
as ‘‘intensity’’ and ‘‘severity,’’ are not explicitly defined,
and are therefore not easily generalized among studies.
For example, ‘‘intensity’’ has been used to describe a
variety of experimental manipulations and variables, such
as penetration depth per bite by limpets (Steneck et al.
1991), type of mechanical scrubbing (McCabe and Gotelli
2000), and degree of oscillation in sediment (Garstecki
and Wickham 2003). Clearly, there is a need for systemspecific concepts, but in order to evaluate general
ecological theories, such as the IDH, it is important that
concepts are commensurable among studies.
Disturbance can be operationally defined as a rate,
i.e., the sum of the size of all disturbance events in a
given area per unit time (Miller 1982). The rate is the
only measure, which accounts for the combined effects
of area and frequency, and thus the total amount of
disturbance actually inflicted upon the community under
study. This is important because information about one
of these components makes no sense without the context
of the other. For instance, the information that a defined
biological community was disturbed once a week is
completely uninformative if we do not know the extent
of the disturbance. Surely, we would predict massive
differences in effects on diversity if the area disturbed
each week was 0.5% of the total area compared to if it
was 20%. This is a fundamental fact that is always
considered in the process of defining appropriate
treatment levels in ecological experiments, but it is later
also commonly disregarded in the interpretation and
discussion of the study. Consequently, the combined
effects of frequency and area are always implicit in
experimental studies on the effects of disturbance, but in
order to put any experimental result into a wider
context, and to allow for direct and meaningful
comparisons among studies, it is always necessary to
transform the measure of disturbance into a rate.
From this it does, however, not follow that the rate of
disturbance is always the most accurate predictor of
diversity. Several theoretical arguments can be made,
which suggest that the effect of a given rate of
disturbance may differ depending on the way the specific
time1
time
area or volume
area
area/time
rate is obtained, i.e., the regime. For instance, Miller
(1982) suggested that small, frequent disturbances favor
species with rapid vegetative growth (‘‘competitive’’
species), whereas large, less frequent disturbances favor
species with high capacity for dispersal (‘‘colonizers’’)
due to the differences in perimeter to area ratios among
patches. Although the area of disturbance have been the
predominant focus for models of rates of disturbance
(i.e., Miller 1982), the other component of the rate,
frequency, is equally important. Differences in frequency and timing of disturbance will, similarly to variations
in area, influence the abundance and composition of
natural communities (Sousa 2001), because species are
likely to increase in abundance when the disturbance
regime matches their preferred recruitment time (Underwood and Anderson 1994, Crawley 2004). The large
variation in temporal distribution of propagules among
species (Roughgarden et al. 1988, Underwood and
Anderson 1994) will also have the consequence that a
single large disturbance cannot be colonized by all
species in the regional species pool, but only by the
propagules that are available at the specific time when
space is made free. Thus, it is evident that variations in
both area and frequency of a disturbance rate are likely
to influence the outcome of studies on effects of
disturbance on diversity. However, despite that area
and frequency have well known distinct effects on
diversity, no study, to our knowledge, has performed
unconfounded testing of these factors for a given rate of
disturbance.
Testing of differences among regimes at equal rates of
disturbance is not straightforward because experimental
manipulations of the frequency or the extent of
disturbance are inevitably associated with a change in
the rate of disturbance. Previous experiments, simultaneously testing hypotheses about effects of frequency
and extent, have done this using orthogonal designs
(e.g., Collins 1987, McCabe and Gotelli 2000). The
interpretation of such experiments is problematic
because tests of the effects of one factor, within one
level of the other factor, are confounded by changes in
498
J. ROBIN SVENSSON ET AL.
TABLE 2. Explanation of each combination of experimental
factors, regime (Re1, small, frequent disturbances; Re2,
large, infrequent disturbances), and frequency and rate of
disturbance.
Regime
Area (cm2)
Frequency
(no./week)
Rate
(cm2/week)
Re1
Re1
Re1
Re1
Re1
45
45
45
45
45
0
2/16
4/16
8/16
16/16
0
5.63
11.25
22.5
45
Re2
Re2
Re2
Re2
Re2
90
90
90
90
90
0
1/16
2/16
4/16
8/16
0
5.63
11.25
22.5
45
the rate. In order to test for differences among regimes,
specific analytical contrasts need to be extracted and
parts of the experiment become redundant. Furthermore, in a direct analogy to generalizations among
studies, any significant interaction resulting from such
an experiment can only be sensibly interpreted if the rate
of disturbance is considered.
Here we present, for the first time, a field experiment
with an experimental design that allows for unconfounded testing of combinations of area and frequency
(i.e., regimes) for a given rate of disturbance. In a
previous manipulative experiment in marine hardsubstratum communities on the west coast of Sweden,
we found that maximum species richness was attained at
intermediate frequencies of disturbance (Svensson et al.
2007). Following from this work, our aim was to use
hard-substratum assemblages to test the hypothesis that,
in these assemblages, the effect of a certain rate of
disturbance depends on the specific combination of area
and frequency. In order to do this mechanical disturbance was applied at five distinct rates of disturbance
under two different regimes (disturbing either a small
proportion of the assemblage, frequently, or a large
proportion, infrequently) at three sites.
MATERIALS
AND
METHODS
Study site
The field experiment was conducted on the west coast
of Sweden in the vicinity of Tjärnö Marine Biological
Laboratory. The experimental sites were three bays
located approximately 1 km apart (58852 0 92 00 N,
1188 0 3100 E; 58852 0 2000 N, 1188 0 7000 E; and 58852 0 7600 N,
1188 0 1500 E for sites 1, 2, and 3, respectively). Site 1 had
an average depth of 8 m and was surrounded by muddy
and rocky shores inhabited by red, green and brown
macroalgae as well as mussels and tunicates. Site 2 had
an average depth of 6 m and was surrounded by sandy
beaches and boulder fields. This site also had an
extensive seagrass (Zostera marina) meadow and the
boulders were commonly overgrown by fucoids, barnacles, and mussels. Site 3 had an average depth of 10 m
Ecology, Vol. 90, No. 2
and was surrounded exclusively by rocky shores with a
steep declining sandy bottom. The nearby rocks were
predominantly occupied by breadcrumb sponge (Halichondria panicea), ephemeral red algae, and fucoids.
Experimental design
Experimental units, made from 2100 3 250 3 4 mm
PVC strips folded into a ring, were placed hanging from
a buoy approximately 0.5 m below the water surface
(Svensson et al. 2007). Ten quadratic PVC panels (150 3
150 3 3 mm), roughened with emery paper, were
attached with cable ties on each of the 24 rings. Eight
rings were deployed at each site on 25 April to allow
settling and establishment of communities before the
experimental manipulation started on 25 May. The
experimental manipulation had a duration of 17 weeks
and was terminated 28 September 2005.
The panels were disturbed by either of five different
rates of disturbance (Ra0Ra4), under two different
disturbance regimes (Re1 and Re2; Table 2). Under the
first regime two randomly selected areas were scraped,
each covering 10% of the panel area, with frequencies
ranging from every week to every eighth week. Under the
second regime, four randomly selected areas, each
covering 10% of the panel area, were scraped at
frequencies ranging from every second week to 16th week
(Table 2). In addition to killing or damaging individuals,
the disturbance treatment also facilitated recruitment by
freeing substratum and is therefore coherent with the
definition of disturbance by Sousa (1984). All rates were
present in all rings with two replicate panels and either of
the two regimes was randomly assigned to rings. Thus,
four rings of regime 1 and four rings of regime 2 was
present at each of the three sites, allowing eight rings with
all five rates replicated per site.
Sampling
The composition and abundance of assemblages were
sampled at the end of the experiment after 17 weeks of
manipulation. Panels were detached and brought into
the laboratory submerged in seawater and then kept
under running seawater in the laboratory during the
entire sampling procedure. Wet weight was measured
and the edges and reverse side of all panels were scraped
clean before sampling. Percent cover of bare space and
sessile species was then estimated in 5% intervals using a
15 3 15 cm plastic grid. A 1-cm margin to all edges of the
panels was not assessed in order to avoid confounding
edge effects. The percentage cover of species with a small
holdfast and wide thallus was estimated from the two
dimensional projection of the organism on the panel.
Cover of epibionts was also estimated, thus, total cover
could exceed 100%.
Statistical analyses
The data obtained from the experiment was analysed
with analysis of variance (ANOVA) using Statistica 6.0
(Statsoft, Tulsa, Oklahoma, USA). Hypotheses about
February 2009
DISTURBANCE RATE AND COMMUNITY DYNAMICS
499
FIG. 1. Abundance of species in undisturbed communities (rate 0) averaged over regime at sites 1, 2, and 3. Different letters
indicate significant difference in community composition (ANOSIM; R . 0.6, P ¼ 0.001).
effects of main factors and interactions were tested using
the following general linear model:
Xijklm ¼ l þ Si þ Rej þ SReij þ Rak þ SRaik þ ReRajk
þ SReRaijk þ RiðSReÞlðijÞ þ RaRiðSReÞklðijÞ þ eijklm
where l is the overall mean, site (Si) is a random factor
with three levels, disturbance regime (Rej) is a fixed
factor with two levels, disturbance rate (Rak) is a fixed
factor with five levels, ring (Ri[SRe]l(ij)) is a nested
random factor with four levels and eijklm is a random
deviation.
Support for the hypothesis that equivalent rates of
disturbance may cause different effects depending on the
specific combination of frequency and area of disturbance, would be shown by a significant interaction (Ra
3 Re). More specifically the hypothesis is supported if a
difference in species richness is found at equivalent rates
of disturbance for regimes with the larger area (regime
2), compared to regimes with the smaller area (regime 1).
Support for the alternative, i.e., that the effect of rate of
disturbance does not depend on the particular combination of frequency and area, is obtained from a
significant effect of rate of disturbance (Ra) if there is
a nonsignificant interaction (Ra 3 Re). Furthermore, to
evaluate whether the response of disturbance is consistent with the predictions from the IDH, two additional
tests are necessary. First, if there is a significant negative
quadratic component in a polynomial regression this
indicates that the response is non-linear and that the rate
of increase declines at larger rates of disturbance.
Second, in order to test whether the maximum species
richness occurs at a rate of disturbance that is different
from the two extreme rates of disturbance in our study,
we performed a Mitchell-Olds and Shaw’s test (MOS
test [Mitchell-Olds and Shaw 1987]). This method has
previously been used to investigate patterns of curvilinear relationships between diversity and productivity,
where significant results of the MOS test shows support
for hump-shaped or U-shaped patterns (Mittelbach et
al. 2001, Chase and Leibold 2002, Fukami and Morin
2003).
Initial tests of differences among sites in undisturbed
assemblages (Ra0) were made using analysis of similarity (ANOSIM) using the PRIMER software package
(PRIMER-E, Plymouth, UK). The abundance data was
transformed to the fourth root, in order to avoid bias by
the greater influence of abundant species, following the
recommendations by Clarke and Warwick (1994). These
analyses revealed whether differences were solely related
to the number of species or also dependent on the
particular species present and their relative abundance.
The individual contributions of different species to the
observed differences among sites were evaluated using
SIMPER. A graphical comparison among sites was
obtained using non-metric multidimensional scaling (nMDS).
RESULTS
Structure of assemblages
A total of 19 species of algae and 16 species of sessile
invertebrates were observed in the experimental communities at the time of sampling. The most abundant
organisms, occupying large areas of the panels, were the
tunicate Ciona intestinalis, the common blue mussel
Mytilus edulis, the red algae Ceramium rubrum, and the
hydroid Laomedea flexuosa, whereas sea anemones,
bryozoans, barnacles, and most ephemeral algae were
found at low cover (Fig. 1). The composition of species
in undisturbed assemblages differed among all three sites
(ANOSIM; global R . 0.6, P , 0.005 for all pairwise
tests; Fig. 2). Difference in cover of C. intestinalis
explained most of the dissimilarity among all sites
(SIMPER; 77%, 54%, and 35%, sites 1 vs. 2, 1 vs. 3, and
2 vs. 3, respectively). The total area covered of sessile
species in the undisturbed assemblages exceeded 100% at
site 1, whereas assemblages at sites 2 and 3 had less than
full coverage of available space, indicating that the
500
J. ROBIN SVENSSON ET AL.
FIG. 2. Multidimensional scaling (MDS) of species composition for all sites averaged over disturbance regime.
strength of competition for space differed substantially
among sites (Fig. 1). The ascidian C. intestinalis covered
over 95% of the space in the undisturbed assemblages at
site 1, and was also the most abundant species at site 2
occupying 33% of the panel. At site 3 no single species
occupied more than 15% of the available space in the
assemblage.
Responses to rates of disturbance
There was no overall effect of the rate of disturbance
on species richness and thus there was no general
support for the IDH (Table 3). Instead, the effects of the
rate of disturbance differed significantly among sites (P
, 0.05 for S 3 Ra; Table 3, Figs. 3 and 4). A graphical
examination of the nature of this interaction suggested
that there was maximum richness at intermediate rates
of disturbance at site 1, a decrease in richness with
increasing rate of disturbance at site 2, while no clear
pattern was distinguishable at site 3 (Fig. 4). Additional
analyses using polynomial regression showed that there
were significant linear and quadratic components at site
1, a significant decreasing linear trend at site 2, and no
significant pattern at site 3 (Table 4). Because significant
quadratic components do not automatically show
evidence for internal maxima in polynomial regressions,
a Mitchell-Olds and Shaw’s test was performed for
assemblages at site 1. This test showed that maximum
richness occurs at an intermediate rate of disturbance
Ecology, Vol. 90, No. 2
(b1/2b2 ¼ 32.8, P , 0.01 for Mitchell-Olds and Shaw’s
test), and comparisons of adjusted R2 showed that the
hump-shaped model had a better fit to our data
compared to the linear model (adjusted R2 ¼ 0.213 and
0.156, respectively). Thus, although the IDH was not
globally supported, the patterns observed at site 1 were
not only statistically significant but precisely those
predicted by the IDH (Table 4).
Analyses of individual species also showed that these
differed in their response to rate of disturbance (Fig. 5).
Inspection of mean cover showed that the highly
abundant tunicate C. intestinalis was negatively affected
by the disturbance treatment, whereas rapid colonizers,
such as the ephemeral algae Enteromorpha intestinalis
and Ectocarpus siliquosis, increased in cover with
increasing disturbance rate. Although disturbance rate
had a negative effect on richness on many panels it is still
noteworthy that assemblages subjected to the highest
rate, which constituted of scraping 320% of the panel
area over a period of 16 weeks, still had between 5 and
11 species at the termination of the experiment.
Responses to regimes of disturbance
Analysis of species richness at the end of the
experiment also showed that there was a significant
interaction among factors site, regime and disturbance
rate (Table 3). Although calculations of variance
components showed that this component was only half
as large as that of the two-way interaction (for S 3 Re 3
Ra, r2 ¼ 0.46, and for S 3 Ra, r2 ¼ 0.98), this
interaction suggests that the effects of the rate of
disturbance differ between regimes, at least at some
sites. In order to test for differences among regimes in
assemblages where maximum richness was observed at
intermediate rates (Fig. 3), post hoc tests for differences
between regimes within rates were done at site 1. SNK
test showed that regime 2 (large and infrequent
disturbances), had significantly greater species richness
than regime 1 (small and frequent disturbances) at rate
2. In order to further investigate the underlying cause for
this difference in richness we compared the abundance
of species among regimes (Table 5). Assemblages in
regime 2 had more species of ephemeral algae (i.e.,
‘‘colonizers’’) and most algal species had higher coverage, whereas the ascidian C. intestinalis and the blue
TABLE 3. Analysis of variance on species richness.
Source
df
MS
F
P
Error term
Site, S
Regime, Re
Rate, Ra
S 3 Re
S 3 Ra
Re 3 Ra
S 3 Re 3 Ra
Ring, Ri(S 3 Re)
Ra 3 Ri(S 3 Re)
Residual
2
1
4
2
8
4
8
17
68
115
89.26
2.03
5.31
7.77
18.26
4.60
6.26
10.71
2.58
2.17
8.34
0.26
0.29
0.73
7.09
0.74
2.43
4.95
1.19
0.003
0.660
0.876
0.499
0.000
0.593
0.023
0.000
0.205
Ri(S 3 Re)
S 3 Re
S 3 Ra
Ri(S 3 R)
Ra 3 Ri(S 3 Re)
S 3 Re 3 Ra
Ra 3 Ri(S 3 Re)
residual
residual
February 2009
DISTURBANCE RATE AND COMMUNITY DYNAMICS
501
FIG. 3. Effects of rate of disturbance on species richness at different sites and disturbance regimes. Data are presented as mean
þ SE. Disturbance rates and regimes are described in Table 2.
mussel M. edulis (i.e., ‘‘competitors’’) had higher cover
in regime 1. Furthermore, inspection of means showed
that maximum richness was attained at lower rates of
disturbance for regimes involving larger areas at site 1
(Fig. 3). Thus, in addition to significantly affecting the
number of species, the specific combination of area and
frequency of disturbance also determines what kind of
species that will be present in assemblages.
DISCUSSION
This experiment was designed to test the hypothesis
that species richness in a hard-substratum assemblage
respond in specific ways to different rates of disturbance
and that this response depends on the particular
combination of area and frequency of the disturbance
events. We found that the rate of disturbance affects
species richness in a hard-substratum assemblage in
different ways at different sites. At one site richness
responded in accordance with the predictions from the
IDH, while at the other sites there was a monotonic
negative effect on richness or richness was not affected
at all. Furthermore, at the site where patterns were
consistent with the IDH, maximum species richness was
observed at lower rates of disturbance when large areas
were scraped less frequently, compared to when small
areas were scraped more frequently. Because the
experiment allows comparisons among regimes at
comparable rates of disturbance, our study provide the
first unconfounded test and empirical evidence for the
hypothesis that different combinations of area and
frequency at equal rates have different effects on
diversity.
TABLE 4. Regressions of species richness on linear and
quadratic rates at individual sites.
FIG. 4. Species richness, averaged over disturbance regime,
as a function of the rate of disturbance at sites 1, 2, and 3. Data
are presented as mean 6 SE. See Table 2 for specific details of
how each rate can be obtained through two different regimes.
Source
df
MS
F
P
Site 1
Rate
Rate2
Residual
1
1
67
61.78
24.78
4.19
14.74
5.91
,0.001
0.018
Site 2
Rate
Rate2
Residual
1
1
77
46.76
0.32
2.87
11.16
0.08
0.001
0.783
Site 3
Rate
Rate2
Residual
1
1
77
2.89
1.03
2.85
0.69
0.25
0.409
0.621
502
J. ROBIN SVENSSON ET AL.
Ecology, Vol. 90, No. 2
FIG. 5. Effects of rate of disturbance on cover of C. intestinalis, E. silicuosus, and E. intestinalis at site 1, averaged over regime.
Data are presented as mean 6 SE.
Miller (1982) argued that the underlying cause for
observing higher diversity in assemblages under regimes
composed of larger, compared to smaller, areas at lower
rates, was that larger disturbances would favor species
with high capacity for dispersal (i.e., ‘‘colonizers’’) due
to the longer persistence of free space. Smaller disturbances should benefit ‘‘competitive’’ species because the
larger ratio of perimeter to area allows more rapid
reoccupation of space by vegetative growth. In our
experiment we investigated the effects of regimes with
different area and frequency with similar size and shape
of disturbance events at equivalent rates of disturbance.
This allowed testing for differences among regimes that
are not based on differences in perimeter area ratios.
Nevertheless, similar to the predictions in the model by
Miller (1982), we observed a larger number of species of
ephemeral algae (i.e., Ceramium rubrum, Cystoclonium
purpureum, and Ectocarpus siliquosis) in assemblages
under the regime composed of larger area at site 1, for
rates where this regime showed maximum richness. We
also found that the ephemeral algae and the hydroid
Laomedea flexuosa (i.e., colonizers [Svensson et al.
2007]) had higher percent cover in regimes with larger
area, while the ascidian Ciona intestinalis and the blue
mussel Mytilus edulis, which are strong competitors for
space (Lenz et al. 2004, Svensson et al. 2007), were more
abundant in the regime composed of smaller areas at
this site. This means that colonizing species may not
only be facilitated by larger ratios of perimeter to area,
as predicted by Miller (1982), but also by removing a
larger part of the assemblage where their propagules
settle. It has previously been shown that post-settlement
survivability of propagules is greatly affected by larval–
adult interactions, such as competition for food (Osman
et al. 1989) and space (Connell 1961), and also that the
survivability of propagules is more sensitive during the
first few days (Gosselin and Qian 1997). A possible
explanation for the difference among regimes in this
study may, therefore, be that post-settlement survivability among colonizers is higher where 40% of an
assemblage is removed, compared to 20%, due to lower
levels of larval–adult competition at the time following
the disturbance events. Although the underlying mechanisms for this pattern need to be unraveled by further
experiments, our results demonstrate that the disturbance regime will affect the outcome of studies
investigating effects of disturbance on diversity.
One striking result of this experiment was that the
assemblages at the three sites all differed substantially in
their response to disturbance. A likely explanation to the
variability among sites in their response to disturbance is
the natural variation in abundance and composition of
assemblages. The total cover in undisturbed assemblages
exceeded 100% at site 1, and the ascidian C. intestinalis
covered more than 95% of the area. At the other sites,
the total cover was less dense and no single taxa covered
more than 35%. The high percentage of total cover and
TABLE 5. Qualitative differences in cover between regime 1
(small area and frequent disturbances) and regime 2 (large
area and infrequent disturbances) at site 1 and rate 2.
Species
Difference
Algae
Cystoclonium purpureum
Dasya baillouviana
Ceramium rubrum
Cladophora rupestris
Ectocarpus siliquosus
Enteromorpha intestinalis
Spermotamnion repens
Osmondea truncata
Polysiphonia fucoides
Ceramium strictum
þþ
þþ
þ
þ
þ
þ
þ
Invertebrates
Electra pilosa
Membranipora membranacea
Ascidiella aspersa
Laomeda flexuosa
Ciona intestinalis
Mytilus edulis
Cryptosula pallasiana
þþ
þþ
þ
þ
Note: Symbols are as follows: þ, larger cover in regime 2; ,
smaller cover in large areas; , species occurs exclusively in
regime 1; and þþ, species occurs only in regime 2.
February 2009
DISTURBANCE RATE AND COMMUNITY DYNAMICS
503
FIG. 6. Effects of rate of disturbance on species richness in a study by Svensson et al. 2007 (experiment 1) and this study
(experiment 2) expressed as (a) relative disturbance and (b) absolute rate. Data are presented as mean 6 SE.
the dominance of the ascidians at site 1, suggests that C.
intestinalis is a strong competitor for space, excluding
other invertebrates and many species of macroalgae.
Such competitive exclusion is a fundamental premise for
observing higher diversity at intermediate levels of
disturbance in natural communities (e.g., Fuentes and
Jaksic 1988, Collins and Glenn 1997). Accordingly, the
lack of support for the IDH at the other sites is most
likely explained by the absence of clear dominance of
strong competitors. By reducing the cover of the strong
dominant C. intestinalis, and thereby preventing or
disrupting competitive exclusion, disturbance can have a
positive effect on diversity in assemblages with intense
competition. This mechanism has previously been
implied in field experiments in Sweden (Svensson et al.
2007) and Chile (Valdivia et al. 2005), where reduction
in cover of the dominant ascidians C. intestinalis and
Pyura chilensis both resulted an increase in diversity at
intermediate frequencies of disturbance. Additional
evidence for the importance of disrupting competitive
exclusion comes from several studies involving dominant organisms, such as mussels (Paine 1966), bryozoans
(Jara et al. 2006), trees (Molino and Sabatier 2001), and
grasses (Collins 1987).
An important factor with potentially large effects on
the outcome of disturbance experiments is the availability of propagules, which is not easily controlled or
measured in field experiments. The availability of
propagules and a large regional species pool, is of great
importance in order for disturbance to have a positive
effect on richness (Osman 1977). In a study on tallgrass
prairie vegetation Collins et al. (1995) pointed out that it
is settlement by propagules, and not disturbance per se,
which increases species richness. If no new species settle
in the cleared space, richness will obviously not increase.
This was clearly shown in an experiment on soft-bottom
intertidal assemblages by Huxham et al. (2000), where
the species settling in areas cleared by disturbance were
the same species that already inhabited the patch. The
different responses to disturbance among experimental
sites in this study, and the large difference in total cover
of undisturbed assemblages among the sites, suggests
that there is large spatial variation in propagule supply.
Jonsson et al. (2004) showed that local hydrodynamics
strongly influenced the settling of planktonic barnacle
larvae and caused highly variable recruitment on panels
at different sites in the archipelago where this study was
conducted. This could possibly explain the surprisingly
low cover (;35%) of substratum in the undisturbed
assemblages at site 3. However, the highly disturbed
panels at this site still did not show a large reduction in
diversity compared to the controls. This suggests that
some propagules were capable of settlement, and that
the amount of propagules that settled successfully could
compensate for the loss in species by the disturbance
treatment.
One important theme of this paper is the distinction
between conceptual and operational terms and its
consequences for the interpretation and synthesis of
empirical results. Comparisons of the result from site 1
in this study and the one by Svensson et al. (2007), which
both provide support for the IDH in the same system,
serves to illustrate some of these consequences (Fig. 6).
If disturbance is not consistently defined (e.g., frequency
the first study and rate in this study) and subsequently
compared on a relative scale, the outcomes of the
experiments appear similar in some respects but
different in others. Maximum diversity was predicted
at similar levels of disturbance (0.5 and 0.7) by both
models. The parameters of the fitted model were
consistent with respect to the intercept, which represents
the diversity in the absence of disturbance, while the
linear (the rate of increase in the absence of disturbance)
and the quadratic components (the curvature) were
504
J. ROBIN SVENSSON ET AL.
roughly doubled in this study, compared to the previous
(Fig. 6a). When levels of disturbance are transformed
into rates, however, it becomes obvious that maximum
diversity was obtained at rates approximately three
times larger in this compared to the previous study (Fig.
6b). Furthermore, an analysis of the parameters of the
fitted model show that the intercept and initial increase
were consistent between studies, but that the curvature,
i.e., the tendency to decline at higher levels of
disturbance, was less pronounced in this study compared
to the previous (Fig. 6b). These detailed analyses have
important consequences not only for the interpretation
of differences in patterns among studies, but also for the
hypotheses about the processes that were causing these
differences. In a relative perspective it appears that the
effect of competitive exclusion at low levels of disturbance was stronger in this compared to the previous
study (Fig. 6a). On an absolute scale, however, the
effects of competitive exclusion were similar in both
experiments, while lack of recruitment of new species in
the first experiment was responsible for the rapid decline
at higher rates of disturbance. Although this analysis
does not provide conclusive evidence about the importance of different processes it is clear that results may be
interpreted in very different ways depending on how
disturbance is represented.
Conclusions
This study provides the first unconfounded experimental design for testing the hypothesis that equal rates
cause different patterns of diversity. Because the
experiment was designed to test effects of regimes at
equivalent rates of disturbance, we were able to show
that effects on species richness depended on the specific
combination of frequency and area and not only on the
rate. Furthermore, the effects of a certain rate of
disturbance differed substantially among sites. At one
site maximum richness was observed at intermediate
rates (i.e., support for the IDH), at another site richness
of assemblages declined with increasing rates of disturbance and at the third site there was no effect of
disturbance. The variable responses among sites were
likely due to differences in degree of competitive
exclusion and rates of recruitment. In summary, the
results suggest that the general understanding of the
magnitude and nature of effects of disturbance would
benefit from more explicit definitions of the components
of disturbance and a stronger focus on the importance of
the inherent properties of natural assemblages.
ACKNOWLEDGMENTS
This study was financially supported by MARICE (an
interdisciplinary research platform at the Faculty of Sciences,
University of Gothenburg) and by the Swedish Research
Council through contract no. 621-2004-2658 to H. Pavia, as
well as by Formas through contracts 21.0/2004-0550 to H.
Pavia and 217-2006-357 to M. Lindegarth. We thank Malin
Karlsson for performing a large part of the fieldwork and
Elisabet Brock for assistance with identification of macroalgal
species.
Ecology, Vol. 90, No. 2
LITERATURE CITED
Armesto, J. J., and S. T. A. Pickett. 1985. Experiments on
disturbance in old-field plant communities: impact on species
richness and abundance. Ecology 66:230–240.
Buckling, A., R. Kassen, G. Bell, and P. B. Rainey. 2000.
Disturbance and diversity in experimental microcosms.
Nature 408:961–964.
Chase, J. M., and M. A. Leibold. 2002. Spatial scale dictates the
productivity–biodiversity relationship. Nature 416:427–430.
Clarke, K. R., and R. M. Warwick. 1994. Change in marine
communities: an approach to statistical analysis and interpretation. Natural Environment Research Council, Swindon,
UK.
Collins, S. L. 1987. Interaction of disturbances in tallgrass
prairie: a field experiment. Ecology 68:1243–1250.
Collins, S. L., and S. M. Glenn. 1997. Intermediate disturbance
and its relationship to within- and between-patch dynamics.
New Zealand Journal of Ecology 21:103–110.
Collins, S. L., S. M. Glenn, and D. J. Gibson. 1995.
Experimental analysis of intermediate disturbance and initial
floristic composition: decoupling cause and effect. Ecology
76:486–492.
Connell, J. H. 1961. The influence of interspecific competition
and other factors on the distribution of the barnacle
Chthamalus stellatus. Ecology 42:710–723.
Connell, J. H. 1978. Diversity in tropical rain forests and coral
reefs. Science 199:1302–1310.
Cowie, P. R., S. Widdicombe, and M. C. Austen. 2000. Effects
of physical disturbance on an estuarine intertidal community:
field and mesocosm results compared. Marine Biology 136:
485–495.
Crawley, M. J. 2004. Timing of disturbance and coexistence in a
species-rich ruderal plant community. Ecology 85:3277–3288.
Eggeling, W. J. 1947. Observations on the ecology of the
Budongo rain forrest, Uganda. Journal of Ecology 34:20–87.
Flöder, S., and U. Sommer. 1999. Diversity in planktonic
communities: an experimental test of the intermediate
disturbance hypothesis. Limnology and Oceanography 44:
1114–1119.
Fuentes, E. R., and F. M. Jaksic. 1988. The hump-backed
species diversity curve: why has it not been found among land
animals? Oikos 53:139–143.
Fukami, T., and P. J. Morin. 2003. Productivity–biodiversity
relationships depend on the history of community assembly.
Nature 424:423–426.
Garstecki, T., and S. A. Wickham. 2003. The response of
benthic rhizopods to sediment disturbance does not support
the intermediate disturbance hypothesis. Oikos 103:528–536.
Gosselin, L. A., and P. Y. Qian. 1997. Juvenile mortality in
benthic marine invertebrates. Marine Ecology Progress Series
146:265–282.
Grime, J. P. 1977. Evidence for the existence of three primary
strategies in plants and its relevance to ecological and
evolutionary theory. American Naturalist 111:1169–1194.
Huston, M. A. 1979. A general hypothesis of species diversity.
American Naturalist 113:81–101.
Huston, M. A. 1994. Biological diversity: the coexistence of
species. Cambridge University Press, Cambridge, UK.
Huxham, M., I. Roberts, and J. Bremner. 2000. A field test of
the intermediate disturbance hypothesis in the soft-bottom
intertidal. International Review of Hydrobiology 85:379–394.
Jara, V. C., J. H. S. Miyamoto, B. A. P. da Gama, M. Molis,
M. Wahl, and R. C. Pereira. 2006. Limited evidence of
interactive disturbance and nutrient effects on the diversity of
macrobenthic assemblages. Marine Ecology Progress Series
308:37–48.
Jonsson, P. R., K. M. Berntsson, and A. I. Larsson. 2004.
Linking larval supply to recruitment: flow-mediated control
of initial adhesion of barnacle larvae. Ecology 85:2850–2859.
February 2009
DISTURBANCE RATE AND COMMUNITY DYNAMICS
Lenz, M., M. Molis, and M. Wahl. 2004. Testing the
intermediate disturbance hypothesis: response of fouling
communities to various levels of emersion intensity. Marine
Ecology Progress Series 278:53–65.
Mackey, R. L., and D. J. Currie. 2001. The diversity–
disturbance relationship: is it generally strong and peaked?
Ecology 82:3479–3492.
McCabe, D. J., and N. J. Gotelli. 2000. Effects of disturbance
frequency, intensity, and area on assemblages of stream
macroinvertebrates. Oecologia 124:270–279.
Miller, T. E. 1982. Community diversity and interactions
between the size and frequency of disturbance. American
Naturalist 120:533–536.
Mitchell-Olds, T., and R. G. Shaw. 1987. Regression analysis of
natural selection: statistical inference and biological interpretation. Evolution 41:1149–1161.
Mittelbach, G. G., C. F. Steiner, S. M. Scheiner, K. L. Gross,
H. L. Reynolds, R. B. Waide, M. R. Willig, S. I. Dodson, and
L. Gough. 2001. What is the observed relationship between
species richness and productivity? Ecology 82:2381–2396.
Molino, J. F., and D. Sabatier. 2001. Tree diversity in tropical
rain forests: a validation of the intermediate disturbance
hypothesis. Science 294:1702–1704.
Osman, R. W. 1977. The establishment and development of a
marine epifaunal community. Ecological Monographs 47:37–
63.
Osman, R. W., R. B. Whitlatch, and R. N. Zajac. 1989. Effects
of resident species on recruitment into a community: larval
settlement versus post-settlement mortality in the oyster
Crassostrea virginica. Marine Ecology Progress Series 54:61–
73.
Padisak, J. 1993. The influence of different disturbance
frequencies on the species richness, diversity and equitability
of phytoplankton in shallow lakes. Hydrobiologia 249:135–
156.
Paine, R. T. 1966. Food web complexity and species diversity.
American Naturalist 100:65–75.
Patricio, J., F. Salas, M. A. Pardal, S. E. Jorgensen, and J. C.
Marques. 2006. Ecological indicators performance during a
re-colonisation field experiment and its compliance with
ecosystem theories. Ecological Indicators 6:43–57.
Peters, R. H. 1991. A critique for ecology. Cambridge
University Press, Cambridge, UK.
505
Pickett, S. T. A., and P. S. White. 1985. The ecology of natural
disturbance and patch dynamics. Academic Press, London,
UK.
Reynolds, C. S. 1995. The intermediate disturbance hypothesis
and its applicability to planktonic communities: comments
on the views of Padisak and Wilson. New Zealand Journal of
Ecology 19:219–225.
Roughgarden, J., S. Gaines, and H. Possingham. 1988.
Recruitment dynamics in complex life-cycles. Science 241:
1460–1466.
Schoener, T. W. 1972. Mathematical ecology and its place
among the sciences. I. The biological domain. Science 178:
389–391.
Sousa, W. P. 1979. Disturbance in marine intertidal boulder
fields: the nonequilibrium maintenance of species diversity.
Ecology 60:1225–1239.
Sousa, W. P. 1984. The role of disturbance in natural
communities. Annual Review of Ecology and Systematics
15:353–391.
Sousa, W. P. 2001. Natural disturbance and the dynamics of
marine benthic communities. Sinauer, Sunderland, Massachusetts, USA.
Steneck, R. S., S. D. Hacker, and M. N. Dethier. 1991.
Mechanisms of competitive dominance between crustose
coralline algae: an herbivore-mediated competitive reversal.
Ecology 72:938–950.
Svensson, J. R., M. Lindegarth, M. Siccha, M. Lenz, M. Molis,
M. Wahl, and H. Pavia. 2007. Maximum species richness at
intermediate frequencies of disturbance: consistency among
levels of productivity. Ecology 88:830–838.
Underwood, A. J., and M. J. Anderson. 1994. Seasonal and
temporal aspects of recruitment and succession in an
intertidal estuarine fouling assemblage. Journal of the
Marine Biological Association of the United Kingdom 74:
563–584.
Valdivia, N., A. Heidemann, M. Thiel, M. Molis, and M. Wahl.
2005. Effects of disturbance on the diversity of hard-bottom
macrobenthic communities on the coast of Chile. Marine
Ecology Progress Series 299:45–54.
Widdicombe, S., and M. C. Austen. 1999. Mesocosm investigation into the effects of bioturbation on the diversity and
structure of a subtidal macrobenthic community. Marine
Ecology Progress Series 189:181–193.
Paper I
Paper II
Paper
III III
PAPER
Paper IV
Paper V
On drawing conclusions from observations:
“Sir Bedevere: What makes you think she's a witch?
Paper VI
Peasant: Well, she turned me into a newt!
Sir Bedevere: A newt?
Peasant: [meekly after a long pause] ... I got better.
Crowd: [shouts] Burn her anyway!”
-the quest for the Holy Grail, Monty Python
Ecology, 91(10), 2010, pp. 3069–3080
Ó 2010 by the Ecological Society of America
Physical and biological disturbances interact differently
with productivity: effects on floral and faunal richness
J. ROBIN SVENSSON,1 MATS LINDEGARTH,
AND
HENRIK PAVIA
Department of Marine Ecology, University of Gothenburg, Tjärnö Marine Biological Laboratory, Strömstad 452 96 Sweden
Abstract. Physical and biological disturbances are ecological processes affecting patterns
in biodiversity at a range of scales in a variety of terrestrial and aquatic systems. Theoretical
and empirical evidence suggest that effects of disturbance on diversity differ qualitatively and
quantitatively, depending on levels of productivity (e.g., the dynamic equilibrium model). In
this study we contrasted the interactive effects between physical disturbance and productivity
to those between biological disturbance and productivity. Furthermore, to evaluate how these
effects varied among different components of marine hard-substratum assemblages, analyses
were done separately on algal and invertebrate richness, as well as richness of the whole
assemblage. Physical disturbance (wave action) was simulated at five distinct frequencies,
while biological disturbance (grazing periwinkles) was manipulated as present or absent, and
productivity was manipulated as high or ambient. Uni- and multivariate analyses both showed
significant effects of physical disturbance and interactive effects between biological
disturbance and productivity on the composition of assemblages and total species richness.
Algal richness was significantly affected by productivity and biological disturbance, whereas
invertebrate richness was affected by physical disturbance only. Thus, we show, for the first
time, that biological disturbance and physical disturbance interact differently with
productivity, because these two types of disturbances affect different components of
assemblages. These patterns might be explained by differences in the distribution (i.e., press
vs. pulse) and degree of selectivity between disturbances. Because different types of
disturbance can affect different components of assemblages, general ecological models will
benefit from using natural diverse communities, and studies concerned with particular subsets
of assemblages may be misleading. In conclusion, this study shows that the outcome of
experiments on effects of disturbance and productivity on diversity is greatly influenced by the
composition of the assemblage under study, as well as on the type of disturbance that is used
as an experimental treatment.
Key words: disturbance; diversity; dynamic equilibrium model (DEM); intermediate disturbance
hypothesis (IDH); marine assemblages; species richness; Tjärnö, Sweden.
INTRODUCTION
Disturbance has long been recognized as an important
structuring force in ecological communities (Darwin
1859). In the mid-1920s Cooper (1926) initiated a
discussion on possible effects of disturbance on succession and biodiversity, a discussion that is still ongoing
and has given rise to several hypotheses and models.
Among the most prominent is the intermediate disturbance hypothesis (IDH; Connell 1978), which predicts
that diversity will be high at intermediate levels of
disturbance and low at both extremes of a disturbance
continuum. Disturbance does, however, not only have
documented effects on the diversity of biological
communities, but also on evolutionary processes (Benmayor et al. 2008), biological invasions (Davis et al.
2000), and ecosystem functions (Cardinale and Palmer
Manuscript received 17 April 2009; revised 27 January 2010;
accepted 10 February 2010. Corresponding Editor: B. J.
Cardinale.
1 E-mail: [email protected]
2002). The ecological literature contains many examples
of agents and definitions of disturbance. One potentially
important distinction is that between agents of physical
and biological disturbance (McGuinness 1987, Sousa
2001, Svensson et al. 2007). Agents of physical
disturbance in previous studies include fire (Eggeling
1947), wind (Molino and Sabatier 2001), wave action
(McGuinness 1987), ice-scouring (Gutt and Piepenburg
2003), drifting logs (Dayton 1971), floods (Lake et al.
1989), sediment movement (Cowie et al. 2000), temperature (Flöder and Sommer 1999), desiccation (Lenz et
al. 2004), trawling (Tuck et al. 1998), pollution
(Benedetti-Cecchi et al. 2001), and even warfare
(Rapport et al. 1985). Biological disturbances are mainly
predation (Talbot et al. 1978) and grazing (Collins
1987), although some authors add trampling (Eggeling
1947) and burrowing (Guo 1996).
Because of the rich variety of agents of disturbance, a
number of definitions of disturbance have been proposed to make experiments commensurable (Sousa
2001). The definitions range from Grime’s (1977)
3069
3070
J. ROBIN SVENSSON ET AL.
straightforward partial or total destruction of biomass
to the more explicit definition of Pickett and White
(1985:356) in which disturbance is ‘‘. . . any relative
discrete event in time that disrupts ecosystems, community, or population structure and changes resources,
substrate availability, or the physical environment.’’
Among the more operational, and therefore more
commonly used definitions, is that of Sousa (1984:7),
in which disturbance not only kills or damages
individuals, but also ‘‘directly or indirectly creates an
opportunity for new individuals (or colonies) to become
established.’’ The recognition that disturbance creates
opportunities for recruitment is crucial, because without
new species recruiting into the freed space, disturbance
cannot increase diversity (Osman 1977, Collins et al.
1995, Huxham et al. 2000).
Disturbance has also been recognized as an important
component in multifactorial models in which interactions among community structuring processes are used
to predict diversity in natural communities. The
dynamic equilibrium model (DEM; Huston 1979,
Kondoh 2001) predicts high diversity at high levels of
disturbance when productivity is high, because a
stronger disturbance is then required to prevent
competitive exclusion. Similarly, diversity will be high
at low levels of disturbance when productivity is low,
because exclusion is then disrupted already by disturbances that are less frequent. The DEM has been tested
using either biological or physical agents of disturbance
in several experiments in aquatic as well as terrestrial
systems (e.g.,Turkington et al. 1993, Worm et al. 2002,
Jara et al. 2006, Svensson et al. 2007). The DEM and the
IDH have, however, received criticism from both
empirical and theoretical studies for being too simplistic
and based on weak theoretical grounds (Pacala and Rees
1998, Huxham et al. 2000, Shea et al. 2004). For
example, Chesson and Huntley (1997) showed that
disturbance may not diminish the importance of
competition, as predicted by Huston (1979), and that
indirect benefits of disturbance may fall short of the
direct negative effects. As a consequence of the
elucidation of the models’ weaknesses, their predictions
have been suggested, by empirical studies, to rely on a
number of prerequisites: competitive exclusion (Fuentes
and Jaksic 1988), a large regional species pool (Osman
1977), multiple stages in succession (Collins and Glenn
1997), and trade-offs between competition and colonization (Wilson 1994). These prerequisites, or assumptions, are in essence very similar to the flaws pointed out
in theoretical studies, i.e., that disturbance alone cannot
stabilize coexistence (Chesson and Huntly 1997) and the
underlying mechanisms for coexistence are in fact
nonlinear responses caused by trade-offs in life history
attributes (Amarasekare et al. 2004) and spatiotemporal
niches (Pacala and Rees 1998). In order to benefit from
this critique in a constructive way, Shea et al. (2004)
proposed that combining the suggestions of improvement from both empirical and theoretical studies will
Ecology, Vol. 91, No. 10
lead beyond mere descriptions of the hump-shaped
pattern. However, despite the many studies suggesting
critique against or improvements of the IDH and DEM,
few studies recognize the potentially large source of
variation caused by differences in the way an assemblage
is disturbed.
The manner in which a disturbance inflicts damage is
important because disturbances that are equal in extent
can nonetheless have significantly different effects on
diversity, depending on how the disturbance is distributed (Bertocci et al. 2005, Svensson et al. 2009). In a
theoretical study, Bender et al. (1984) identified two
different types of disturbance, pulse and press, and
evaluated their different effects on species’ interactions.
This distinction may also be useful for predictions of
patterns of diversity. Instantaneous alteration of species
number (pulse) and the sustained alteration of species
densities (press) are two clearly different mechanisms
that may still fall under the same general definition of
disturbance. Consequently, inconsistencies in outcomes
may occur if disturbances with dissimilar distributions
are treated without distinction in manipulative experiments (Svensson et al. 2009). Furthermore, it may be
important to make a general distinction between
biological and physical agents, because they commonly
differ in the degree of selectiveness (McGuinness 1987,
Sousa 2001). Such selectivity may be increasingly
complex under interactions with productivity, because
consumers often prefer prey with higher nutrient content
(Emlen 1966, Onuf et al. 1977, Pavia and Brock 2000).
There are many studies from various environments that
show interactive effects between biological disturbance
and productivity (see Proulx and Mazumder 1998),
while tests of the DEM using physical disturbance have
more variable outcomes (e.g., Turkington et al. 1993,
Jara et al. 2006, Svensson et al. 2007). No study, to our
knowledge, has explicitly contrasted differences between
biological and physical disturbances and their interactions with productivity. The only study to apply all three
factors simultaneously found a significant interaction
between biological, but not physical, disturbance and
productivity (Kneitel and Chase 2004). These findings
suggest that the choice of disturbance agent may
influence the outcome of experiments on interactive
effects of disturbance and productivity.
In this study we contrast the interactive effects
between a physical disturbance (i.e., wave action) and
productivity to those between a biological disturbance
(i.e., grazing) and productivity, in natural marine
benthic communities placed in mesocosms. We predict
that: (1) the biological disturbance will have a stronger
impact on macroalgal species and the physical disturbance will have a stronger impact on invertebrate species
and (2) the biological and the physical disturbances
therefore will interact differently with productivity.
Furthermore, we attempt to identify underlying mechanisms by investigating changes in composition of
assemblages among levels of biological and physical
October 2010
DISTURBANCES AND PRODUCTIVITY
disturbance and productivity, as well as evaluating
differences in species’ trade-offs in life history attributes
(sensu Pacala and Rees 1998, Shea et al. 2004). Hardsubstratum communities, such as the epilithic assemblages in this study, are generally considered to be
suitable for studies on disturbance, because sessile
species compete for the limiting resource space (e.g.,
Sousa 1984). More specifically, manipulative experiments conducted in this system have previously observed
the pattern predicted by the IDH, as well as strong
competition for space among macroalgae and invertebrates within one season (Svensson et al. 2007, 2009).
MATERIALS
AND
METHODS
Experimental assemblages
The marine hard-substratum assemblages that were
used to study the effects of physical and biological
disturbance and productivity were collected from semiexposed boulder fields in the Tjärnö archipelago at
depths of 0.5–1.5 m (58852.17 0 N, 1188.82 0 E). The
epilithic assemblages were composed exclusively of
sessile species, and the collected boulders, on which the
assemblages resided, were all of similar size (;20 cm
diameter). Common species in this system include red,
green, and brown macroalgae, as well as mussels,
tunicates, bryozoes, and sea anemones (Svensson et al.
2007, 2009; see Plate 1). The 43 different species present
in the experimental assemblages are listed in Table 1.
Associated mobile invertebrates (i.e., amphipods and
isopods) were not collected and included in the
experiment, since our aim was to add grazers of similar
density in assemblages subjected to the biological
disturbance treatment. Natural conditions of assemblages were maintained to a high extent by a constant
supply of unfiltered seawater from the Tjärnö bay,
allowing natural conditions of salinity, temperature,
food for filter feeders, nutrient availability, and propagules for colonization.
Experimental design
The experiment was carried out in the Ecotrone
mesocosm facility at Tjärnö Marine Biological Laboratory (TMBL), Strömstad, Sweden. One hundred boulders with epilithic assemblages were placed separately in
10-L plastic containers filled with seawater and covered
with mosquito nets (mesh size ¼ 1 mm) to prevent
grazers from escaping and/or entering. Water volume in
the containers was maintained using a constant flow
seawater supply, drawn from the bay adjacent to TMBL
at 0.5 m depth. The experimental manipulation started
on 5 June, had a duration of 17 weeks, and was
terminated on 10 October 2005.
The physical disturbance treatment was replicated five
times in each treatment combination, and the boulders
were subjected to one of five different frequencies: twice
per week (DP1), once per week (DP2), once every
second (DP3) or every fourth week (DP4), or left
undisturbed (DP0). The physical disturbance was caused
3071
by rolling each boulder by hand for one minute with
equal force in a tub filled with seawater and clean
boulders, in order to mimic effects of wave action in
boulder fields, at each disturbance event. In addition to
killing or damaging individuals, the rolling also facilitates recruitment by freeing substratum, and the
disturbance is therefore coherent with the definition by
Sousa (1984).
The productivity treatment consisted of two levels:
ambient (PR0) and increased (PR1). Bags with 100 g
slow-release fertilizer were attached to 50 containers
with boulders subjected to the increased productivity
treatment (PR1), while boulders of the ambient treatment (PR0) were not experiencing increased nutrient
availability. Fertilizer bags were changed every eighth
week in order to have constant nutrient release
throughout the experiment. Plantacote Depot 6-M
(5.7% NO3, 8.3% NH4, 9% P2O5, and 15% K2O;
AGLUKON, Düsseldorf, Germany) was used as
fertilizer due to its steady release rate (Worm et al.
2000).
The common periwinkle, Littorina littorea, is a very
abundant and important grazer in the Tjärnö archipelago (Wikström et al. 2006, Toth et al. 2007) where the
boulders were collected and was therefore used in the
grazing treatment. Littorina littorea periwinkles were
collected from the same areas as the boulders. In order
to achieve an ecologically relevant grazing pressure we
conducted a pilot study that suggested that ;10
periwinkles would be appropriate for the size of the
boulders in this experiment. Accordingly, 10 L. littorea
of similar size were added to each of the 50 boulders
subjected to the biological disturbance treatment (DB1),
and no periwinkles were added to the remaining 50
boulders, which were not experiencing grazing (DB0).
Sampling
Sampling of abundance and composition of the
communities was done at the end of the experiment
after 17 weeks of manipulation. Boulders were brought
into the laboratory submerged in seawater and also kept
under running seawater in the laboratory during the
entire sampling procedure. An area of 100 cm2 was
randomly chosen on each boulder for sampling, in order
to sample an equal area from all boulders regardless of
differences in actual area and size. Percent cover of bare
space and sessile species was then estimated at 5%
intervals using a 10 3 10 cm plastic grid, and a dissecting
microscope (magnification 123) was used for species
identification. Percent cover of species with a small
holdfast and wide thallus was estimated from the twodimensional projection of the organism on the panel.
Sessile epibionts were also accounted for. Thus, total
cover was allowed to exceed 100%.
Statistical analyses
The data obtained from the experiment were analyzed
with a three-way factorial ANOVA using Statistica 6.0
3072
J. ROBIN SVENSSON ET AL.
Ecology, Vol. 91, No. 10
TABLE 1. Abundance (percent cover, mean 6 SE) of sessile invertebrate and algal species present in the experimental communities
after 24 weeks, averaged over nutrient treatment for all levels of physical disturbance (DP0–DP4).
Species
Chlorophyceae
Chaetomorpha melagonium
Cladophora albido
Cladophora rupestris
Codium fragile
Enteromorpha prolifera
Enteromorpha intestinalis
Ulva lactuca
Phaeophyceae
Ahnfeltia plicata
Dumontia incrassata
Ectocarpus siliculosus
Fucus serratus
Fucus vesiculosus
Ralfsia tenuis
Sargassum muticum
Sphacelaria cirrosa
Sphacelaria plumosa
Rhodophyceae
Bonnemaisonia hamifera
Ceramium rubrum
Ceramium strictum
Chondrus crispus
Corallina officinalis
Cystoclonium purpureum
Furcellaria lumbricalis
Hildenbrandia rubra
Lithothamnion sp.
Osmundea truncata
Polysiphonia nigrescens
Polysiphonia urceolata
Polysiphonia violacea
Spermothamnion repens
DP0
0.5 6
0
0.9 6
0
0.4 6
1.4 6
1.65 6
DP1
0.11
0.54
0.26
0.67
1.09
0.4 6 0.26
1.3 6 1.25
0
1.0 6 1.0
21.55 6 7.59
0
0.5 6 0.50
7.35 6 3.38
0.25 6 0.25
1.6
0.1
0.1
19.6
0.05
2.75
1.25
4.1
15.6
1.5
1
2.45
0.15
6
6
6
6
6
6
6
6
6
6
6
6
6
0
1.50
0.07
0.07
7.04
0.05
1.56
1.25
2.52
3.97
1.50
0.78
1.99
0.08
DP2
DP3
0.6 6 0.11
0.1 6 0.07
1.2 6 0.57
0
0.45 6 0.26
3.7 6 3.24
1.45 6 0.60
0.85
0.3
0.55
0.25
1.75
2.8
0.75
0.24
0.25
0.26
0.25
1.25
1.54
0.34
0.8
0.05
0.3
0.05
1.45
5.4
1.1
2.5
0.05
0.05
1.55
0.75
0.05
0.1 6 0.07
0.25 6 0.10
0.1 6 0.07
0
0
0
0
1.75 6 0.71
0
0.8
0.55
0.05
0.05
6
6
6
6
6
6
0
5.1 6
0
1.38
0.05
0.05
1.03
0.55
0.05
2.61
0.25 6 0.25
0.1 6 0.07
0
17.1 6 8.03
0
1.35 6 1.25
0.25 6 0.25
2.6 6 1.11
8.15 6 2.25
0
0.05 6 0.05
1.75 6 1.22
0.25 6 0.25
0
7.25
0.05
2.85
7.65
0.05
0.1
0.05
6
6
6
6
6
6
6
0
0
0
6
0
6
0
6
6
0
6
6
6
0
3.46
0.05
1.11
2.23
0.05
0.05
0.05
DP4
6
6
6
6
6
6
6
0.25
0.05
0.11
0.05
0.60
3.18
0.45
1.6 6 0.54
0.25 6 0.10
0.5 6 0.26
0
0.9 6 0.33
4.15 6 1.59
1.9 6 0.70
6
6
6
6
0
0
0
1.2 6
0
0.55
0.26
0.05
0.05
0.05 6 0.05
0.05 6 0.05
0
0
0
0
0
0.9 6 0.33
0
0.57
0.05 6 0.05
0.1 6 0.07
0
7.35 6 2.94
0
0
0
4.25 6 1.59
8.9 6 3.63
0
0
0
0
0.05 6 0.05
0
0.1 6
0
2.0 6
0
0
0
3.05 6
3.7 6
0
0.1 6
0.1 6
0.05 6
0
0.07
0.58
1.04
1.18
0.07
0.07
0.05
Porifera
Leucosolenia botryides
0.05 6 0.05
0
0.05 6 0.05
0
0
Cnidaria
Dynamena pumila
Laomedea flexuosa
Metridium senile
0.05 6 0.05
0.05 6 0.05
0.5 6 0.26
0
0.05 6 0.05
0
0.05 6 0.05
0.1 6 0.07
0
0
0.25 6 0.10
0.05 6 0.05
0
0.65 6 0.34
0
Annelida
Pomatoceros triqueter
Spirorbis spirorbis
0.2 6 0.09
0.75 6 0.50
0.4 6 0.26
0.05 6 0.05
0.05 6 0.05
0.4 6 0.26
0.1 6 0.07
0.15 6 0.08
0.35 6 0.25
0.15 6 0.08
Crustacea
Balanus crenatus
Semibalanus balanoides
0.05 6 0.05
0.05 6 0.05
0.1 6 0.07
0
0.1 6 0.07
0
0.05 6 0.05
0
0.15 6 0.08
0
Mollusca
Acanthocardia sp.
Leptochiton sp.
Mytilus edulis
0.05 6 0.05
0.25 6 0.10
0.4 6 0.26
0
0
0.05 6 0.05
0
0
0
0
0
0
0
0
0
2.4 6 0.75
1.45 6 0.79
0.6 6 0.34
0.45 6 0.26
0.55 6 0.26
Bryozoa
Cryptosula pallasiana
Hemichordata
Ciona intestinalis
Total coverage
0.6 6 0.05
0
0
0
0
90.1 6 8.86
49.8 6 9.52
29.2 6 4.98
31.6 6 6.25
21.3 6 3.13
Note: The marine hard-substratum assemblages that were used to study the effects of physical and biological disturbance and
productivity were collected from semi-exposed boulder fields in the Tjärnö archipelago, Sweden.
(Statsoft, Tulsa, Oklahoma, USA) and with permutational multivariate analysis of variance (PERMANOVA) using the Permanova software (Anderson 2001).
The multivariate analyses were performed to reveal
whether differences were solely related to the number of
species or also dependent upon the particular species
present and their relative abundance. Hypotheses about
effects of main factors and interactions were tested using
October 2010
DISTURBANCES AND PRODUCTIVITY
3073
TABLE 2. ANOVA on total species richness, algal richness, and invertebrate richness.
Total richness
Algal richness
Invertebrate richness
Source
df
MS
P
MS
P
MS
P
PR
DB
DP
PR 3 DB
PR 3 DP
DB 3 DP
PR 3 DB 3 DP
Residual
1
1
4
1
4
4
4
80
171.28
61.76
16.82
16.74
0.79
8.54
3.26
3.83
0.000
0.000
0.003
0.040
0.935
0.073
0.496
132.49
99.34
2.41
8.88
1.31
4.55
1.44
3.15
0.000
0.000
0.552
0.097
0.797
0.228
0.767
2.49
4.44
7.30
1.24
1.94
1.07
1.75
1.20
0.154
0.058
0.000
0.313
0.178
0.473
0.223
Notes: Values in boldface indicate significance. Abbreviations are: PR, productivity; DB, biological disturbance; DP, physical
disturbance.
the following general linear model:
Xijkl ¼ l þ PRi þ DBj þ PRDBij þ DPk þ PRDPik
þ DBDPjk þ PRDBDPijk þ eijkl
where l is the overall mean, PRi is the ith level of
productivity, DBj is the jth level of biological disturbance, DPk is the kth level of physical disturbance, and
eijkl is a random deviation. Productivity, biological
disturbance, and physical disturbance are fixed factors
with two, two, and five levels, respectively. For the
univariate analysis post hoc tests of differences among
means were analyzed using the Student-Newman-Keuls
test (SNK), and t tests were used for the multivariate
analysis following the recommendations by Anderson
(2001). For all analyses data were tested for meeting the
assumptions of the statistical methods.
The hypothesis that physical and biological disturbance will interact differently with productivity is
supported if there is an interaction between productivity
and either biological disturbance or physical disturbance
exclusively or if a three-way interaction reveals different
patterns for different factorial combinations. Support
for the hypothesis that physical and biological disturbance will have different effects on different components
of assemblages is found if algal and invertebrate richness
are affected by either type of disturbance exclusively or
if the effects of the treatments show different patterns
(e.g., increase vs. decline) for algal compared to
invertebrate richness. In order to visualize patterns of
the multivariate tests and identify differences in species’
life history trade-offs, canonical analysis of principal
coordinates (CAP; Anderson and Willis 2003) was used
as a constrained ordination procedure on appropriate
terms found to be significant using PERMANOVA.
RESULTS
General observations
During the experiment a total of 13 sessile invertebrates and 31 algal species were observed in the
experimental communities. The macroalgae were not
only numerous in species, they also covered most of the
area in the experimental assemblages, and the most
abundant organisms were the red alga Chondrus crispus,
the brown alga Fucus vesiculosus, and the green algae
Enteromorpha intestinalis and Chaetomorpha melagonium. Unlike the algae, the ascidians, bryozoans,
crustaceans, molluscs, and sea anemones were usually
infrequent and had low percent cover in the assemblages
(Table 1).
Efficiency of the productivity treatment
In order to detect effects on productivity as a
consequence of the nutrient addition, differences in
percent cover among levels of nutrient availability were
tested. The ANOVA showed that there was a significant
effect of nutrient availability on total cover (F1,80 ¼ 26.9,
P , 0.001). Inspection of means (total cover for PR0
and PR1 were 29.4 6 4.0 and 58.7 6 4.0 [mean 6 SE],
respectively) showed that the total percent cover in
assemblages experiencing increased nutrient availability
was significantly higher than total cover under ambient
nutrient availability. This large difference in coverage
was mainly caused by increases in percent cover of algal
species (cover of algae for PR0 and PR1 were 27.9 6 4.0
and 56 6 4.0, respectively).
Effects of frequency of physical disturbance
Analysis of total species richness and community
composition at the end of the experiment showed that
there were significant effects of physical disturbance
(Tables 2 and 3 and Figs. 1a and 2), but no interactive
effects with productivity (Tables 2 and 3). A tendency
TABLE 3. Permutation multivariate analysis of variance
(PERMANOVA) with data transformed to presence/absence.
Source
df
MS
P
PR
DB
DP
PR 3 DB
PR 3 DP
DB 3 DP
PR 3 DB 3 DP
Residual
1
1
4
1
4
4
4
80
14 240
39 085
2899
4044
972
2313
1237
1746
0.001
0.001
0.034
0.028
0.944
0.159
0.829
Notes: Values in boldface indicate significance. Abbreviations are: PR, productivity; DB, biological disturbance; DP,
physical disturbance.
3074
J. ROBIN SVENSSON ET AL.
Ecology, Vol. 91, No. 10
FIG. 1. (a) Effects of frequency of physical disturbance on total species richness, algal richness, and invertebrate richness (mean
6 SE). (b) Effects of frequency of physical disturbance on mean cover of Chaetomorpha melagonium, Chondrus crispus,
Enteromorpha intestinalis, and Fucus vesiculosus averaged over nutrient and grazing treatments (mean 6 SE). The marine hardsubstratum assemblages that were used to study the effects of physical and biological disturbance and productivity were collected
from semi-exposed boulder fields in the Tjärnö archipelago, Sweden.
for an interaction between physical and biological
disturbance was found in the univariate analysis (P ¼
0.07; Table 2), indicating that the effects of biological
disturbance on richness were stronger in the presence of
physical disturbance (i.e., grazers had no impact on
undisturbed assemblages: richness for DP0DB0 and
DP0DB1 were 8.1 6 0.50 and 8.1 6 0.71, respectively).
Post hoc analysis on the multivariate test showed that
undisturbed assemblages were significantly different
from those experiencing higher levels of physical
disturbance (post hoc, DP0 ¼ DP1 6¼ DP2 6¼ DP3 6¼
DP4, at a ¼ 0.05) and in the univariate test all levels of
disturbance had lower species richness compared to the
undisturbed treatments (post hoc, DP0 . DP1 ¼ DP2 ¼
DP3 ¼ DP4, at a ¼ 0.05). Graphical examination of the
multivariate test using CAP revealed that undisturbed
assemblages and assemblages experiencing high levels of
physical disturbance were distributed at opposite sides
along the first axis with overlap around the origo (Fig.
2). Inspection of the mean cover of the most abundant
species suggested that there was a strong negative effect
on the cover of the perennial red alga Chondrus crispus
and the brown alga Fucus vesiculosus, whereas the
ephemeral green alga Enteromorpha intestinalis and
Chaetomorpha melagonium, capable of rapid growth
and colonization, were positively affected by disturbance
(Fig. 1b). Investigation of the effects of physical
disturbance on richness of algal and invertebrate species
showed that only the invertebrate species were significantly affected (Table 2). Graphical analysis also
showed that the number of invertebrate species was
higher in assemblages that were not subjected to
physical disturbance compared to assemblages that
experienced physical disturbance, whereas the algal
richness remained fairly constant over the disturbance
continuum (Fig. 1a).
October 2010
DISTURBANCES AND PRODUCTIVITY
3075
FIG. 2. Constrained canonical analysis of principal coordinates (CAP) plot on Bray-Curtis similarity comparing assemblages
among levels of physical disturbance (twice per week [DP1], once per week [DP2], once every second [DP3] or every fourth week
[DP4], or left undisturbed [DP0]). Values shown for d2 are the squared canonical correlation coefficients.
Effects of biological disturbance
There was a significant interaction between biological
disturbance and productivity for total species richness,
whereas algal richness was significantly altered by
biological disturbance and productivity independently
and the invertebrate species was not affected by either
factor (Table 2 and Fig. 3a–c). Inspection of means
showed that the effect of biological disturbance on total
richness was greater in assemblages in which nutrients
were not added (Fig. 3a). Thus richness was significantly
higher in assemblages with both grazers and nutrient
additions than in assemblages with only grazers (post
hoc; PR0, DB0 . DB 1; PR1, DB0 ¼ DB1; DB0, PR0 ,
PR1; DB1, PR0 , PR1; at a ¼ 0.05), which suggested
that the productivity treatment counteracted the effects
of biological disturbance on richness. The interaction
between biological disturbance and productivity was
also significant in the multivariate analysis (Table 3) and
post hoc analysis showed that all factor combinations
differed significantly (post hoc; PR0, DB0 6¼ DB1; PR1,
DB0 6¼ DB1; DB0, PR0 6¼ PR1; DB1, PR0 6¼ PR1; at a
¼ 0.05). Further graphical examination using CAP
revealed that, similar to the univariate analysis, there
was a more distinct separation between communities
with high compared to low productivity when grazers
were present (Fig. 4). Inspection of mean cover at the
species level showed that the green alga Chaetomorpha
melagonium was positively affected by biological disturbance, whereas Sphacelaria cirrosa and Enteromorpha
intestinalis decreased in the presence of grazers, and that
Hildenbrandia rubra was more frequent at high levels of
productivity (Fig. 4). Thus, in addition to significantly
affecting the number of species, the levels of biological
disturbance and productivity interactively determine
what species will be present in the assemblages.
DISCUSSION
In accordance with the first hypothesis, biological and
physical disturbances had different effects on the
marcoalgal and invertebrate species in the assemblages.
Biological disturbance significantly reduced only macroalgal richness, whereas physical disturbance exclusively
reduced invertebrate richness. Our second prediction,
that interactive effects of biological disturbance and
productivity on species richness were different from
those of physical disturbance and productivity, was also
supported. Increased productivity had a positive effect
on the number of algal, but not invertebrate, species.
Total species richness was negatively affected by
biological disturbance under ambient productivity, but
richness was generally larger and not affected by
biological disturbance when productivity was increased.
Physical disturbance, on the other hand, had a negative
effect on richness irrespective of whether productivity
was high or ambient. Similarly, the multivariate analyses
showed that increased productivity affected species
composition interactively with biological disturbance
but not with physical disturbance. Furthermore, evalu-
3076
J. ROBIN SVENSSON ET AL.
ation of responses of individual species to both types of
disturbance and productivity indicate that differences in
assemblage composition may be attributed to differences
in species’ trade-offs in life history characteristics.
The results of this study are consistent with those of
Kneitel and Chase (2004) and Svensson et al. (2007),
who found that the effect of physical disturbance on
richness was not affected by levels of productivity.
Svensson et al. (2007) suggested that their results could
be explained by the lack of a positive effect of the
productivity treatment on the dominant invertebrate
species. However, this cannot explain the lack of an
interactive effect in our experiment, since the dominant
organisms were macroalgae, which benefit directly from
increased nutrient availability (e.g., Worm et al. 2002).
The outcome is instead more likely explained by the
different effects of treatments on different components
of the assemblages. Physical disturbance affected the
number of invertebrate species, but not algal richness.
The invertebrate species were not affected by either
productivity or biological disturbance. Thus, the decimation of invertebrate species by physical disturbance
could not be counteracted by productivity, and,
consequently, there were no interactive effects between
the two treatments. There was, however, a tendency for
interactive effects between physical and biological
disturbance, although physical disturbance did not show
the quadratic function predicted by the IDH, which has
previously been observed in this environment (Svensson
et al. 2007, 2009). The lack of support for the IDH is
likely explained by the low rate of competitive exclusion,
despite the presence of competitive species, and the
tendency for interactive effects is possibly caused by
consumers inhibiting recolonization of free substratum
(e.g., Underwood 1980, Robles 1982). Unlike physical
disturbance, biological disturbance had interactive
effects with productivity on total species richness, which
is in accordance with previous studies from many
different environments (Proulx and Mazumder 1998).
The number of algal species was higher in assemblages
subjected to both grazing and productivity than in
assemblages subjected only to grazing. This indicates
that the positive effect of the productivity treatment
counteracted the negative effects of the biological
disturbance. Thus, it would appear that productivity,
rather than biological or physical disturbance, was the
factor promoting diversity in this experiment, in contrast
to previous manipulative experiments in the same system
(Svensson et al. 2007, 2009).
It has been shown in both theory (Chesson and
Huntly 1997) and practice (Huxham et al. 2000) that
disturbance, by either physical or biological agents, is
not in itself a diversity-promoting mechanism. Advancements have, however, been made in response to this
criticism by the suggestions of specific prerequisites that
are necessary for disturbance-mediated coexistence (e.g.,
Collins and Glenn 1997) and alternative theoretical
models, such as the ‘‘storage effect’’ (Roxburgh et al.
Ecology, Vol. 91, No. 10
FIG. 3. Interactive effects of productivity (PR0, PR1) and
biological disturbance, averaged over levels of physical
disturbance, on (a) total species richness, (b) algal richness,
and (c) invertebrate richness (mean 6 SE). Productivity 0 is
ambient productivity, and PR1 is increased productivity
(increased growth rates in the experimental assemblages
through the addition of nutrients: 100 g of slow-release fertilizer
per experimental assemblage).
2004) and ‘‘successional niche’’ (Pacala and Rees 1998).
The common ground in both the suggested prerequisites
and the alternative models is that disturbance can
promote coexistence in spatially homogeneous competitive environments with large species pools if the species
show differences in life history trade-offs. Coexistence
may then occur through creation of spatiotemporal
niches by disturbance, which may allow inferior species
competitive advantages over dominants in different
niches. Interpretation of results from multivariate
analyses within this framework may allow identification
of underlying mechanisms of disturbance–diversity
patterns through investigation of changes in community
October 2010
DISTURBANCES AND PRODUCTIVITY
3077
FIG. 4. Constrained canonical analysis of principal coordinates (CAP) plot on Bray-Curtis similarity comparing assemblages
among levels of productivity (high [PR1] or ambient [PR0]) and biological disturbance (present [DB1] or absent [DB0]). Treatment
combinations plotted are: productivity and biological disturbance (PR1DB1), solely biological disturbance (PR0DB1) or
productivity (PR1DB0), and controls (PR0DB0). Values shown for d2 are the squared canonical correlation coefficients.
composition and life history trade-offs (Shea et al.
2004). In this study, there were significant differences in
composition of species in the assemblages among levels
of physical disturbance. Biological disturbance and
productivity interactively affected the composition of
assemblages to form four distinct groups based on the
treatment combination (PR0DB0, PR0DB1, PR1DB0,
and PR1DB1). Although these four groups were all
significantly different, there were larger differences
among levels of productivity in the presence of grazers,
which is in accordance with the results from the
univariate analyses. Evaluation of responses of individual species allows for speculations on whether differences in species’ trade-offs in life history characteristics
may be the underlying cause for differences in assemblage composition. For instance, the green alga Chaetomorpha melagonium was positively affected by both
physical and biological disturbance, but not by productivity, possibly suggesting life history attributes toward
environmental tolerance rather than competition or fast
growth (Dial and Roughgarden 1998). Conversely,
Enteromorpha intestinalis, another green alga, was
positively affected by physical disturbance and productivity, but not by biological disturbance, thus indicating
a trade-off for fast growth compared to grazer tolerance
or competitive capacity (Petraitis et al. 1989). Two algal
species, Fucus vesiculosus and Chondrus crispus, had
coverage of up to 100% in individual undisturbed
assemblages and were negatively affected by physical
disturbance, while not benefiting from either productivity or biological disturbance. This shows that there were
species present in the assemblages that have characteristics of competitive dominants, even though the strong
competitive exclusion seen in this system in previous
experiments (Svensson et al. 2007, 2009) was not
apparent here. In accordance with the theories of Pacala
and Rees (1998), trade-offs in species’ life history
attributes allowed assemblages to differ in composition,
depending on the levels and treatment combinations of
biological and physical disturbance and productivity.
Thus, it would appear that combinations of different
regimes of disturbances and productivity enable different species to thrive under different conditions, ultimately maintaining regional and/or local coexistence.
The underlying mechanisms of the interaction between biological, but not physical, disturbance and
productivity, is most likely a combination of differences
in growth rates of species among levels of productivity
and mechanical differences in the way the damage is
inflicted on the assemblages. In addition, productivity
interacted with biological but not physical disturbance,
reflecting qualitative differences between the damage
exerted by the two types of disturbance. It has
previously been shown that similar disturbances can
3078
J. ROBIN SVENSSON ET AL.
Ecology, Vol. 91, No. 10
PLATE 1. (Main image) A typical boulder field on the west coast of Sweden. (Inset) Submerged boulders with benthic
assemblages composed of green, brown, and red macroalgae, as well as bryozoans, hydroids, mussels, tunicates, polychaetes, and
sea anemones. A color version of this plate is available in the Appendix. Photo credits: J. R. Svensson.
have different effects on diversity, depending on the way
the damage is distributed (Bertocci et al. 2005, Svensson
et al. 2009). In a theoretical study, Bender et al. (1984)
have described differences between two kinds of
mechanical disturbances, pulse and press. The herbivorous periwinkles in our experiment could be characterized as a press disturbance because they exert a
continuous small-scale reduction of biomass in algal
species. When biomass is slowly reduced, this effect can
more easily be counteracted by increased growth of the
affected organisms (Huston 1979, Kondoh 2001). The
productivity treatment probably had such a positive
effect, since diversity was higher in assemblages that
experienced grazing and nutrient addition compared to
assemblages subjected solely to grazing. The frequency
of physical disturbance shows characteristics similar to
pulse disturbance, which instantaneously kills, or
removes, a fraction of the community (Bender et al.
1984, Sousa 1984). Increased individual growth rate
cannot easily compensate for instantaneous loss of
individuals, which could explain the lack of interactive
effects between productivity and physical disturbance in
this experiment. In accordance with these arguments and
our results, Kneitel and Chase (2004), in the only
previous study that has tested for interactions of all
three factors, also found that biological disturbance
(predation), but not physical disturbance (drying), and
productivity interactively affected species richness.
Although Kneitel and Chase (2004) did not discuss
their treatments in terms of press and pulse disturbances,
the predatory mosquito larvae in their study could be
characterized as a press disturbance, whereas drying
every third or eighth day is similar to a pulse
disturbance. Thus, not only the selectivity, but also the
way that the damage caused by disturbance is applied,
may differ between agents of biological and physical
disturbance and determine the outcome of multifactorial
experiments.
In this study we have shown that the outcome of
experiments on disturbance and productivity is greatly
influenced by the type of disturbance that is used as a
treatment and also on the composition of the assemblage upon which the disturbance is inflicted. Previous
studies testing the DEM have commonly looked at
specific parts of natural communities, such as annelids
(Widdicombe and Austen 2001), macroalgae (Worm et
October 2010
DISTURBANCES AND PRODUCTIVITY
al. 2002), or periphytes (Cardinale et al. 2006), or used
artificial assemblages composed of bacteria and protozoans (Rashit and Bazin 1987), protozoans and rotifers
(Kneitel and Chase 2004), or bacteria and ciliates
(Scholes et al. 2005). The effects of treatments in such
experiments may be overestimated if the specific group
of species in the study are strongly affected or,
conversely, underestimated if species strongly affected
by the process in nature are not present in the
experimental assemblages. We have also shown, for
the first time, that an agent of biological disturbance and
an agent of physical disturbance interacted differently
with productivity due to their different effects on
different components of assemblages. Differences in
community composition and in responses of individual
species also indicate that the underlying mechanism for
the observed effects of both types of disturbance and
productivity may be traced back to species’ trade-offs in
life history attributes. In conclusion, our findings suggest
that experiments testing hypotheses on interactive effects
between disturbance and productivity, such as the
DEM, benefit from working with natural diverse
communities and should consider the ecological relevance of manipulative treatments in relation to both the
explanatory model and the system under study.
ACKNOWLEDGMENTS
This study was financially supported by MARICE (an
interdisciplinary research platform at the Faculty of Sciences,
Göteborg University) and by the Swedish Research Council
through contract number 621-2007-5779 to H. Pavia, as well
as by Formas through contracts 21.0/2004-0550 to H. Pavia
and 217-2006-357 to M. Lindegarth. We thank Malin
Karlsson for performing a large part of the fieldwork
and Anneli Lindgren for assistance with identification of
macroalgal species. We also thank two anonymous reviewers,
whose comments greatly improved an earlier version of the
manuscript.
LITERATURE CITED
Amarasekare, P., M. F. Hoopes, N. Mouquet, and M.
Holyoak. 2004. Mechanisms of coexistence in competitive
metacommunities. American Naturalist 164:310–326.
Anderson, M. J. 2001. A new method for non-parametric
multivariate analysis of variance. Austral Ecology 26:32–46.
Anderson, M. J., and T. J. Willis. 2003. Canonical analysis of
principal coordinates: a useful method of constrained
ordination for ecology. Ecology 84:511–525.
Bender, E. A., T. J. Case, and M. E. Gilpin. 1984. Perturbation
experiments in community ecology: theory and practice.
Ecology 65:1–13.
Benedetti-Cecchi, L., F. Pannacciulli, F. Bulleri, P. S. Moschella, L. Airoldi, G. Relini, and F. Cinelli. 2001. Predicting
the consequences of anthropogenic disturbance: large-scale
effects of loss of canopy algae on rocky shores. Marine
Ecology Progress Series 214:137–150.
Benmayor, R., A. Buckling, M. B. Bonsall, M. A. Brockhurst,
and D. J. Hodgson. 2008. The interactive effects of parasites,
disturbance, and productivity on experimental adaptive
radiations. Evolution 62:467–477.
Bertocci, I., E. Maggi, S. Vaselli, and L. Benedetti-Cecchi. 2005.
Contrasting effects of mean intensity and temporal variation
of disturbance on a rocky seashore. Ecology 86:2061–2067.
3079
Cardinale, B. J., H. Hillebrand, and D. F. Charles. 2006.
Geographic patterns of diversity in streams are predicted by a
multivariate model of disturbance and productivity. Journal
of Ecology 94:609–618.
Cardinale, B. J., and M. A. Palmer. 2002. Disturbance
moderates biodiversity–ecosystem function relationships:
experimental evidence from caddisflies in stream mesocosms.
Ecology 83:1915–1927.
Chesson, P., and N. Huntly. 1997. The roles of harsh and
fluctuating conditions in the dynamics of ecological communities. American Naturalist 150:519–553.
Collins, S. L. 1987. Interaction of disturbances in tallgrass
prairie: a field experiment. Ecology 68:1243–1250.
Collins, S. L., and S. M. Glenn. 1997. Intermediate disturbance
and its relationship to within- and between-patch dynamics.
New Zealand Journal of Ecology 21:103–110.
Collins, S. L., S. M. Glenn, and D. J. Gibson. 1995.
Experimental analysis of intermediate disturbance and initial
floristic composition: decoupling cause and effect. Ecology
76:486–492.
Connell, J. H. 1978. Diversity in tropical rain forests and coral
reefs. Science 199:1302–1310.
Cooper, W. S. 1926. The fundamentals of vegetational change.
Ecology 7:391–413.
Cowie, P. R., S. Widdicombe, and M. C. Austen. 2000. Effects
of physical disturbance on an estuarine intertidal community:
field and mesocosm results compared. Marine Biology 136:
485–495.
Darwin, C. 1859. On the origin of species by means of natural
selection. John Murray, London, UK.
Davis, M. A., J. P. Grime, and K. Thompson. 2000.
Fluctuating resources in plant communities: a general theory
of invasibility. Journal of Ecology 88:528–534.
Dayton, P. K. 1971. Competition, disturbance, and community
organization: the provision and subsequent utilization of
space in a rocky intertidal community. Ecological Monographs 41:351–389.
Dial, R., and J. Roughgarden. 1998. Theory of marine
communities: the intermediate disturbance hypothesis. Ecology 79:1412–1424.
Eggeling, W. J. 1947. Observations on the ecology of the
Budongo rain forest, Uganda. Journal of Ecology 34:20–87.
Emlen, J. M. 1966. The role of time and energy in food
preference. American Naturalist 100:611–617.
Flöder, S., and U. Sommer. 1999. Diversity in planktonic
communities: an experimental test of the intermediate
disturbance hypothesis. Limnology and Oceanography 44:
1114–1119.
Fuentes, E. R., and F. M. Jaksic. 1988. The hump-backed
species diversity curve: Why has it not been found among
land animals? Oikos 53:139–143.
Grime, J. P. 1977. Evidence for the existence of three primary
strategies in plants and its relevance to ecological and
evolutionary theory. American Naturalist 111:1169–1194.
Guo, Q. F. 1996. Effects of bannertail kangaroo rat mounds on
small-scale plant community structure. Oecologia 106:247–
256.
Gutt, J., and D. Piepenburg. 2003. Scale-dependent impact on
diversity of Antarctic benthos caused by grounding of
icebergs. Marine Ecology Progress Series 253:77–83.
Huston, M. A. 1979. A general hypothesis of species diversity.
American Naturalist 113:81–101.
Huxham, M., I. Roberts, and J. Bremner. 2000. A field test of
the intermediate disturbance hypothesis in the soft-bottom
intertidal. International Review of Hydrobiology 85:379–394.
Jara, V. C., J. H. S. Miyamoto, B. A. P. da Gama, M. Molis,
M. Wahl, and R. C. Pereira. 2006. Limited evidence of
interactive disturbance and nutrient effects on the diversity of
macrobenthic assemblages. Marine Ecology Progress Series
308:37–48.
3080
J. ROBIN SVENSSON ET AL.
Kneitel, J. M., and J. M. Chase. 2004. Disturbance, predator,
and resource interactions alter container community composition. Ecology 85:2088–2093.
Kondoh, M. 2001. Unifying the relationships of species richness
to productivity and disturbance. Proceedings of the Royal
Society B 268:269–271.
Lake, P. S., T. J. Doeg, and R. Marchant. 1989. Effects of
multiple disturbance on macroinvertebrate communities in
the Acheron River, Victoria. Australian Journal of Ecology
14:507–514.
Lenz, M., M. Molis, and M. Wahl. 2004. Testing the
intermediate disturbance hypothesis: response of fouling
communities to various levels of emersion intensity. Marine
Ecology Progress Series 278:53–65.
McGuinness, K. A. 1987. Disturbance and organisms on
boulders. 2. Causes of patterns in diversity and abundance.
Oecologia 71:420–430.
Molino, J. F., and D. Sabatier. 2001. Tree diversity in tropical
rain forests: a validation of the intermediate disturbance
hypothesis. Science 294:1702–1704.
Onuf, C. P., J. M. Teal, and I. Valiela. 1977. Interactions of
nutrients, plant growth and herbivory in a mangrove
ecosystem. Ecology 58:514–526.
Osman, R. W. 1977. The establishment and development of a
marine epifaunal community. Ecological Monographs 47:37–
63.
Pacala, S. W., and M. Rees. 1998. Models suggesting field
experiments to test two hypotheses explaining successional
diversity. American Naturalist 152:729–737.
Pavia, H., and E. Brock. 2000. Extrinsic factors influencing
phlorotannin production in the brown alga Ascophyllum
nodosum. Marine Ecology Progress Series 193:285–294.
Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989.
The maintenance of species diversity by disturbance.
Quarterly Review of Biology 64:393–418.
Pickett, S. T. A., and P. S. White. 1985. The ecology of natural
disturbance and patch dynamics. Academic Press, London,
UK.
Proulx, M., and A. Mazumder. 1998. Reversal of grazing
impact on plant species richness in nutrient-poor vs. nutrientrich ecosystems. Ecology 79:2581–2592.
Rapport, D. J., H. A. Regier, and T. C. Hutchinson. 1985.
Ecosystem behavior under stress. American Naturalist 125:
617–640.
Rashit, E., and M. Bazin. 1987. Environmental fluctuations,
productivity, and species-diversity: an experimental-study.
Microbial Ecology 14:101–112.
Robles, C. 1982. Disturbance and predation in an assemblage
of herbivorous diptera and algae on rocky shores. Oecologia
54:23–31.
Roxburgh, S. H., K. Shea, and J. B. Wilson. 2004. The
intermediate disturbance hypothesis: patch dynamics and
mechanisms of species coexistence. Ecology 85:359–371.
Ecology, Vol. 91, No. 10
Scholes, L., P. H. Warren, and A. P. Beckerman. 2005. The
combined effects of energy and disturbance on species
richness in protist microcosms. Ecology Letters 8:730–738.
Shea, K., S. H. Roxburgh, and E. S. J. Rauschert. 2004. Moving
from pattern to process: coexistence mechanisms under
intermediate disturbance regimes. Ecology Letters 7:491–508.
Sousa, W. P. 1984. The role of disturbance in natural
communities. Annual Review of Ecology and Systematics
15:353–391.
Sousa, W. P. 2001. Natural disturbance and the dynamics of
marine benthic communities. Sinauer, Sunderland, Massachusetts, USA.
Svensson, J. R., M. Lindegarth, and H. Pavia. 2009. Equal rates
of disturbance cause different patterns of diversity. Ecology
90:496–505.
Svensson, J. R., M. Lindegarth, M. Siccha, M. Lenz, M. Molis,
M. Wahl, and H. Pavia. 2007. Maximum species richness at
intermediate frequencies of disturbance: consistency among
levels of productivity. Ecology 88:830–838.
Talbot, F. H., B. C. Russel, and G. R. V. Anderson. 1978.
Coral reef fish communities: Unstable, high diversity
systems? Ecological Monographs 48:425–440.
Toth, G. B., M. Karlsson, and H. Pavia. 2007. Mesoherbivores
reduce net growth and induce chemical resistance in natural
seaweed populations. Oecologia 152:245–255.
Tuck, I. D., S. J. Hall, M. R. Robertson, E. Armstrong, and
D. J. Basford. 1998. Effects of physical trawling disturbance
in a previously unfished sheltered Scottish sea loch. Marine
Ecology Progress Series 162:227–242.
Turkington, R., E. Klein, and C. P. Chanway. 1993. Interactive
effects of nutrients and disturbance: an experimental test of
plant strategy theory. Ecology 74:863–878.
Underwood, A. J. 1980. Grazing and disturbance: an experimental analysis of patchiness in recovery from a severe storm
by the intertidal alga Hormosira banksii on rocky shores in
New South Wales. Journal of Experimental Marine Biology
and Ecology 231:291–306.
Widdicombe, S., and M. C. Austen. 2001. The interaction
between physical disturbance and organic enrichment: an
important element in structuring benthic communities.
Limnology and Oceanography 46:1720–1733.
Wikström, S. A., M. B. Steinarsdottir, L. Kautsky, and H. Pavia.
2006. Increased chemical resistance explains low herbivore
colonization of introduced seaweed. Oecologia 148:593–601.
Wilson, J. B. 1994. The intermediate disturbance hypothesis of
species coexistence is based on patch dynamics. New Zealand
Journal of Ecology 18:176–181.
Worm, B., H. K. Lotze, H. Hillebrand, and U. Sommer. 2002.
Consumer versus resource control of species diversity and
ecosystem functioning. Nature 417:848–851.
Worm, B., T. B. H. Reusch, and H. K. Lotze. 2000. In situ
nutrient enrichment: methods for marine benthic ecology.
International Review of Hydrobiology 85:359–375.
APPENDIX
A color version of Plate 1 showing a typical boulder field on the west coast of Sweden (Ecological Archives E091-213-A1).
Paper I
Paper II
PAPER
Paper
III IV
Paper IV
Paper V
”Du utför ditt slitgöra, match efter match, och
håller käften. Då kommer du dit du vill.”
- Håkan Mild
Paper VI
An Epic Journal 1(1), 2010, pp. 1-8
© 2010 by the El Grande Society
The Intermediate Disturbance Hypothesis predicts different
effects on species richness and evenness
J. Robin Svensson1, Mats Lindegarth, Per R. Jonsson, and Henrik Pavia
Department of Marine Ecology - Tjärnö, University of Gothenburg, SE-452 96 Strömstad, Sweden
Abstract. The Intermediate Disturbance Hypothesis (IDH) is among the most influential theories in ecology. Yet,
the aspect of biodiversity predicted to peak at intermediate disturbance is not explicitly defined, or even discussed,
in the literature. An issue that reaches beyond the scientific community, since the IDH also influences management
of national parks and reserves. As a consequence of this apparent lapse in disturbance theory, tests of the IDH and
later extensions are often based on unclear hypotheses and ambiguous measures of biodiversity. We used one
established model and one new, spatially explicit model to compare the responses to disturbance between the two
major aspects of biodiversity: species richness and the evenness of species abundance. Both models support the IDH
when biodiversity is measured as species richness, but predict that evenness instead increases monotonically with
increasing levels of disturbance. In order to investigate the generality of this discrepancy, we performed an extensive
meta-analysis of studies that use more than one measure of diversity and support the IDH. In accordance with the
predictions of the models, two-thirds of the published studies in the survey present different results for different
diversity measures. More specifically, comparisons between richness and evenness showed an even higher degree of
dissimilarity. Hence, based on the logic behind the underlying mechanism of the IDH, the predictions of our two
models and the meta-analysis, we argue that species richness is the most straightforward and appropriate response
variable in tests of the IDH and its associated models.
Key words: disturbance; diversity indices; evenness; ecological models; IDH; species richness.
INTRODUCTION
the last five years. The IDH, and the related dynamic
equilibrium model (DEM; Huston 1979, Kondoh 2001),
have received criticism in both empirical and theoretical
studies (Pacala and Rees 1998, Huxham et al. 2000, Shea
et al. 2004). Variable outcomes of empirical tests have led
to the awareness that the models rely on a number of
assumptions: competitive exclusion (Fuentes and Jaksic
1988), large regional species pool (Osman 1977), multiple
stages in succession (Collins and Glenn 1997) and tradeoffs between competition and colonization (Wilson 1994).
Similarly, theoretical studies argue that the underlying
mechanisms for coexistence are nonlinear responses to
competition (Amarasekare et al. 2004) and spatiotemporal
differences in niches (Pacala and Rees 1998). This critique
has, thus, led to a more thorough understanding of the
underlying coexistence mechanisms (Shea et al. 2004).
However, regardless of these improvements of the models,
tests of the IDH will inevitably also depend on the choice
of diversity measure and this has not yet received
attention.
Arguably, the most fundamental steps in science are the
formulation and testing of hypotheses (Popper 1959,
Underwood 1997, Quinn and Keough 2002). This involves
the logical linking of results from observations or
experiments to the hypothesis under test. Without clear and
explicit definitions of response variables, empirical studies
cannot unambiguously test the predictions of the model. In
ecological sciences, a central concept with such an elusive
definition is biodiversity (e.g. CBD Rio 1992). Because of
the great inconsistency in the way scientists define and
measure biodiversity (e.g. Hurlbert 1971), hypotheses
aiming to predict patterns of diversity, and the subsequent
tests, can be unclear. The well-known intermediate
disturbance hypothesis (IDH; Connell 1978) constitute,
together with its related models (Huston 1979, Miller
1982, Kondoh 2001), a keystone in ecological theory, but
it is also a case where many tests are based on unclear
predictions and ambiguous measures of biodiversity. The
original formulation of IDH predicts maximum
biodiversity to occur at an intermediate level of
disturbance due to coexistence of competitive dominants
and rapid colonizers, while diversity will be low at both
extremes due to competitive exclusion and local
extinction. The IDH is one of few well established
ecological theories and has influenced management and
conservation of nature reserves (Wootton 1998) as well as
grass- and pasture-land (Olff and Ritchie 1998). The
original work by Connell has received more than 3000
citations and continues to generate scientific papers at an
increasing rate, with over one third of all articles published
1
Considering the large body of literature on the IDH and its
later extensions it is surprising that it is still unclear what
aspects of species diversity that are predicted or measured.
This is even more remarkable, given that many other
aspects of the IDH have received ample attention, such as
alternative mechanisms underlying coexistence (Pacala
and Rees 1998), influence of characteristics of
communities (Fuentes and Jaksic 1988), interactive effects
of disturbances (Collins 1987), specific traits of individual
species (Haddad et al. 2008), temporal variation of
disturbance (Bertocci et al. 2005), how disturbance is
applied (Svensson et al. 2009) as well as the important
Corresponding author: J. Robin Svensson
Email: [email protected]
1
J. ROBIN SVENSSON ET AL.
apparent and unrecognized lapse in the tests of IDH and
related hypotheses/models jeopardizes the applicability of
disturbance-diversity theory in both basic and applied
ecology.
discussion on definitions of ecological disturbance (Pickett
and White 1985). In contrast, explicit discussions of how
to measure species diversity for appropriate tests of the
IDH is lacking in even the most extensive and influential
reviews on disturbance (cf. Sousa 1984, Mackey and
Currie 2001, Sousa 2001, Shea et al. 2004). Hence, it is not
surprising that there is no consensus on which measure of
diversity to use. A consequence of the lack of such a
consensus is that the IDH is tested with a plethora of
measures and indices of diversity, such as, Margalef’s
Richness, Simpson’s D, clonal diversity, functional
diversity, 1-lambda, and the more well known Shannon
index H’ (a.k.a the Shannon-Wiener or Shannon-Weaver
index, eqn 3; Shannon 1948, Shannon and Weaver 1963),
Pielou’s Evenness (eqn 4; Pielou 1966), and Species
Richness (i.e. number of species). It is likely that much of
the confusion about how to measure diversity stems from
the lack of clarity in the original formulations of IDH and
related models.
The confusion about what aspects of biodiversity that are
predicted by IDH led us to explore in detail the logical link
between IDH and biodiversity. We first show that models
of the IDH generate qualitatively different predictions for
different biodiversity measures. Secondly, we apply a
meta-analysis of the published tests of IDH to show that
support of IDH indeed depends on how diversity is
measured. Finally, we discuss the need for hypotheses
about mechanisms explaining the relationship between
disturbance intensity and specific measures of biodiversity.
METHODS
Model predictions of how disturbance affects species
richness and evenness
In the original article by Connell (1978), the word
diversity is frequently used without being defined in the
text, while species richness is the only specific measure of
diversity used in graphs and tables. This indicates that the
IDH primarily was intended to predict changes in the
number of species. The dynamic equilibrium hypothesis
(DEM; (Huston 1979, Kondoh 2001), an extension of the
IDH, predicts that the level of disturbance where
maximum diversity is observed will depend on the level of
productivity. Huston (1979) defines diversity as only
richness and evenness, rejecting various diversity indices,
but makes no distinction in predictions between effects of
disturbance on richness and evenness. Kondoh (2001)
discusses only species richness and does not consider
specific effects of productivity and disturbance on
evenness. In yet another extension of the IDH, on
differences in effects depending on distribution of
disturbance, Miller (1982) stated that the highest diversity
will occur at an intermediate rate of disturbance “…if
diversity is a measure of both species abundance and
number”. The addition of species abundance to the
hypothesis is, however, not explained or motivated in the
article. The only articles to our knowledge that discuss the
relevance of different diversity measures in tests of the
IDH, are those by Sommer (1995) and Weithoff et al.
(2001). Both articles mainly concern phytoplankton
communities, but while Weithoff et al. (2001) argues that
functional diversity, rather than species diversity, is the
most suitable response variable, Sommer (1995) maintains
that theories about coexistence principally predict changes
in species number, not abundances or diversity indices. In
their review, Shea et al. (2004) argue that the IDH cannot
be tested by studies using only one species, because the
IDH does not make predictions about abundances of
species. If the IDH does not predict differences in
abundance, not only single species studies but also tests
using compound diversity indices or the evenness of
species distributions as response variables are
inappropriate. The lack of an explicit definition of
diversity in the original presentation of the models has,
together with the variety of response variables that have
been used in subsequent experimental tests, made the
status of the IDH and related models unclear. This
The two models, spatially implicit and explicit, both
involve one-sided competition, occupancy as a function of
colonization ability, competitive strength and local
extinction, which increases with disturbance. A pool of 20
species was used in all modeling runs and colonization
rates of the ith species, ci, were modeled as ci=0.1/0.9i:
(Kondoh 2001). The first model (A) is a spatially implicit
patch-occupancy model proposed by Kondoh (2001) and
later used by Worm et al. (2002). The model was solved
using an ordinary differential equation solver in Matlab®
7.6 (MathWorks Inc) and colonization rates of the ith
species, ci, was modeled as ci=0.1/0.9i: (Kondoh 2001).
For more details see Kondoh (2001).
Since spatial relationships are well-known to affect
population- and community dynamics (e.g. Hassell et al.
1991, Molofsky 1994), the second modeling approach (B)
Table 1. a) Number of studies supporting the IDH and
the measures used in tests; Species richness (S), ShannonWiener index (H') and Pielous Evenness (J). b) Number of
studies where different measures of diversity differed in
their support of the IDH.
a) Measure of diversity
Supported
Tested
Species richness (S)
109
123
Shannon index (H')
49
56
Pielous Evenness (J)
16
38
Other measures/indices
27
32
Total number of studies
160
Dissimilarity among
b) measures
Studies testing >1 diversity
measure
Studies testing both
richness and evenness
2
Total
Dissimilar
60
42 (70%)
33
25 (76%)
THE IDH PREDICTS; RICHNESS ≠ EVENNESS
Fig. 1 Quadratic components for Species richness and Evenness, calculated through regression analyses after z-transformation
of data extracted from studies in the meta-analysis that used both measures (see methods), a) plotted together for each study for
comparisons within studies and b) plotted separately for general comparisons among measures. Support for the IDH, i.e. a humpshaped relationship, is indicated by high negative values of quadratic components for each measure of diversity.
For the examination of dissimilarities in outcomes of tests
of the IDH using more than one measure of diversity, we
specifically contrasted the number of species (i.e. species
richness) to the evenness of species distributions (i.e.
Pilou’s Evenness). This was done because (i) these two
measures are the key components in all indices of
diversity, and (ii) they represent two very different
components of the concept of diversity. In order to
compare differences in outcomes between species richness
and evenness we calculated the quadratic coefficient in
regression models describing the relationship between
disturbance and richness. The quadratic components were
calculated through regression analyses after ztransformation of data extracted from publications using
the graph digitizer GrabIt!© (Datatrend Software, Raleigh,
North Carolina, USA). The z-transformations were done in
order to allow comparisons between component values for
richness and evenness. Disturbance levels were normalized
between 0 and 1. Data extraction was possible in 28
studies from the articles reviewed (Fig. 1, Appendix S1).
The strength and polarity of the quadratic coefficient was
then plotted with species richness on the x-axis and
evenness on the y-axis. A high negative quadratic
coefficient indicates a strong hump-shaped relationship
between disturbance and diversity, thus supporting the
IDH.
was spatially explicit using a cellular automaton model
(e.g. Silvertown et al. 1992, Ermentrout and
Edelsteinkeshet 1993). The model was set up as a onedimensional universe with 100 cells. At each time step, a
proportion of the cells were subjected to a random, local
extinction. Thereafter, transition of each cell was achieved
either by competition or by recruitment. In the event of
competition, the state (i.e. the occupying species), a, of the
jth cell at time t+1, was determined by the state of
neighboring cells by:
(1)
a tj+1 = max([a tj−1
a tj
a tj +1])
Recruitment occurred with a probability of 0.1 in
unoccupied cells. The probability of recruitment of the ith
species was modeled as:
(2)
pi =
ci
20
∑c
i
i=1
Meta-analysis of diversity measures and support for
IDH
In the meta-analysis of previous tests of IDH and choice of
diversity measure we followed Shea et al. (2004), only
included studies reporting support for proceeded from the
list of papers provided in Shea et al. (2004) and
complemented it by searching in ISI Web of Science for
recent articles (2003-2010)citing Connell’s original paper
(Connell 1978). Of the over one thousand articles initially
reviewed, 143 studies in 132 publication were found which
reported support for the IDH (Table 1, Appendix S1).
Among these, 60 studies included more than one measure
of diversity (Table 1b, Appendix S1), mainly species
richness, Shannon’s index H’ (eqn 3; Shannon 1948,
Shannon and Weaver 1963), and evenness (eqn 4; Pielou
1966).
S
(3)
'
H max
= −∑
(4)
H'
E= '
H max
i =1
RESULTS
Model predictions of how disturbance affects species
richness and evenness
We applied a modeling approach to explore how
disturbance affects different measures of biodiversity. Here
we report effects on species richness and Pielou’s evenness
since these measures extract the two main components of
species-abundance distributions. Other compound indices
(e.g. Shannon’s H’) yielded intermediate results. The
response of species richness and evenness to different rates
of disturbance was explored with two different models,
one well-established (Kondoh 2001, Worm et al. 2002)
spatially implicit (model A) and one spatially explicit
(model B). Both models involve one-sided competition,
and occupancy of a particular species is a function of
colonization ability, competitive strength and local
1 1
ln
s s
3
J. ROBIN SVENSSON ET AL.
Fig. 2 Species richness (solid line) and Evenness (dashed line) as functions of magnitude of disturbance predicted
by a) the spatially implicit model A and b) the spatially explicit model B. Parameters in A are: productivity level=2,
extinction rate=0.05, threshold for local extinction=0.01, time steps=500. Data are presented as mean ± SE
consistent with the model predictions as species richness
yielded stronger hump-shaped relationships between
disturbance and diversity, than did evenness. This also
corresponds with our finding that two-thirds of the
published studies supporting the IDH present different
results for different diversity measures. Specifically, when
both species richness and evenness were used the
relationship between disturbance and diversity showed an
even higher degree of dissimilarity. It is surprising that the
use of different diversity measures and implications for
how to interpret tests of the IDH has not received any
previous attention. Mackey and Currie (2001) reviewed
tests of IDH and they found a hump-shaped relationship
for species richness, the Shannon index H’ and evenness
with disturbance in 19, 10 and 3 out of 85 analyzed
articles, respectively. They did not, however, discuss
possible causes of the different outcomes based on the
selected measure of diversity. This potentially confounding
factor in tests of the IDH is also neglected in the otherwise
excellent review by Shea et al. (2004), where they focus on
the mechanisms of coexistence underlying the humpshaped pattern.
extinction, which increases with disturbance (see
Methods). In both model A and model B richness shows a
unimodal hump-shaped pattern, whereas evenness is
asymptotically increasing with increasing disturbance
levels (Figs. 1a and b). Thus, both mathematical models of
the IDH predict qualitatively different effects on species
richness and evenness.
Meta-analysis of diversity measures and support for
IDH
Of the over one thousand articles initially reviewed, 143
studies in 132 publications reported support for the IDH
and 60 of these studies included more than one measure of
diversity (Table 1). In studies including more than one
measure of diversity the support for the IDH was often
inconsistent between different diversity measures. When
outcomes among all measures are compared they show
dissimilar support in 70% of the cases (Table 1b). In
comparisons specifically contrasting outcomes among tests
using both richness and evenness, these two measures
differed in their support in over 75 % of the cases. The
support for the IDH in 28 previous studies using both
species richness and evenness as biodiversity measures is
shown in Fig. 2. Negative values of the quadratic
component in the statistical model of the effect of
disturbance on diversity indicate a hump-shaped (unimodal
peak) relationship and thus support for the IDH. Only
when diversity is measured as species richness is there a
consistent hump-shaped relation supporting IDH (Fig. 2a),
and the cumulative distributions in Fig. 2b show that the
range of the quadratic coefficients is narrower for the tests
using species richness compared to when evenness is used.
Why then do different measures of diversity differ in
response to disturbance? According to the original
formulation of the IDH by Connell (1978), it is the number
of species that will increase when disturbance prevents
competitive exclusion to occur and allows new species to
colonize, up to a certain point when disturbance becomes
too severe for species to persist (Eggeling 1947, Osman
1977, Connell 1978). Thus, the prediction that the number
of species should show a hump-shaped response to
disturbance rests on logic arguments, and the hypothesis is
easily tested with species richness as the most evident
response variable. It is, however, less logical that this
prediction should automatically also apply to various
diversity indices, such as Shannon’s H’ and evenness, that
also consider how abundance is distributed among species.
Species do not need to be more evenly distributed at
intermediate disturbance just because the number of
species is large. If the predictions are logical for the
number of species, but not for species-abundance
Discussion
We here show that an established model as well as a new,
spatially explicit model only support IDH when biodiversity is measured as species richness. Both models
predict that evenness instead increases monotonically with
increasing levels of disturbance. Our extensive metaanalysis of published empirical tests of the IDH is also
4
THE IDH PREDICTS; RICHNESS ≠ EVENNESS
contaminated with mine tailings (Gregory and Bradshaw
1965). However, because of the general lack of discussion
of what is predicted about evenness, we recommend that
studies choosing evenness as the response variable in tests
of the IDH and DEM should present logical arguments, a
priori, to why the predicted pattern can be observed in
natural communities.
distributions (Shea et al. 2004), there is no clear reason for
H’ to be a preferable index in disturbance studies, as has
previously been suggested (Worm et al. 2002). On a more
general level, Stirling and Wilsey (2001) argued that H’
was the best measure of diversity because it considers both
the separate effects of richness and evenness and also their
interrelations. Although this may be advantageous under
certain circumstances, it may be less so in efforts to
unravel specific changes in diversity, because the
underlying ecological process or mechanism causing
changes in H’ can be traced back to effects on either
richness or evenness (Hurlbert 1971). Thus, a more
interesting and challenging question is why patterns of
richness and evenness differ, and if a logical pattern
between evenness and disturbance can be conceived within
the framework of the IDH.
In conclusion, we argue that the logic behind the
underlying mechanism of the IDH, the predictions of our
two models and the meta-analysis, all suggest that species
richness is the most straightforward and appropriate
response variable in tests of the IDH and its associated
models. Furthermore, since the IDH is also utilized in
management of marine and terrestrial national reserves and
parks (e.g. Yellowstone National Park, USA), a consensus
on appropriate response variables will have benefits
reaching beyond the scientific community.
The IDH relies on the assumption that one or a few species
will dominate the community in the absence of disturbance
(Fuentes and Jaksic 1988, Collins and Glenn 1997,
Svensson et al. 2007). An uneven distribution of species is
therefore to be expected at low levels of disturbance,
which is also commonly observed in marine and terrestrial
field experiments (Eggeling 1947, Molis et al. 2003, Lenz
et al. 2004b, a). According to the compensatory mortality
hypothesis(Janzen 1970), mortality from causes unrelated
to the competitive interactions falls heaviest on whichever
species that ranks highest in competitive ability. The
reduction of a highly abundant basal species (i.e.
dominant) by disturbance may lead to colonization of new
species in the free space (e.g. Connell 1978).
Consequently, both the number of species and the
evenness of species distributions are likely to initially
increase following a disturbance in an already uneven
community. In accordance with this, the presented
mathematical models (Figs. 2a and b), as well as previous
field experiments from both marine and terrestrial systems
(Vujnovic 2002, Kimbro and Grosholz 2006), support
these patterns.
Acknowledgements
This study was financially supported by MARICE (an
interdisciplinary research platform at the Faculty of
Sciences, Göteborg University), by the Swedish Research
Council through contract no. 621-2007-5779 to HP and
621-2008-5456 to PRJ, and by Formas through contracts
21.0/2004-0550 to HP and 217-2006-357 to ML.
References
Amarasekare, P., M. F. Hoopes, N. Mouquet, and M.
Holyoak. 2004. Mechanisms of coexistence in competitive
metacommunities. American Naturalist 164:310-326.
Bertocci, I., E. Maggi, S. Vaselli, and L. Benedetti-Cecchi.
2005. Contrasting effects of mean intensity and temporal
variation of disturbance on a rocky seashore. Ecology
86:2061-2067.
Collins, S. L. 1987. Interaction of Disturbances in Tallgrass
Prairie - a Field Experiment. Ecology 68:1243-1250.
Collins, S. L. and S. M. Glenn. 1997. Intermediate
disturbance and its relationship to within- and betweenpatch dynamics. New Zealand Journal of Ecology 21:103110.
Connell, J. H. 1978. Diversity in tropical rain forests and
coral reefs. Science 199:1302 – 1310.
Eggeling, W. J. 1947. Observations on the ecology of the
Budongo rain forrest, Uganda. Journal of ecology 34:20 –
87.
Ermentrout, G. B. and L. Edelsteinkeshet. 1993. Cellular
Automata Approaches to Biological Modeling. Journal of
Theoretical Biology 160:97-133.
Fuentes, E. R. and F. M. Jaksic. 1988. The Hump-Backed
Species Diversity Curve: Why Has It Not Been Found
among Land Animals? Oikos 53:139-143.
Gregory, R. P. G. and A. D. Bradshaw. 1965. Heavy metal
tolerance in populations of Agrostis tenuis Sibth. and other
grasses. New Phytologist 64:131-&.
Haddad, N. M., M. Holyoak, T. M. Mata, K. F. Davies, B. A.
Melbourne, and K. Preston. 2008. Species' traits predict the
effects of disturbance and productivity on diversity.
Ecology Letters 11:348-356.
Hassell, M. P., H. N. Comins, and R. M. May. 1991. Spatial
Structure and Chaos in Insect Population-Dynamics. Nature
353:255-258.
Hurlbert, S. H. 1971. The nonconcept of species diversity: a
critique and alternative parameters. Ecology 52:577 – 586.
Following the plausible increase in evenness at low to
intermediate disturbance levels, logical predictions and
patterns are less clear for high levels of disturbance.
Commonly, high disturbance is associated with larger
areas of free substratum (e.g. Paine and Levin 1981, Miller
1982, Valdivia et al. 2005). This hinders dominants to
achieve large abundances, or even exist, and allows the
few rapid colonizers able to withstand the disturbance to
settle in the free space. These colonizers are all likely to
initially be low in abundance, which might lead to a high
level of evenness despite low total coverage in
assemblages at high levels of disturbance (Kimbro and
Grosholz 2006), which is in accordance with our model
predictions. Maximum evenness at intermediate levels of
disturbance has, however, been found in a few
manipulative experiments (Lenz et al. 2004b, a). Logical
arguments explaining the subsequent decrease in evenness
are not given in these studies, possibly because
clarification of patterns thought to conform to an existing
model seemed redundant. One possible explanation for
low evenness at intense disturbance is caused by the
dominance of a few disturbance specialists, where a wellknown example is metal-tolerant grasses on soils
5
J. ROBIN SVENSSON ET AL.
Huston, M. A. 1979. A general hypothesis of species
diversity. American Naturalist 113:81 – 101.
Huxham, M., I. Roberts, and J. Bremner. 2000. A field test of
the intermediate disturbance hypothesis in the soft-bottom
intertidal. International Review of Hydrobiology 85:379394.
Janzen, D. H. 1970. Herbivores and the Number of Tree
Species in Tropical Forests. . American Naturalist
104:501-528.
Kimbro, D. L. and E. D. Grosholz. 2006. Disturbance
influences oyster community richness and evenness, but not
diversity. Ecology 87:2378.
Kondoh, M. 2001. Unifying the relationships of species
richness to productivity and disturbance. Proceedings of the
Royal Society of London Series B-Biological Sciences
268:269-271.
Lenz, M., M. Molis, and M. Wahl. 2004a. Experimental test
of the intermediate disturbance hypothesis: frequency
effects of emersion on fouling communities. Journal of
Experimental Marine Biology and Ecology 305:247-266.
Lenz, M., M. Molis, and M. Wahl. 2004b. Testing the
intermediate disturbance hypothesis: response of fouling
communities to various levels of emersion intensity.
Marine Ecology-Progress Series 278:53-65.
Mackey, R. L. and D. J. Currie. 2001. The diversitydisturbance relationship: Is it generally strong and peaked?
Ecology 82:3479-3492.
Miller, T. E. 1982. Community diversity and interactions
between the size and frequency of disturbance. American
Naturalist 120:533 – 536.
Molis, M., M. Lenz, and M. Wahl. 2003. Radiation effects
along a UV-B gradient on species composition and
diversity of a shallow-water macrobenthic community in
the western Baltic. Marine Ecology-Progress Series
263:113-125.
Molofsky, J. 1994. Population-Dynamics and PatternFormation in Theoretical Populations. Ecology 75:30-39.
Olff, H. and M. E. Ritchie. 1998. Effects of herbivores on
grassland plant diversity. Trends in Ecology & Evolution
13:261-265.
Osman, R. W. 1977. The establishment and development of a
marine epifaunal community. Ecological Monographs
47:37 – 63.
Pacala, S. W. and M. Rees. 1998. Models suggesting field
experiments to test two hypotheses explaining successional
diversity. American Naturalist 152:729-737.
Paine, R. T. and S. A. Levin. 1981. Intertidal landscapes:
disturbance and the dynamics of pattern. Ecological
Monographs 51:145 – 178.
Pickett, S. T. A. and P. S. White. 1985. The ecology of
natural disturbance and patch dynamics. Academic press,
London.
Pielou, E. C. 1966. Species-diversity and pattern-diversity in
the study of ecological succession. J. Theoret. Biol. 10:370
- 383.
Popper, K. R. 1959. The logic of scientific discovery. Basic
Books, New York.
Quinn, G. P. and M. J. Keough. 2002. Experimental design
and data analysis for biologists. Cambridge university
press, Cambridge.
Shannon, C. E. 1948. A Mathematical Theory of
Communication. The Bell System Technical Journal Vol.
27:379–423, 623–656.
Shannon, C. E. and W. Weaver. 1963. The mathematical
theory of communication. University of Illinois Press,
Urbana.
Shea, K., S. H. Roxburgh, and E. S. J. Rauschert. 2004.
Moving from pattern to process: coexistence mechanisms
under intermediate disturbance regimes. Ecology Letters
7:491-508.
Silvertown, J., S. Holtier, J. Johnson, and P. Dale. 1992.
Cellular Automaton Models of Interspecific Competition
for Space - the Effect of Pattern on Process. Journal of
Ecology 80:527-534.
Sommer, U. 1995. An experimental test of the intermediate
disturbance hypothesis using cultures of marine
phytoplankton. Limnology and Oceanography 40:12711277.
Sousa, W. P. 1984. The role of disturbance in natural
communities. Annual Review of Ecology and Systematics
15:353 – 391.
Sousa, W. P. 2001. Natural disturbance and the dynamics of
marine benthic communities. Sinauer, Sunderland,
Massachusetts, USA.
Stirling, G. and B. Wilsey. 2001. Empirical relationships
between species richness, evenness, and proportional
diversity. American Naturalist 158:286-299.
Svensson, J. R., M. Lindegarth, and H. Pavia. 2009. Equal
rates of disturbance cause different patterns of diversity.
Ecology 90:496-505.
Svensson, J. R., M. Lindegarth, M. Siccha, M. Lenz, M.
Molis, M. Wahl, and H. Pavia. 2007. Maximum species
richness at intermediate frequencies of disturbance:
Consistency among levels of productivity. Ecology 88:830838.
Underwood, A. J. 1997. Experiments in ecology- their logical
design and interpretation using analysis of variance
Cambridge University Press, Cambridge.
Valdivia, N., A. Heidemann, M. Thiel, M. Molis, and M.
Wahl. 2005. Effects of disturbance on the diversity of hardbottom macrobenthic communities on the coast of Chile.
Marine Ecology-Progress Series 299:45-54.
Weithoff, G., N. Walz, and U. Gaedke. 2001. The
intermediate, disturbance hypothesis - species diversity or
functional diversity? Journal of Plankton Research
23:1147-1155.
Wilson, J. B. 1994. The Intermediate Disturbance Hypothesis
of Species Coexistence Is Based on Patch Dynamics. New
Zealand Journal of Ecology 18:176-181.
Wootton, J. T. 1998. Effects of disturbance on species
diversity: A multitrophic perspective. American Naturalist
152:803-825.
Worm, B., H. K. Lotze, H. Hillebrand, and U. Sommer. 2002.
Consumer versus resource control of species diversity and
ecosystem functioning. Nature 417:848-851.
Vujnovic, K. 2002. Predicting plant species diversity in
response to disturbance magnitude in grassland remnants of
central Alberta. Canadian Journal of Botany 80:504.
6