Content uploaded by Tano Gutiérrez-Cánovas
Author content
All content in this area was uploaded by Tano Gutiérrez-Cánovas on May 22, 2017
Content may be subject to copyright.
Biological invasion modifies the co-occurrence
patterns of insects along a stress gradient
Jos
e Antonio Carbonell*
,1
, Josefa Velasco
1
, Andr
es Mill
an
1
, Andy J. Green
2
,
Cristina Coccia
3
, Simone Guareschi
1
and Cayetano Guti
errez-C
anovas
4
1
Department of Ecology and Hydrology, Regional Campus of International Excellence ‘Campus Mare Nostrum’,
University of Murcia, Murcia, Spain;
2
Department of Wetland Ecology, Do~
nana Biological Station (EBD-CSIC), Am
erico
Vespucio 26, 41092 Seville, Spain;
3
Departamento de Ecolog
ıa, Facultad de Ciencias Biol
ogicas, Pontificia
Universidad Cat
olica de Chile, Santiago 3542000, Chile; and
4
Catchment Research Group, Cardiff University, School of
Biosciences, The Sir Martin Evans Building, Museum Avenue, Cardiff CF10 3AX, UK
Summary
1. Biological invasions have become one of the most important drivers of biodiversity loss and
ecosystem change world-wide. However, it is still unclear how invasions may interact with
local abiotic stressors, which are expected to increase as global change intensifies. Further-
more, we know little about the response to biological invasions of insects, despite their dispro-
portionate contribution to global animal biodiversity.
2. The aim of the present work is to investigate the impact of an invasive aquatic insect on the
co-occurrence patterns of native species of insects along a salinity gradient, and determine
which assembly rules are driving these patterns.
3. First, we characterised the habitat specialisation and functional niches of each species from
physiological and biological traits, respectively, and their degree of overlap. Second, we used
field data to compare the co-occurrence patterns of native and invasive species in invaded and
non-invaded areas of southern Iberia and northern Morocco. Finally, we tested if habitat fil-
tering or niche differentiation assembly rules mediate their co-occurrence.
4. In non-invaded areas, habitat filtering drives habitat segregation of species along the salinity
gradient, with a lower contribution of niche differentiation. The presence of the invasive insect
modifies the distribution and co-occurrence patterns of native species. In invaded areas, niche
differentiation seems to be the main mechanism to avoid competition among the invasive and
native species, enabling coexistence and resource partitioning.
5. The combined study of functional niche similarity and abiotic stressor tolerance of invasive
and native species can improve our understanding of the effects of invasive species along
abiotic stress gradients. This approach may increase our capacity to predict the outcomes of
biological invasion in a global change context.
Key-words: alien species, coexistence, community assembly, Corixidae, habitat filtering, niche
differentiation, predictive ecology
Introduction
Biological invasion is one of the most important drivers of
biodiversity loss and ecosystem change (Simberloff et al.
2013). So far, the study of the invasiveness of alien species
has been focused mainly on isolated traits (Van de Meutter
et al. 2010; Carbonell et al. 2016) and the specific ecosys-
tem impacts induced by alien species (Gherardi & Acquis-
tapace 2007; Gallardo et al. 2015). However, it remains
unclear how ecological and functional similarity between
native and alien species may influence the success of bio-
logical invasions, especially in the presence of intense envi-
ronmental stressors.
Abiotic stressors that occur at unprecedented rates or
magnitudes, or those that are novel for the regional pool
(i.e. anthropogenic stressors), may enhance biological inva-
sion success (e.g. MacDougall & Turkington 2005)
through the extinction of native species (e.g. Guti
errez-
C
anovas et al. 2013). Such depauperate communities are
more vulnerable to invasion due to the reduction of niche
overlap and, in turn, the percentage of competitive
interactions (Maestre et al. 2009). This scenario could be
*Correspondence author. E-mail: joseantonio.carbonell@um.es
©2017 The Authors. Functional Ecology ©2017 British Ecological Society
Functional Ecology 2017 doi: 10.1111/1365-2435.12884
especially relevant for invaders with a generalised niche
(Davidson, Jennions & Nicotra 2011). Furthermore,
historically persistent environmental filters (i.e. those of
natural origin) are assumed to sort species from the
regional pool in a different manner. Along gradients of
environmental stress, the reduction in diversity is usually
accompanied by a turnover of specialist species (Guti
errez-
C
anovas et al. 2013). A specialised niche occupying a
specific portion of the stress gradient is a result of the
adaptation process over a long time, leading to specific
traits, with associated trade-offs (e.g. Carbonell, Mill
an &
Velasco 2012b). Although some correlative studies found
that natural stress might constrain biological invasions
(e.g. Gerhardt & Collinge 2007), the ecological mecha-
nisms behind these empirical observations remain poorly
understood.
Theories rooted in the community assembly framework
proposed that both neutral (e.g. demography, dispersal
limitation) and niche-based (e.g. habitat filtering and niche
differentiation) processes may explain the patterns of spe-
cies co-occurrence (Weiher et al. 2011). Regarding niche-
based processes, abiotic and biotic pressures can modify
species occurrence through different mechanisms (Fig. 1).
Intense abiotic stress may increase habitat filtering (i.e.
segregation of species into different habitats) by selecting
certain non-random trait combinations that would be
more successful under such stressful conditions (Shipley,
Vile & Garnier 2006) (Fig. 1b). As a result, habitat filter-
ing reduces taxonomic and functional diversities and
increases the functional similarity of the remaining species
more than expected by chance (Weiher et al. 2011;
Guti
errez-C
anovas et al. 2015). Moreover, the stress gradi-
ent hypothesis predicts a decrease in the strength of compe-
tition among species as abiotic stress increases, which is
contingent on species life histories and the type of focal
stress (Maestre et al. 2009). Under low-stress conditions,
niche differentiation of coexisting species within a given
habitat could play an important role for stabilising com-
munities (Fig. 1c). This mechanism enhances differences in
the way that species exploit resources or habitats, facilitat-
ing the coexistence of ecologically diverse species (Chesson
2000). In this context, invasive species may compete
strongly with ecologically related native species (Violle
et al. 2011), affecting the performance and coexistence pat-
terns of native species through resource limitation and
interference (Chase & Leibold 2003). When competition
occurs, the degree of resource/niche overlap among inva-
sive and native species can determine the outcome of such
interactions (De Roos et al. 2008). In general, when two
species have a similar niche, the weaker competitors are
Fig. 1. Predicting species co-occurrence from niche-based processes and a conceptual model highlighting the importance of niche differen-
tiation (ND) and habitat filtering (HF) assembly rules in competitive communities. (a) Functional niche trait space: species traits vary
along different axes of specialisation (functional axes 1, 2, n) describing independent dimensions of a species’ functional niche, as
represented by circles with different patterns within the trait space. Community assembly rules: HF and ND processes determine species
co-occurrence, richness and abundance through convergence or divergence of functional traits: HF (b) sorts species via the presence of
certain traits which tend to be dominant. Increased HF results in species segregation along environmental gradients and progressively
lower species richness and higher relative abundance; ND (c) selects species with dissimilar traits, increases diversity and stabilises the
community through resource partitioning. This assembly rule favours species coexistence in communities with higher species richness and
lower relative abundance of a given species.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
2J. A. Carbonell et al.
often excluded from the community (Reitz & Trumble
2002). Therefore, invaders could alter the assembly rules of
native communities along abiotic gradients, depending on
their niche similarities and their tolerance to the abiotic
stressor (Iacarella et al. 2015).
Despite recent improvement in the understanding of
which single traits enhance invasive capacity and impact
on aquatic ecosystems (e.g. Statzner, Bonada & Dol
edec
2008; Gallardo et al. 2015; Carbonell et al. 2016), we
know little about how ecological and functional similarity
influences the coexistence of insect species and their
response to biological invasion and environmental change.
Improving our knowledge is crucial considering the dispro-
portionate contribution of insects to global biodiversity.
Applying the community assembly framework to well-
studied insects undergoing biological invasions along gra-
dients of abiotic stress (e.g. Carbonell et al. 2016; Coccia
et al. 2016a) could reveal new insights into how these pro-
cesses interact to shape species coexistence and their
response to biological invasions.
Here, we investigate the impact of the invader
Trichocorixa verticalis verticalis (Fieber, 1851), a water
boatman subspecies originally from North America, on
co-occurrence patterns among three Palaearctic native
boatman species (Sigara, Corixidae) along a salinity gradi-
ent, and determine which mechanisms are driving these
patterns. First, we characterised different aspects of the
ecological niche (habitat specialisation and functional
niche) of each species and their pairwise overlap, using
physiological and biological traits. Second, we compared
the co-occurrence patterns of the boatman species within
non-invaded and invaded areas in southern Iberia and
northern Morocco, using field data. Finally, we used null
models to explore how habitat filtering and niche differen-
tiation assembly rules may explain the co-occurrence pat-
terns of the corixid species.
Materials and methods
STUDY SPECIES
Trichocorixa verticalis verticalis is a small (c. 5 mm) euryhaline
corixid (Hemiptera) native to North America and the Caribbean,
where it mainly lives in coastal habitats such as brackish and sal-
ine lentic waterbodies. This water boatman has been recognised
globally as one of the few aquatic alien insects (Fenoglio et al.
2016) and recorded as an alien species in South Africa, New Cale-
donia, Morocco and the Iberian Peninsula, being the only invasive
aquatic hemipteran in Europe. Its distribution in the introduced
range in south-west Spain and Portugal has been expanding in
recent years (Carbonell et al. 2012a; Guareschi et al. 2013), and it
is predicted to spread widely across Europe and the Mediter-
ranean basin in future years (Guareschi et al. 2013). Thus, T. v.
verticalis is a good candidate for pilot studies of how invasions
influence aquatic communities.
In the invaded Iberian and Moroccan distribution ranges, T. v.
verticalis is found at conductivities from 1 to 120 mS cm
1
.
Although it coexists with native corixid species along a wide salin-
ity gradient, it only breeds at conductivities exceeding 16 mS cm
1
(Rodr
ıguez-P
erez et al. 2009). In order to study co-occurrence
with native corixids, we selected three commonly co-occurring
Sigara species of similar size that can potentially compete with T.
v. verticalis: (i) Sigara lateralis (Leach, 1817), an opportunistic spe-
cies that frequently inhabits temporary freshwater pools (Boda &
Csabai 2009; Carbonell et al. 2011); (ii) Sigara scripta (Rambur,
1840), common in hyposaline waters (Carbonell et al. 2011); and
(iii) Sigara selecta (Fieber, 1848), which inhabits brackish and sal-
ine coastal lentic water bodies (Carbonell et al. 2011).
CHARACTERISATION AND OVERLAPPING OF THE
HABITAT SPECIALISATION AND FUNCTIONAL NICHES
We measured two different facets of the ecological niche: the
habitat specialisation niche and the functional niche (Table S1,
Supporting Information), following two basic niche approaches
(Devictor et al. 2010). The habitat specialisation niche repre-
sents a fundamental Grinnellian niche, related to species perfor-
mance along the salinity gradient. To describe this niche, we
used the physiological tolerance of the different life stages
(adult, nymph and egg) to different conductivities reflecting the
range of conditions where each species can potentially live and
reproduce. The functional niche describes an Eltonian niche
based on specific biological traits, such as fecundity, dispersal
ability, trophic role, life cycle and body size. These traits are
mechanistically related to the ability of species to exploit
resources (Devictor et al. 2010) and cope with stress (Guti
errez-
C
anovas et al. 2015).
We gathered trait data describing physiological tolerance,
fecundity, dispersal ability and trophic features from previous
studies (for methodological details, see Carbonell, Mill
an &
Velasco 2012b; Carbonell et al. 2016; Coccia et al. 2016a). Briefly,
in these studies, individuals were brought from localities within
the study area (i.e. non-invaded and invaded areas) to the labora-
tory for experimental measurements. These analyses included
salinity tolerance for adults, nymphs and eggs (determined as sur-
vival/hatching time in days under different salinity treatments),
fecundity as egg production per day at 10 and 25 g L
1
, dispersal
ability as wing loading (body mass/wing area) and wing aspect
ratio (wing length/wing width), and the isotopic values of carbon
(d13C) and nitrogen (d15N) as trophic traits, which characterise
the basal food source and trophic position.
Remaining traits (life cycle and body size) were obtained from
the available literature (Jackson et al. 1986; Barahona, Millan &
Velasco 2005; Rodr
ıguez-P
erez et al. 2009; Guti
errez-C
anovas
et al. 2012): wintering life cycle stage (adult, nymph, egg), number
of generations per year (≤1or>1), development time (≤1 month
or >1 month), life cycle duration (≤1 year or >1 year) and maxi-
mal body size in mm. Traits were classified into ‘grouping fea-
tures’ following the terminology used by Schmera et al. (2015).
Moreover, each trait was split into categories (e.g. salinity toler-
ances ranged from 0 to 100, and were separated into five cate-
gories of survival percentage ≤16, 17–37, 38–74, 75–100) to
represent the affinity of each species for each trait category. This
data arrangement provides a distribution of the probabilities for
each category for each trait and species. Thus, we characterised
the intraspecific ecological and functional variability displayed in
the study area (Carbonell et al. 2016; Coccia et al. 2016a). This
trait formatting is called a fuzzy coding approach (Chevenet,
Dol
edec & Chessel 1994), and it is widely used in ecological stud-
ies (e.g. Statzner, Bonada & Dol
edec 2008; Guti
errez-C
anovas
et al. 2015). We generated two trait 9species matrices including
species as rows and trait categories as columns. Appendix S1 illus-
trates the procedure followed to build a niche space from
intraspecific traits.
Typically, to model a niche space it is necessary to represent
multiple ecological features via a reduced number of independent
dimensions (e.g. Dol
edec, Chessel & Gimaret-Carpentier 2000;
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
Invasion modifies co-occurrence patterns 3
Carbonell et al. 2011). However, representing intraspecific vari-
ability from heterogeneous traits is more challenging and requires
simulating an appropiate number of ‘pseudo-individuals’ showing
likely trait combinations so as to capture species niche breadth
(Guti
errez-C
anovas et al. 2015). Here, to address this challenge,
we simulated 75 ‘pseudo-individuals’ per species assigning them
traits according to the probabilities that a given trait was assigned
to each category, such that the most likely trait combinations were
best represented (Appendix S1b). We chose 75 ‘pseudo-indivi-
duals’, as this number is adequate to represent the intraspecific
variability of each taxon (C. Guti
errez-C
anovas & S. Vill
eger,
unpubl. data). Take, for instance, a hypothetical trait named c
(Appendix S1). For S. lateralis, this trait has a probability of 0%
for the category c1, 50% for c2 and 50% for c3 (Appendix S1a).
For each simulated pseudo-individual, one category is sampled
based on such probabilities. The chosen category was set to 1,
while the others were set as 0, which finally results in a binary trait
matrix (Appendix S1b). Note that none of the simulations will
render a value of 1 for c1, whose probability was 0%. This process
was repeated 75 times for each trait and species, and separately
for the physiological and biological trait matrices.
Once we had the resulting physiological and biological binary
trait matrices containing the simulated pseudo-individuals
(Appendix S1b), we derived Gower dissimilarity matrices that were
projected into habitat specialisation and functional niche spaces,
respectively, through a Principal coordinates analysis (PCoA) tech-
nique (Appendix S1c) (Vill
eger, Mason & Mouillot 2008; Pavoine
et al. 2009). For this, following the method proposed in Maire
et al. (2015), we retained the first three axes for the habitat speciali-
sation space (mean squared deviation =0006) and the first two
axes for the functional space (mean squared deviation =0009).
The habitat specialisation and functional niche breadths of each
species were estimated separately using the ecological and func-
tional spaces respectively. Niche breadths were estimated as the
sum of the ranges covered by each species for each niche axis
(Appendix S1d). Niche breadths were standardised by maximum
values, and thus ranged from 0 to 1. In addition, we quantified
the pairwise similarity for all species pairs based on a species pair’s
overlap either for habitat specialisation or functional niches
(Appendix S1e). For each niche space, this metric was quantified
as the mean of the percentage of niche overlap between each spe-
cies pair along each axis. These variables also ranged from 0 to 1.
DESCRIPTION OF THE FIELD DATASETS
A total of 338 presence/absence records of the four studied spe-
cies, together with salinity data, were gathered from 179 localities,
including ponds, wetlands and other lentic habitats. These locali-
ties are in the south of the Iberian Peninsula (Spain and Portugal)
and in northern Morocco (see detailed map in Fig. S1). From
those sites, a subset of 106 localities was selected within the area
invaded by T. v. verticalis, comprised of the Iberian Southwest
and North Morocco (hereafter, ‘invaded dataset’), along a gradi-
ent of salinity ranging from 0 to 100 g L
1
. A second subset of 73
sampling points was selected from localities showing comparable
salinity features (salinity range: 01–120 g L
1
), but located out-
side of the invaded area (hereafter, ‘non-invaded dataset’). The
localities of the non-invaded subset were from the Iberian South-
east, which is a well-prospected area for aquatic organisms (Bruno
et al. 2012), but where there is currently no evidence of T. v. verti-
calis occurrence (Guareschi et al. 2013) (Fig. S1). However, this
area has been predicted as highly suitable for the future presence
of T. v. verticalis in terms of climate and habitat suitability (Guar-
eschi et al. 2013). Indeed, T. v. verticalis expansion along the
Mediterranean coast from west to east has recently been observed
in North Morocco (L’Mohdi 2016). Most localities studied (both
invaded and non-invaded by T. v. verticalis) are located in the
Baetic System (Southeast of Spain) and Rif Mountains (North of
Morocco), included in the Betic-Rif Mountain Belt (Lonergan &
White 1997) with similar geological and climatic characteristics,
which results in similar aquatic habitats, biological communities
and salinity range. The presence of at least one of the four studied
species was the criterion to select the localities. Field data were
gathered by sampling in spring and/or summer from the date of
the first observation of the invasive species in the study area
(G€
unther 2004) until 2012, and were stored in the Aquatic Ecol-
ogy Research Group’s Biodiversity database (University of Mur-
cia) and the Wetland Ecology Department’s database (Estaci
on
Biol
ogica de Do~
nana-CSIC, Seville). To complete our dataset,
additional records were extracted from published literature on
corixids (Sala & Boix 2005; Kment 2006; L’Mohdi et al. 2010).
Corixid diversity and abundance were always surveyed using hand
nets (20–30 cm deep and 05–1 mm mesh), sampling all the aqua-
tic microhabitats suitable for corixids (i.e. multihabitats approach,
J
aimez-Cuellar, Vivas & Bonada 2002), and individuals were
immediately preserved in ethanol for posterior identification in
laboratory. However, due to the diverse sources used to gather
field information, we were unable to use abundance data to anal-
yse co-occurrence patterns in both non-invaded and invaded areas.
Finally, we classified the study localities into four salinity classes
following Guti
errez-C
anovas (2014): fresh/subsaline (<3gL
1
),
hyposaline (3–20 g L
1
), mesosaline (20–50 g L
1
) and hyper-
saline (>50 g L
1
).
PATTERNS OF SPECIES CO-OCCURRENCE
The incidence of each species (frequency of occurrence) in the dif-
ferent salinity categories was quantified for both invaded and non-
invaded areas. For each area, we tested if each species pair showed
a significant positive, negative or random co-occurrence pattern
using the COOCCUR R package (Veech 2013, 2014) based on binary
presence/absence data. This analysis uses a probabilistic model to
estimate the likelihood of each species pair occurring less or more
often than expected if each species was distributed independently.
Given that salinity classes were unevenly represented in the
original datasets, we tested if this may influence the observed co-
occurrence patterns. Using simulations where sites were resampled
to achieve a more balanced representation of the salinity classes,
we found no difference in patterns between the results obtained
for simulated and original datasets. Appendix S2 contains the
details and results of these simulations.
NULL MODELS TO TEST THE ASSEMBLY RULES
To explore which assembly rules are explaining the observed co-
occurrence patterns in the non-invaded and invaded areas, we
tested the actual co-occurrence values against the patterns found
in simulated matrices created under null model scenarios of habi-
tat filtering and niche differentiation (e.g. Weiher et al. 2011). For
each simulated matrix, we estimated the number of co-occurrences
for each species pair using the COOCCUR function. These values
were used to create the null distributions to be compared with the
observed co-occurrence values for each species pair for both non-
invaded and invaded areas (for a similar approach, see Sfen-
thourakis, Giokas & Tzanatos 2004 and Gotelli & Ulrich 2010).
Null matrices were created using a filling algorithm which ran-
domly selected one species per site, and then added species until
reaching the original number of species (when richness was >1), so
that simulated communities had the same species richness and
occurrence as the empirical matrices (fixed rowsums and colsums).
This method is a constrained way to randomise presence–absence
community matrices, which resulted in a moderate to low risk of
finding false positives (mean Type I error in simulations was 023
for the non-invaded dataset and 002 for the invaded dataset; for
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
4J. A. Carbonell et al.
more details see Table S4). The order in which species entered the
artificial communities was different depending on the assembly
rule being simulated and species similarity. The habitat filtering
scenario assumes that environmental filtering is the dominant
force, selecting species showing viable traits. As a result, the
assembly rule was that the new species entering the community
should be the one most ecologically similar to the first species,
which itself was selected by chance. On the other hand, the niche
differentiation scenario assumes that negative biotic interactions
prevail in such a way that ecologically similar species will compete
strongly for resources. Thus, the assembly rule was the opposite,
the species entering the community after the first, randomly
selected one being the one with a greater niche dissimilarity
respect to the first species. We used the matrices of pairwise over-
all overlap for each niche space and for individual PCoA axes as
measures of habitat specialisation and functional similarity
between species. For each scenario, species pair and dataset, we
created 999 artificial matrices assembled through either habitat fil-
tering or niche differentiation.
We examined the null model’s statistical significance by deter-
mining the proportion of simulated co-occurrence values that were
below or above the observed co-occurrence value at a=005 [i.e.
two-tailed test: observed values are significantly different from the
null model when they are lower than the 25 lowest values of the
null distribution (left tail), or higher than the 25 greatest null val-
ues (right tail)]. The null hypothesis is that the empirical co-occur-
rence patterns are caused by the assembly rule used in the null
model (i.e. habitat filtering or niche differentiation). A co-occur-
rence value that significantly departs from the null model suggests
that this assembly rule is not driving the co-occurrence patterns
for this species pair.
All statistical analyses were performed with R statistical soft-
ware (libraries: ‘ADE4’, ‘APE’, ‘CLUE’, ‘CLUSTER’, ‘COOCCUR’, ‘FD’, ‘GE-
OMETRY’, ‘GGPLOT2’, ‘GTOOLS’, ‘PLYR’, ‘SPLANCS’ and ‘VEGAN’; R
Development Core Team 2012). The code to run these analyses is
available in Appendix S3.
Results
HABITAT SPECIALISATION, FUNCTIONAL NICHES AND
NICHE OVERLAP
The first three axes of the habitat specialisation niche were
correlated with the salinity tolerance of nymphs, adults
and eggs (axes I, II and III respectively) (see Table 1). The
salinity tolerance of eggs was the trait that best discrimi-
nated the niche differences between T. v. verticalis and the
native species (high correlation with PCoA axis III; Fig. 2a
and Table 1).
The axis I of the functional niche was described by egg
production at 10 g L
1
salinity, maximum body length,
trophic position and wintering life cycle stage (Table 1),
whereas the axis II was described by feeding strategies,
egg production at 25 g L
1
salinity, maximum body
length and trophic position (Table 1). The invasive spe-
cies’ functional niche did not overlap with those of native
species, when both axes were considered (Fig. 2b). The
axis I was best at discriminating between the native spe-
cies and T. v. verticalis, the invader being on the negative
side of this axis (Fig. 2b). Differences in the functional
niche of T. v. verticalis were due principally to its inter-
mediate trophic position, its omnivorous feeding strategy,
higher egg production at 25 g L
1
salinity, its smaller size
Table 1. Principal coordinates analysis (PCoA) results for the habitat specialisation niche (displaying the first three axes) and functional
niche (displaying the first two axes)
PCoA (habitat specialisation niche)
max.pc1 min.pc1 max.pc2 min.pc2 max.pc3 min.pc3
Salinity tolerance of adults 031 026 065 057 050 058
Salinity tolerance of eggs 027 017 059 069 046 063
Salinity tolerance of nymphs 088 081 029 036 052 031
PCoA (functional niche)
max.pc1 min.pc1 max.pc2 min.pc2
d
13
C020 030 060 033
d
15
N032 037 051 043
Development time NA NA NA NA
Number of generations per year NA NA NA NA
Egg production at 10 g L
1
057 046 009 012
Egg production at 25 g L
1
004 038 088 014
Life cycle duration NA NA NA NA
Maximum body length 051 059 051 038
Trophic position 060 050 051 059
Wing aspect ratio 047 030 017 032
Wing loading 031 037 021 018
Wintering stage 069 069 009 009
As each trait has several categories, to identify which traits are explaining each of the niche axes, we show the minimum and maximum
values of the Pearson’s correlation coefficients between each axis and trait categories (see Appendix S1 to see trait categories; min.pc, mini-
mum Pearson’s coefficient for a given trait and axis; max.pc, maximum Pearson’s coefficient for a given trait and axis). NA indicates corre-
lation coefficients were not available for traits with variance equal zero. Values in bold indicate significant associations between trait
categories and PCoA axes. See details in the main text and in Fig. 2.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
Invasion modifies co-occurrence patterns 5
Fig. 2. Principal coordinates analysis (PCoA) results for habitat specialisation (a) and functional (b) niches.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
6J. A. Carbonell et al.
and the presence of nymphs in winter (Fig. 2b and
Table 1).
Trichocorixa verticalis verticalis showed wider habitat
specialisation (i.e. being able to occupy a broader range of
salinity) and functional niche breadth than the native
species, although the greatest differences were observed
among the species’ functional niche breadths (Table 2).
Species showed higher similarity for the habitat specialisa-
tion axes than for axes defining the functional space. In
general terms, the species pairs T. v. verticalis/S. selecta
and S. lateralis/S. scripta showed the greatest similarities
for the habitat specialisation niche (Fig. 2a and Table 3).
The highest similarities in functional niche were found
between S. lateralis and S. scripta (Fig. 2b and Table 3).
PATTERNS OF SPECIES CO-OCCURRENCE
Native species occurrence along the salinity gradient
showed a clear turnover pattern in the non-invaded area.
Sigara lateralis had highest occurrence in fresh/subsaline
waters, while S. scripta occurred with a higher frequency
in fresh/subsaline and hyposaline waters (Fig. 3a). With
increasing salinity, the presence of both S. lateralis and
S. scripta was less likely, whereas S. selecta occurred
with a higher frequency in mesosaline and hypersaline
waters (Fig. 3a). In the invaded area, T. v. verticalis was
present all along the gradient, showing the highest inci-
dence in mesosaline waters. In contrast, S. scripta and
S. selecta showed evidence of reduced occurrence in
fresh/subsaline–hyposaline waters and mesosaline waters
respectively (Fig. 3b).
Analysis of species co-occurrence showed a dominance
of negative co-occurrences between the native species in
the non-invaded area (Fig. 4a and Table S2), whereas both
negative (pairs S. lateralis/S. selecta and S. scripta/S.
selecta) and positive (pair S. lateralis/S. scripta) significant
co-occurrences were found in the invaded area.
Trichocorixa verticalis verticalis showed positive
co-occurrences with S. scripta and S. selecta, but negative
co-occurrence with S. lateralis (Fig. 4b and Table S2).
COMMUNITY ASSEMBLY RULES
The results of the null models for the non-invaded area
showed a relatively balanced contribution of habitat filter-
ing and niche differentiation in explaining the observed co-
occurrences of the species pair S. scripta/S. selecta (Table 4;
Table S3.1 contains the significance and standardised effect
sizes for the null models). Nonetheless, the contribution of
habitat filtering is higher overall, especially for the species
pairs S. lateralis/S. scripta and S. lateralis/S. selecta.
In the invaded area, niche differentiation was more rele-
vant than habitat filtering for the native species pair S.
scripta/S. selecta (Table 5; Table S3.2 contains the signifi-
cance and standardised effect sizes for the null models). On
the other hand, the observed co-occurrence for S. lateralis/
S. selecta and S. lateralis/T. v. verticalis was closer to val-
ues based on both habitat filtering and niche differentia-
tion assumptions. However, the observed co-occurrences
of the invader with the other native species (S. scripta and
S. selecta) were not explained by any model tested. A simi-
lar result was recorded for the native pair S. lateralis/S.
scripta. Although the observed co-occurrence was signifi-
cantly different from simulated values, habitat filtering
models were consistently closer to observed values for the
pair S. lateralis/S. scripta, whereas niche differentiation
models were consistently closer to observed values for the
pair S. scripta/T. v. verticalis (Table 5).
Table 2. Habitat specialisation and functional niche breadths (Tri-
chocorixa verticalis verticalis niche breadth taken as reference =1)
1D estimated niche breadth
Sigara
lateralis
Sigara
scripta
Sigara
selecta T. v. verticalis
Habitat specialisation 085 089 096 100
Functional 060 061 055 100
Table 3. Habitat specialisation and functional niche similarities among species based on significant principal coordinates analysis (PCoA) axes
Habitat specialisation similarity Functional similarity
PCoA axis S. lateralis S. scripta S. selecta PCoA axis S. lateralis S. scripta S. selecta
Overall S. scripta 090 Overall S. scripta 062
Overall S. selecta 075 081 Overall S. selecta 020 014
Overall T. v. verticalis 074 074 090 Overall T. v. verticalis 011 023 015
Axis I S. scripta 084 Axis I S. scripta 049
Axis I S. selecta 087 097 Axis I S. selecta 091 053
Axis I T. v. verticalis 095 084 087 Axis I T. v. verticalis 000 008 000
Axis II S. scripta 099 Axis II S. scripta 081
Axis II S. selecta 098 099 Axis II S. selecta 000 000
Axis II T. v. verticalis 093 092 091 Axis II T. v. verticalis 031 039 042
Axis III S. scripta 084
Axis III S. selecta 041 052
Axis III T. v. verticalis 039 049 091
S. lateralis,Sigara lateralis;S. scripta,Sigara scripta;S. selecta,Sigara selecta;T. v. verticalis,Trichocorixa verticalis verticalis.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
Invasion modifies co-occurrence patterns 7
Discussion
Our results suggest that the presence of T. v. verticalis can
modify co-occurrence patterns and assembly rules of
native Sigara species along the salinity stress gradient. In
the non-invaded area, negative co-occurrences associated
with habitat filtering prevailed among native species. In
contrast, in the invaded area, we found a greater
proportion of positive co-occurrences linked with niche
differentiation. In the case of T. v. verticalis, such positive
co-occurrence indicates that the invasive species can coex-
ist with the more dissimilar species, even at high levels of
abiotic stress.
Habitat filtering seems to explain the negative co-occur-
rences among native species, whereas niche differentiation
has less influence. Previous studies demonstrated that habi-
tat filtering shapes taxonomic and functional patterns
along natural stress gradients (e.g. Weiher et al. 2011;
Guti
errez-C
anovas et al. 2013, 2015). In saline environ-
ments, species generally share a high proportion of traits
(D
ıaz, Alonso & Guti
errez 2008), especially when they are
phylogenetically related as is the case of Sigara species.
However, S. selecta tends to show a segregated distribu-
tion probably due to the differences in osmoregulation
capacity with congeners, but also due to different competi-
tive abilities (Carbonell, Mill
an & Velasco 2012b). Sigara
selecta, the most halo-tolerant species, does not co-occur
with S. lateralis or S. scripta in fresh/subsaline or hypos-
aline waters, despite being able to tolerate hypoosmotic
media. This may be related to the functional similarity of
these species, as suggested by the secondary role of func-
tional niche differentiation in explaining their co-occur-
rence in the non-invaded area. Therefore, biological
interactions such as resource competition, predation or
parasitism might also play a role in explaining the
Fig. 3. Incidence of corixid species along
the salinity gradient in non-invaded (a) and
invaded (b) areas.
Fig. 4. Co-occurrence patterns of corixid species in non-invaded
(a) and invaded (b) areas.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
8J. A. Carbonell et al.
observed exclusion pattern (Reitz & Trumble 2002; Car-
bonell, Mill
an & Velasco 2012b; S
anchez et al. 2015).
The outcome of the invader–stressor interaction depends
on the species tolerance to the stressor, and biotic pressure
at the given level of stress (Gerhardt & Collinge 2007;
Alexander et al. 2011). In our case, salinity did not con-
strain the invasion success of T. v. verticalis, because of its
wide salinity and perturbation tolerance (Van de Meutter
et al. 2010; Carbonell et al. 2016). The invasive species
occurred along the entire salinity gradient studied, peaking
in mesosaline and hypersaline waters, where species rich-
ness tends to be lower (Mill
an et al. 2011) and biotic inter-
actions are potentially weaker (Maestre et al. 2009).
Moreover, higher rates of parasitism by water mites com-
pared to native Sigara may limit the occurrence of T. v.
verticalis in fresh waters (S
anchez et al. 2015).
Niche differentiation has been acknowledged to play an
important role in facilitating the establishment of alien
species (Fargione, Brown & Tilman 2003). Successful plant
invaders are often functionally distinct from species within
the recipient community, reducing competition by filling
an empty niche (Hierro, Maron & Callaway 2005). How-
ever, in other native plant communities, niche differentia-
tion has been found to play a limited role in the
establishment of alien plants (Price & P€
artel 2013). In our
case, the successful establishment and spread of T. v. verti-
calis in the introduced range could have been driven by
functional niche differences, especially its higher fecundity,
ability to breed throughout the year, a different trophic
niche (as shown by stable isotopes, Coccia et al. 2016a)
and its smaller size. The size difference among species has
Table 4. Null model results for the non-invaded area
PCoA axis
Null
model
Non-invaded area
Co-occurrences
(mean values from simulations)
S. lateralis–
S. scripta
S. lateralis–
S. selecta
S. scripta–
S. selecta
Hab overall HF 5200137
Hab1 HF 0030167
Hab2 HF 3100160
Hab3 HF 5100137
Fun overall HF 4824120
Fun1 HF 0052146
Fun2 HF 4823120
Hab overall ND 0051146
Hab1 ND 4824120
Hab2 ND 2547125
Hab3 ND 0051147
Fun overall ND 0030168
Fun1 ND 5200137
Fun2 ND 0051146
Obs. 4116
The mean co-occurrence values from simulations are shown,
together with observed values. Principal coordinates analysis
(PCoA) axis (Hab, habitat specialisation niche; Fun, functional
niche) shows the habitat specialisation or functional axis used to
compute species pairwise dissimilarity. Null model indicates
whether habitat filtering (HF) or niche differentiation scenario
(ND) were used. Values in bold represent co-occurrences falling
within simulated values, which were not significant at P=005.
Values in italic represent actual values observed.
S. lateralis,Sigara lateralis;S. scripta,Sigara scripta;S. selecta,
Sigara selecta.
Table 5. Null model results for the invaded area
PCoA axis
Null
model
Invaded area
Co-occurrences (mean values from simulations)
S. lateralis–S.
scripta
S. lateralis–S.
selecta
S. lateralis–T. v.
verticalis
S. scripta–S.
selecta
S. scripta–T. v.
verticalis
S. selecta–T. v.
verticalis
Hab overall HF 17621697428977
Hab1 HF 13242819786321
Hab2 HF 16737783984400
Hab3 HF 17521694429078
Fun overall HF 159707223310
233
Fun1 HF 146767325111300
Fun2 HF 17414707149493
Hab overall ND 86697553316734
Hab1 ND 104346653315769
Hab2 ND 63256741219197
Hab3 ND 85687563316533
Fun overall ND 62217777813042
Fun1 ND 85187331918576
Fun2 ND 95697734214817
Obs.36 3 79 0 27 14
The mean co-occurrence values from simulations are shown, together with observed values. Principal coordinates analysis (PCoA) axis
(Hab, habitat specialisation niche; Fun, functional niche) shows the habitat specialisation or functional axis used to compute species pairwise
dissimilarity. Null model indicates whether habitat filtering (HF) or niche differentiation scenario (ND) were used. Values in bold represent
co-occurrences falling within simulated values, which were not significant at P=005. Values in italic represent actual values observed.
S. lateralis,Sigara lateralis;S. scripta,Sigara scripta;S. selecta,Sigara selecta;T. v. verticalis,Trichocorixa verticalis verticalis.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
Invasion modifies co-occurrence patterns 9
previously been considered as important in the niche dif-
ferentiation of corixids. Hutchinson (1959) studied a com-
munity of corixids and proposed the ‘limiting similarity’
theory (analogous to ‘niche differentiation’), based on the
idea that size differences among species allow them to
coexist in a shared habitat, by differentially exploiting
resources. In a different community, the coexistence in
rock pools of two corixid species, with similar reproductive
rate and phenology, is mediated by different abilities to
colonise refilled pools, the better disperser having the
advantage of earlier reproduction (Pajunen 1979). Thus,
functional differences detected in T. v. verticalis may be
driving the coexistence with the native species (S. selecta
and S. scripta) in the invaded area via resource partition-
ing and different life-history strategies.
Functionally novel invaders may have dramatic impacts
on ecosystems and communities (Martin et al. 2010). For
example, the crayfish Procambarus clarkii, the zebra mus-
sel Dreissena polymorpha (Pallas 1771) and aquatic plants
such as water hyacinth Eichhornia crassipes (Martius)
Solms-Laubach are strong modifiers of invaded ecosys-
tems, which then become more vulnerable to further alter-
ations (Gherardi & Acquistapace 2007; Villamagna &
Murphy 2010). Given our findings, the omnivorous
regime shown by T. v. verticalis in the invaded area may
have limited impacts on aquatic ecosystems, in concor-
dance with the general trend observed in terrestrial inva-
ders that omnivores have less impact on native
ecosystems than herbivores and predators (Cameron, Vil
a
& Cabeza 2016).
In our study, although no obvious environmental modi-
fications of the habitat have been observed in the invaded
area (Coccia et al. 2016b), the presence of T. v. verticalis
seems to modify both co-occurrence and incidence pat-
terns of native Sigara species. The observed pattern could
be the result of higher competition in fresh/subsaline and
hyposaline waters. Competitive exclusion has been fre-
quently observed in insects and arachnids (Reitz & Trum-
ble 2002). However, the presence of T. v. verticalis did
not seem to produce a strong displacement of the coexist-
ing native Sigara species in mesosaline and hypersaline
waters, although its dominance in permanent, mesosaline
wetlands in Do~
nana where it breeds all year long suggests
that T. v. verticalis can out-compete native corixids in
some habitats (Rodr
ıguez-P
erez et al. 2009). Further stud-
ies that include abundance data on coexisting species are
necessary to clarify the competitive strength of the inva-
der (Mason et al. 2008) and determine the possible dis-
placement of native halotolerant species quantitatively as
well as qualitatively. In particular, a comparative before–
after invasion approach in aquatic ecosystems that have
yet to be invaded but which have a high probability of
invasion in the near future (e.g. along the Atlantic and
Mediterranean coasts of the Iberian Peninsula and Mor-
occo), and along natural and anthropogenic perturbation
gradients, would clarify the impacts of the invasion at the
species, community and ecosystem levels.
The present work makes a novel contribution to the
study of the impacts of invasive species at the community
level through the integration of habitat specialisation and
functional niche approaches with field occurrence data.
We showed how the presence of the invasive species T. v.
verticalis can modify the distribution and co-occurrence
patterns of native Sigara species along the salinity gradi-
ent, as well as the main assembly rules that shape the
assemblages in non-invaded and invaded areas. We found
that in non-invaded habitats, environmental filtering drives
habitat segregation of species along the salinity gradient,
with a lower contribution of niche differentiation. On the
other hand, in the invaded area niche differentiation seems
to be the primary mechanism reducing competition
between the invasive and native species, favouring coexis-
tence and resource partitioning. A similar pattern could be
predicted for future invasions in other taxa, in cases where
the invasive species shows a functional niche different
enough to that of native species to enable resource parti-
tioning. The approach employed here can also be useful to
anticipate the consequences of ecologically novel invaders
for native communities at structural and functional levels
in a global change context.
Authors’contributions
J.A.C., A.M., J.V. and C.G.-C. conceived and designed the work. J.A.C.,
A.M., A.J.G., C.C. and S.G. gathered data. J.A.C., J.V., A.M. and C.G.-C.
analysed the data. J.A.C., J.V., A.M., A.J.G., C.C., S.G. and C.G.-C. wrote
and discussed the paper.
Acknowledgements
We thank Laura Monteagudo and the members of the Aquatic Ecology
Research Group (Universidad de Murcia, Spain) for their help and sug-
gestions at various stages of this study, as well as the group of Ouas-
sima L0Mohdi and Nard Bennas (Abdelmalek Essa^
adi University,
T
etouan, Morocco) for providing field data from Morocco. This work
was partially supported by funding from a predoctoral FPU grant to
J.A.C. C.C. was supported in part by a JAE predoctoral grant from
CSIC and by the postdoctoral grant 3160330 from FONDECYT. was
supported by the MARS project (Managing Aquatic ecosystems and
water Resources under multiple Stress), funded by the European Union
under the 7th Framework Programme (contract no. 603378). This work
was also supported by the projects P10-RNM-6262 (A.J.G.) (Consejer
ıa
de Innovaci
on, Ciencia y Empresa, Junta de Andaluc
ıa), ‘Atlas de los
cole
opteros acu
aticos de Espa~
na peninsular’ (A.M.) (Ministerio de Agri-
cultura, Alimentaci
on y Medio Ambiente) and CGL2013-48950-C2-2-P
(J.V.) (Ministerio de Econom
ıa y Competitividad).
Data accesibility
Data available from the Dryad Digital Repository https://doi.org/10.5061/
dryad.1v035 (Carbonell et al. 2017).
References
Alexander, J.M., Kueffer, C., Daehler, C.C., Edwards, P.J., Pauchard, A.
& Seipel, T. & MIREN Consortium (2011) Assembly of nonnative floras
along elevational gradients explained by directional ecological filtering.
Proceedings of the National Academy of Sciences,108, 656–661.
Barahona, J., Millan, A. & Velasco, J. (2005) Population dynamics, growth
and production of Sigara selecta (Fieber, 1848) (Hemiptera, Corixidae) in
a Mediterranean hypersaline stream. Frehswater Biology,50, 2101–2113.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
10 J. A. Carbonell et al.
Boda, P. & Csabai, Z. (2009) Seasonal and diel dispersal activity character-
istics of Sigara lateralis (Leach, 1817) (Heteroptera: Corixidae) with spe-
cial emphasis on possible environmental factors and breeding state.
Aquatic Insects,31, 301–314.
Bruno, D., S
anchez-Fern
andez, D., Mill
an, A., Picazo, F., Carbonell, J.A.
& Velasco, J. (2012) Predicting the richness of aquatic beetles and bugs
in a semi-arid mediterranean region. Limnetica,31,23–36.
Cameron, E.K., Vil
a, M. & Cabeza, M. (2016) Global meta-analysis of the
impacts of terrestrial invertebrate invaders on species, communities and
ecosystems. Global Ecology and Biogeography,25, 596–606.
Carbonell, J.A., Guareschi, S., Coccia, C., S
anchez-Fern
andez, D., Velasco,
J., Boyero, L., Green, A.J. & Mill
an, A. (2012a) Distribuci
on de Tri-
chocorixa verticalis verticalis (Fieber, 1851) (Heteroptera: Corixidae) a
nivel mundial y su expansi
on en la Pen
ınsula Ib
erica. (ed. EEI 2012
Notas Cient
ıficas). GEIB Serie T
ecnica N°,5, 148–152.
Carbonell, J.A., Guti
errez-C
anovas, C., Bruno, D., Abell
an, P., Velasco, J.
& Mill
an, A. (2011) Ecological factors determining the distribution and
assemblages of the aquatic hemiptera (Gerromorpha & Nepomorpha) in
the Segura river basin (Spain). Limnetica,30,59–70.
Carbonell, J.A., Mill
an, A., Green, A.J., C
espedes, V., Coccia, C. &
Velasco, J. (2016) What traits underpin the successful establishment and
spread of the invasive water bug Trichocorixa verticalis verticalis (Fieber,
1851)? Hydrobiologia,768, 273–286.
Carbonell, J.A., Mill
an, A. & Velasco, J. (2012b) Concordance between
realised and fundamental niches in three Iberian Sigara species (Hemi-
ptera: Corixidae) along a gradient of salinity and anionic composition.
Freshwater Biology,57, 2580–2590.
Carbonell, J.A., Velasco, J., Mill
an, A., Green, A.J., Coccia, C., Guareschi,
S. & Guti
errez-C
anovas, C. (2017) Data from: Biological invasion modi-
fies the co-occurrence patterns of insects 1 along a stress gradient. Dryad
Digital Repository, https://doi.org/10.5061/dryad.1v035
Chase, J.M. & Leibold, M.A. (2003) Ecological Niches: Linking Classical
and Contemporary Approaches. University of Chicago Press, Chicago,
IL, USA and London, UK.
Chesson, P. (2000) Mechanisms of maintenance of species diversity. Annual
Review of Ecology and Systematics,31, 343–366.
Chevenet, F., Dol
edec, S. & Chessel, D. (1994) A fuzzy coding approach
for the analysis of long-term ecological data. Freshwater Biology,31,
295–309.
Coccia, C., Fry, B., Ram
ırez, F., Boyero, L., Bunn, S.E., Diz-Salgado, C.,
Walton, M., Le Vay, L. & Green, A.J. (2016a) Niche partitioning
between invasive and native corixids (Hemiptera, Corixidae) in south-
west Spain. Aquatic Sciences,78, 779–791.
Coccia, C., Vanschoenwinkel, B., Brendonck, L., Boyero, L. & Green, A.J.
(2016b) Newly created ponds complement natural waterbodies for
restoration of macroinvertebrate assemblages. Freshwater Biology,61,
1640–1654.
Davidson, A.M., Jennions, M. & Nicotra, A.B. (2011) Do invasive species
show higher phenotypic plasticity than native species and if so, is it
adaptive? A meta-analysis. Ecology Letters,14, 419–431.
De Roos, A.M., Schellekens, T., Van Kooten, T. & Persson, L. (2008)
Stage-specific predator species help each other to persist while competing
for a single prey. Proceedings of the National Academy of Sciences,105,
13930–13935.
Devictor, V., Clavel, J., Julliard, R., Lavergne, S., Mouillot, D., Thuiller,
W., Venail, P., Vill
eger, S. & Mouquet, N. (2010) Defining and measur-
ing ecological specialisation. Journal of Applied Ecology,47,15–25.
D
ıaz, A.M., Alonso, M.L.S. & Guti
errez, M.R.V.A. (2008) Biological traits
of stream macroinvertebrates from a semi-arid catchment: patterns along
complex environmental gradients. Freshwater Biology,53,1–21.
Dol
edec, S., Chessel, D. & Gimaret-Carpentier, C. (2000) Niche separation
in community analysis: a new method. Ecology,81, 2914–2927.
Fargione, J., Brown, C.S. & Tilman, D. (2003) Community assembly and
invasion: an experimental test of neutral versus niche processes. Proceed-
ings of the National Academy of Sciences,100, 8916–8920.
Fenoglio, S., Bonada, N., Guareschi, S., L
opez-Rodr
ıguez, M.J., Mill
an,
A. & Tierno de Figueroa, J.M. (2016) Freshwater ecosystems and aqua-
tic insects: a paradox in biological invasions. Biology Letters,12,
20151075.
Gallardo, B., Clavero, M., S
anchez, M.I. & Vil
a, M. (2015) Global ecologi-
cal impacts of invasive species in aquatic ecosystems. Global Change
Biology,22, 151–163.
Gerhardt, F. & Collinge, S.K. (2007) Abiotic constraints eclipse biotic resis-
tance in determining invasibility along experimental vernal pool gradi-
ents. Ecological Applications,17, 922–933.
Gherardi, F. & Acquistapace, P. (2007) Invasive crayfish in Europe: the
impact of Procambarus clarkii on the littoral community of a Mediter-
ranean lake. Freshwater Biology,52, 1249–1259.
Gotelli, N.J. & Ulrich, W. (2010) The empirical Bayes approach as a tool
to identify non-random species associations. Oecologia,162, 463–477.
Guareschi, S., Coccia, C., S
anchez-Fern
andez, D., Carbonell, J.A., Velasco,
J., Boyero, L., Green, A.J. & Mill
an, A. (2013) How far could the alien
boatman Trichocorixa verticalis verticalis Spread? Worldwide estimation
of its current and future potential distribution. PLoS ONE,8, e59757.
G€
unther, H. (2004) Trichocorixa verticalis verticalis (Fieber), eine nearktis-
che Ruderwanze in Europa (Heteroptera: Corixidae). Mitteilungen des
Internationalen Entomologischen Verereins,29,45–49.
Guti
errez-C
anovas, C. (2014) Ecosystem responses to natural and anthro-
pogenic stress. From biomonitoring tools to predictive ecology. PhD thesis.
University of Murcia, Spain.
Guti
errez-C
anovas, C., Hern
andez, J., Mill
an, A. & Velasco, J. (2012)
Impact of chronic and pulse dilution disturbances on metabolism and
trophic structure in a saline Mediterranean stream. Hydrobiologia,686,
225–239.
Guti
errez-C
anovas, C., Mill
an, A., Velasco, J., Vaughan, I.P. & Ormerod,
S.J. (2013) Contrasting effects of natural and anthropogenic stressors on
beta diversity in river organisms. Global Ecology and Biogeography,22,
796–805.
Guti
errez-C
anovas, C., S
anchez-Fern
andez, D., Velasco, J., Mill
an, A. &
Bonada, N. (2015) Similarity in the difference: changes in community
functional features along natural and anthropogenic stress gradients.
Ecology,96, 2458–2466.
Hierro, J.L., Maron, J.L. & Callaway, R.M. (2005) A biogeographical
approach to plant invasions: the importance of studying exotics in their
introduced and native range. Journal of Ecology,93,5–15.
Hutchinson, G.E. (1959) Homage to Santa Rosalia or why are there so
many kinds of animals? The American Naturalist,93, 145–159.
Iacarella, J.C., Dick, J.T., Alexander, M.E. & Ricciardi, A. (2015) Ecological
impacts of invasive alien species along temperature gradients: testing the
role of environmental matching. Ecological Applications,25, 706–716.
Jackson, M.C., Donohue, I., Jackson, A.L., Britton, J.R., Harper, D.M. &
Jansson, A. (1986) The Corixidae (Heteroptera) of Europe and some
adjacent regions. Acta Entolologica Fennica,47,1–94.
J
aimez-Cuellar, P., Vivas, S., Bonada, N. et al. (2002) Protocolo GUA-
DALMED (PRECE). Limnetica,21, 187–204.
Kment, P. (2006) A contribution to the faunistic of aquatic and semiaquatic
bugs (Heteroptera: Nepomorpha, Gerromorpha) in Portugal, with the
review of biology of the Nearctic corixid Trichocorixa verticalis (Fieber,
1851). Boletin Sociedad Entomol
ogica Aragonesa,38, 359–361.
L’Mohdi, O. (2016). Les H
emipt
eres aquatiques du Maroc: Altas, Biogeogra-
phie et degree de vuln
erabilit
e. PhD thesis. Universit
e Abdelmalec
Essaadi, T
etouan, Morocco.
L’Mohdi, O., Bennas, N., Himmi, O., Hajji, K., El Haissoufi, M., Her-
nando, C., Carbonell, J.A. & Mill
an, A. (2010) Trichocorixa verticalis
verticalis (Fieber 1851) (Hemiptera, Corixidae): une nouvelle especie exo-
tique au Maroc. Bolet
ın de la Sociedad Entomol
ogica Aragonesa,46,
395–400.
Lonergan, L. & White, N. (1997) Origin of the Betic-Rif mountain belt.
Tectonics,16, 504–522.
MacDougall, A.S. & Turkington, R. (2005) Are invasive species the drivers
or passengers of change in degraded ecosystems? Ecology,86,42–55.
Maestre, F.T., Callaway, R.M., Valladares, F. & Lortie, C.J. (2009) Refin-
ing the stress-gradient hypothesis for competition and facilitation in
plant communities. Journal of Ecology,97, 199–205.
Maire, E., Grenouillet, G., Brosse, S. & Vill
eger, S. (2015) How many
dimensions are needed to accurately assess functional diversity? A prag-
matic approach for assessing the quality of functional spaces. Global
Ecology and Biogeography,24, 728–740.
Martin, J.L., Stockton, S., Allombert, S. & Gaston, A. (2010) Top-down
and bottom-up consequences of unchecked ungulate browsing on plant
and animal diversity in temperate forests: lessons from a deer introduc-
tion. Biological Invasions,12, 353–371.
Mason, N.W., Lanoisel
ee, C., Mouillot, D., Wilson, J.B. & Argillier, C.
(2008) Does niche overlap control relative abundance in French lacus-
trine fish communities? A new method incorporating functional traits.
Journal of Animal Ecology,77, 661–669.
Mill
an, A., Velasco, J., Guti
errez-C
anovas, C., Arribas, P., Picazo, F.,
S
anchez-Fern
andez, D. & Abell
an, P. (2011) Mediterranean saline
streams in southeast Spain: what do we know? Journal of Arid Environ-
ments,75, 1352–1359.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
Invasion modifies co-occurrence patterns 11
Pajunen, V.I. (1979) Quantitative analysis of competition between Arcto-
corisa carinata (Sahlb.) and Callicorixa producta (Reut.) (Hemiptera,
Corixidae). Annales Zoologici Fennici,16, 195–200.
Pavoine, S., Vallet, J., Dufour, A.B., Gachet, S. & Daniel, H. (2009) On
the challenge of treating various types of variables: application for
improving the measurement of functional diversity. Oikos,118, 391–402.
Price, J.N. & P€
artel, M. (2013) Can limiting similarity increase invasion
resistance? A meta-analysis of experimental studies. Oikos,122, 649–656.
R Development Core Team (2012) R: A Language and Environment for Sta-
tistical Computing. R Foundation for Statistical Computing, Vienna,
Austria.
Reitz, S.R. & Trumble, J.T. (2002) Competitive displacement among insects
and arachnids. Annual Review of Entomology,47, 435–465.
Rodr
ıguez-P
erez, H., Florencio, M., G
omez-Rodr
ıguez, C., Green, A.J.,
D
ıa-Paniagua, C. & Serrano, L. (2009) Monitoring the invasion of the
aquatic bug Trichocorixa verticalis verticalis (Hemiptera: Corixidae) in
the wetlands of Do~
nana National Park (SW Spain). Hydrobiologia,634,
209–217.
Sala, J. & Boix, D. (2005) Presence of the nearctic water boatman Tri-
chocorixa verticalis verticalis (Fieber, 1951) (Heteroptera, Corixidae) in
the Algarve region (S Portugal). Graellsia,61,31–36.
S
anchez, M.I., Coccia, C., Valdecasas, A.G., Boyero, L. & Green, A.J.
(2015) Parasitism by water mites in native and exotic Corixidae: are
mites limiting the invasion of the water boatman Trichocorixa verticalis
(Fieber, 1851)? Journal of Insect Conservation,19, 433–447.
Schmera, D., Podani, J., Heino, J., Er€
os, T. & Poff, N.L. (2015) A pro-
posed unified terminology of species traits in stream ecology. Freshwater
Science,34, 823–830.
Sfenthourakis, S., Giokas, S. & Tzanatos, E. (2004) From sampling stations
to archipelagos: investigating aspects of the assemblage of insular biota.
Global Ecology and Biogeography,13,23–35.
Shipley, B., Vile, D. & Garnier,
E. (2006) From plant traits to plant com-
munities: a statistical mechanistic approach to biodiversity. Science,314,
812–814.
Simberloff, D., Martin, J.L., Genovesi, P. et al. (2013) Impacts of biological
invasions: what’s what and the way forward. Trends in Ecology and Evo-
lution,28,58–66.
Statzner, B., Bonada, N. & Dol
edec, S. (2008) Biological attributes discrim-
inating invasive from native European stream macroinvertebrates. Bio-
logical Invasions,10, 517–530.
Van de Meutter, F., Trekels, H., Green, A. & Stoks, R. (2010) Is salinity
tolerance the key to success for the invasive water bug Trichocorixa verti-
calis?Hydrobiologia,649, 231–238.
Veech, J.A. (2013) A probabilistic model for analysing species co-occur-
rence. Global Ecology and Biogeography,22, 252–260.
Veech, J.A. (2014) The pairwise approach to analyzing species co-occur-
rence. Journal of Biogeography,41, 1029–1035.
Villamagna, A.M. & Murphy, B.R. (2010) Ecological and socio-economic
impacts of invasive water hyacinth (Eichhornia crassipes): a review.
Freshwater Biology,55, 282–298.
Vill
eger, S., Mason, N.W. & Mouillot, D. (2008) New multidimensional
functional diversity indices for a multifaceted framework in functional
ecology. Ecology,89, 2290–2301.
Violle, C., Nemergut, D.R., Pu, Z. & Jiang, L. (2011) Phylogenetic limiting
similarity and competitive exclusion. Ecology Letters,14, 782–787.
Weiher, E., Freund, D., Bunton, T., Stefanski, A., Lee, T. & Bentivenga, S.
(2011) Advances, challenges and a developing synthesis of ecological
community assembly theory. Philosophical Transactions of the Royal
Society B-Biological Sciences,366, 2403–2413.
Received 23 June 2016; accepted 28 March 2017
Handling Editor: Kailen Mooney
Supporting Information
Details of electronic Supporting Information are provided below.
Fig. S1. Study area showing the selected localities in areas that
have been invaded by Trichocorixa verticalis verticalis and areas
that have not been invaded.
Table S1. Physiological and biological traits of corixids and their
categories.
Table S2. Co-occurrence results.
Table S3. Null model standardised effects for the invaded and
non-invaded areas.
Table S4. Null model Type I error and P-values for the invaded
and non-invaded areas.
Appendix S1. Procedure used to calculate habitat specialisation
and functional niche features.
Appendix S2. Simulations to assess the effect of unbalanced repre-
sentation of the salinity classes in the study of co-occurrence pat-
terns.
Appendix S3. R code.
©2017 The Authors. Functional Ecology ©2017 British Ecological Society, Functional Ecology
12 J. A. Carbonell et al.
