PREDICTING FUTURE SPECIES DISTRIBUTION OF ODONATA … · PREDICTING FUTURE SPECIES DISTRIBUTION OF ODONATA IN WESTERNMOST MEDITERRANEAN REGION UNDER CLIMATE CHANGE Master in Ecology,
Post on 03-Apr-2018
216 Views
Preview:
Transcript
PREDICTING FUTURE SPECIES
DISTRIBUTION OF ODONATA IN
WESTERNMOST MEDITERRANEAN
REGION UNDER CLIMATE CHANGE
Master in Ecology, Environmental Management and Restoration
Author: Aida Viza Sánchez Tutor: Dr. Cesc Múrria i Farnós
Department of Evolutionary Biology, Ecology and Environmental Sciences
University of Barcelona 28th of September of 2016
PREDICTING FUTURE SPECIES
DISTRIBUTION OF ODONATA IN
WESTERNMOST MEDITERRANEAN
REGION UNDER CLIMATE CHANGE
Master in Ecology, Environmental Management and Restoration
Author: Aida Viza Sánchez
Tutor: Dr. Cesc Múrria i Farnós Department of Evolutionary
Biology, Ecology and Environmental
Sciences University of Barcelona
28th of September of 2016
Internal Advisor: Dr. Núria Bonada, FEM Research Group, University of Barcelona
Main Advisor: Dr. Cesc Múrria, FEM Research Group, University of Barcelona
Author: Aida Viza, MSc student, FEM Research Group, University of Barcelona
ABSTRACT
A critic question in biodiversity conservation is how species will response in front of current rates
of Climate Change. Such environmental alterations have the potential to modify habitat
characteristics and, consequently, it is predicted that many species may shift their ranges to
higher latitudes or altitudes to remain in a constant environmental niche. On another hand,
those species with high evolutionary adaptation, phenotypic acclimation or plasticity are
expected to have the ability to face new conditions. Finally, species with poor strategies are
vulnerable and can become extinct. In this project, I focus on the evolutionary history, functional
traits characteristics and Species Distribution Models (SDM) of Odonata to elucidate how
species distribution of odonates in Iberian Peninsula and Morocco will be affected by Climate
Change and the role that traits would play in future species responses. In general, I found that
odonates potential distribution will be altered by an increase of temperature seasonality and
drought events, as a result of anthropogenic impact. High emissions scenarios were dominated
by a reduction of species potential distribution, while low emissions scenarios showed a trend to
subtile displacement from current species distribution. The ecological distance between species
including also closely related species was decoupled to their phylogenetic divergence.
Therefore, phylogeny cannot predict the ecological requirements of species. Moreover, none
clear pattern was found between traits (ecological and life-history), current habitat occupancy
and future potential distribution under several models of climate change. Hence, I cannot
elucidate species response based on the probability of their lineage to neither extinction,
northward range expansion nor shift in its distribution range. Further studies modelling multi-
species distribution considering intraspecific traits and genetic variability will be needed to infer
future species-specific distribution and extinction risk in order to do a correct management of
freshwater biodiversity under climate change.
RESUM
Una qüestió crítica en la conservació de la biodiversitat és com les espècies respondran davant
del Canvi Climàtic. Les alteracions ambientals poden modificar les característiques de l'hàbitat
i, en conseqüència, s’espera que moltes espècies canviïn la seva distribució a latituds o altituds
més elevades, per tal de romandre en un nínxol ambiental constant. S'espera que les espècies
amb una alta capacitat d’adaptació evolutiva, d’aclimatació o de plasticitat fenotípica puguin fer
front a les noves condicions. En canvi, les espècies amb estratègies limitades són vulnerables i
poden arribar a extingir-se. Aquest projecte pretén entendre com afectarà el Canvi Climàtic a la
distribució dels odonats de la Península Ibèrica i el Marroc, i quin paper juguen els trets
biològics en la resposta futura de les espècies. En general, la distribució potencial dels odonats
serà alterada com a resultat del canvi climàtic antropogènic. Els escenaris futurs amb majors
emissions estan dominats per la reducció de la distribució potencial de les espècies, mentre
que en els escenaris de baixes emissions aquest tendeix a desplaçar-se. No obstant això, la
distància ecològica entre espècies no està acoplada a la seva divergència filogenètica, per tant
la filogènia no pot predir els requeriments ecològics de les espècies. D'altra banda, no s'ha
trobat cap pauta clara entre les característiques funcionals, l’ocupació actual i la predicció de la
distribució potencial futura sota diversos models de canvi climàtic. Per tant, no puc aclarir com
respondrà cada espècie ni atribuir a un llinatge la probabilitat de canvi en la seva àrea de
distribució. Per tant, calen més estudis de modelització de distribució de múltiples espècies
tenint en compte les característiques i la variabilitat genètica intraespecífica, ja que són
necessaris per a inferir la futura distribució de les espècies i el grau d'amenaça, per tal de
realitzar una correcta gestió de la biodiversitat d’ecosistemes fluvials sota el canvi climàtic.
INDEX
Introduction ......................................................................................................... 2
Methods .............................................................................................................. 6
Study area, species occurrences and data specifications ......................... 6
Climatic models, bioclimatic variables and future species distribution ...... 6
Compilation of DNA sequences and Phylogenetic analyses ..................... 9
Species-specific habitat preferences and trait conservatism ................... 10
Results .............................................................................................................. 11
Discussion ......................................................................................................... 22
Conclusions ....................................................................................................... 25
Acknowledgements ............................................................................................ 25
References ........................................................................................................ 26
2
INTRODUCTION
Current predicted rates of climate warming will likely modify current habitat
characteristics (Walther et al., 2002; Travis, 2003). As consequence, it is expected that
many species may shift their ranges to higher latitudes and/or altitudes, where the
temperature conditions will be more suitable to remain in a constant environmental niche;
may locally adapt or phenotypically acclimatise to the new ecological conditions; or will
go to extinct (Parmesan, 2006; Markovic et al., 2014; Stoks et al., 2014; Buckley &
Kingsolver, 2016).
Aquatic ecosystems have showed high vulnerability to global change due to losses of
habitat heterogeneity, reduction of connectivity and additional stressors such as
pollution, river regulation, over-abstraction of water, and unpredictable consequences of
alien species introduction (Sala et al., 2000; Woodward et al., 2010). Moreover, the
expected increase of temperature will influence physiological processes of freshwater
macroinvertebrates species such as increases in body size, development rate and
growth rate (Burgmer et al., 2007; Markovic et al., 2014; Stoks et al., 2014). As a result,
habitat suitability will decrease or shift for many species (Markovic et al., 2014) and
freshwater macroinvertebrates communities’ composition will change (Daufresne et al.,
2007). For this reason, to predict how habitats will shift in the future, dispersive abilities
of species and their evolutionary potentials for adapting in new conditions is critical for
the conservation and management of freshwater ecosystems and their associated
freshwater biodiversity.
Functional traits and habitat preferences play an important role in species survival facing
Climate Change because determinate if species is able to shift toward future suitable
habitats or, on the contrary, can locally adapt to the new environmental conditions. Only
those species with a high phenotypic plasticity are expected to have the ability to change
its ecological preferences in response to environmental fluctuations (Parmesan, 2006;
Wellenreuther et al., 2012). Then, determining what biological traits are favourable to
locally face the global warming and whose that may promote a possible expansion to
new habitats (Schloss et al., 2012), can provide critical information for elucidate the
future of species. Once within the new habitat, population persistence is likely to be
driven by traits that determines the species strategy, such that generalists might be more
successful in meeting their needs for food and shelter than specialists (Jeschke &
Strayer, 2006). Estimates of an organism’s fundamental niche following Climate Change
3
can be compared to its current realized niche to predict whether a species will need to
move or adjust its phenotype to avoid extinction (Buckley & Kingsolver, 2016).
The ecological niche of a species is the set of biotic and abiotic conditions in which a
given species is able to persist and maintain stable population sizes (Hutchinson, 1957).
What determine which niche is likely to be occupied by a certain species are the
biological and ecological traits that characterize a species. Species functional traits
evolve over time as a response to species interaction and ecological conditions, but there
is the tendency of species to fix their ancestral ecological characteristics and to retain
aspects of their fundamental niche through time, which is known as niche conservatism
(Webb et al., 2002; Wiens & Graham, 2005). Studying simultaneously niche
conservatism and future potential distribution of species is possible to estimate extinction
risk, northward range expansions or shifts in range distribution, and local adaptation
(plasticity). In this project, I focus on the evolutionary history, functional traits
characteristics and Species Distribution Models (SDM) of Odonata to elucidate how
species distribution in the westernmost Mediterranean region will be affected by Climate
Change and to determine the role that traits would play in future species responses of
odonates.
Odonata is a good model taxon to predict effects of climate warning on freshwater
biodiversity because of (1) their tropical evolutionary origins may limit their distribution
by temperature, (2) medium-high local abundances that facilitates sampling, captures
and experiments, (3) high specialization of larvae and adult for habitat usages and
evident niche partition among species, (4) a long history of scientific research in ecology,
behaviour and evolution because many species can be reared and crossed successfully
in captivity, and (5) the extensive recording and abundant historical datasets, mostly by
volunteer work (Hassall & Thompson, 2008; Ott, 2010).
Given their high mobility, Odonata are currently experiencing a clear trend of northward
range expansion from Morocco to the Iberian Peninsula favoured by climate warming
and facilitated by the increasing frequency of the Saharan southern winds (Herrera-Grao
et al., 2012). The first North-African species arrival registered was Orthetrum nitidinerve
in 1842 (Jaquemin & Boudout, 1999). In the last 60 years, other species that arrived to
Iberian Peninsula from North-Africa were Brachythemis impartita (Compte Sart, 1962),
Paragomphus genei (Testard, 1975), Trithemis annulata (López, 1983), Diplacodes
lefebvrei (Conesa García, 1985), Orthetrum trinacria (Hartung, 1985), and the most
recently registered Trithemis kirbyi (Chelmick & Pickess, 2008), which was recorded in
South-Catalonia in 2012 (Herrera-Grao et al., 2012).
4
Since climate warming is expected to promote changes in the geographical distribution
of odonates, I predicted 5 different categories of changes in geographical species range
that differ in whether the potential future distribution area will increase, decrease or
remain insignificantly alterable (fig.1). (1)”Displaced” potential distribution. Under this
model, the mean of distribution range should displace to northern latitudes. (2)
“Expansive” potential distribution that implies an overall increase of appropriate
environmental areas keeping current distribution. (3) “Non-change” potential distribution
is predicted when the future bioclimatic niche will remain subtlety altered, thus, the
current and future potential range will overlap. (4) “Reduced” potential distribution when
the current distribution will decline due to a reduction of suitable habitat. (5) “Extinct”
potential distribution is expected when the potential area and suitable habitats could
disappear in the future and, therefore, those species will be especially vulnerable to
extinction.
This study focuses in understanding how the distribution of Odonata species currently
located at the westernmost Mediterranean region will change in 2050 and 2080 using
environmental niche models approach and considering the most pessimistic (high
emissions) and optimistic (low emissions) predictions of future climate. Since species
responses facing climate warming are driven by ecological traits and evolutionary
potential, this project also determines the species ecological niches and evolutionary
Fig. 1. Scheme of the main hypothesis of the study. Species can be classified in 5 different
categories depending on how their potential distribution will change in the future. (a)
“Displaced”: potential area shifts northwards, (b) “Expansive”: expansion of the potential
area, (c) “Non-change”: few changes in the location of potential area, (d) “Reduced”: potential
area regression, (e) “Extinct”: potential area loss.
5
history (phylogeny and niche conservatism) of the odonates. I note three points that sum
up the aims of the study:
(1) To predict future environmental conditions in the Iberian Peninsula and Morocco and
to assess how Odonata species distribution will be altered when fitting predicted future
conditions. In order to model current potential species distribution and after perform
future predictions of species potential distribution on future climatic conditions, I used a
compilation of occurrence data from most of the Iberian Peninsula and Morocco species
in which this study focuses.
(2) To determine the evolutionary history of Odonata and the composition of the
functional traits of each species for assessing niche conservatism. Odonata species
appear in almost all type of freshwater habitats, but species showed high habitat-
specificity (Suhling et al., 2015). If traits are phylogenetically conserved, i.e., the niche
conservatism is supported, the relatedness of species should preserve the signature of
habitat preference. As a consequence, I can assess which traits and lineages will be
favoured by Climate Change, and thus, the predictions of extinction risk will be
straightforward. Also, trait conservatism will allow me distinguish if the 5 different
categories of changes in geographical species range are based on phylogenetic
relations.
(3) To associate habitat-specificity with species traits for determining species
vulnerability. Since the unanimity in predictions indicate an increase of temperature and
seasonality of precipitations, I expect species that preferred temporary habitats such as
ponds or streams, which should adapt to these habitat by fast development and growth
rates (Perry et al., 2005), will be favoured by Climate Change. Greater ecological
generalization may release species from being constrained by the distribution or
phenology of species they associate with. Moreover, large body species, which
commonly showed also large geographical range, may correlate directly with dispersal
ability and long life-cycle duration, in some cases also with environmental tolerance and
ecological generalization, and, inversely, with reproductive rates (Davies et al., 2009). If
all these considerations are supported in odonates, large body species and species that
preferred temporary habitats will expand their distribution and they should belong to
categories 1 or 2 (future potential area shift or expansion, respectively). On the other
hand, species with poor dispersal abilities or species located in higher elevations should
be more sensible to Climate Change and they must adapt to great magnitude of warming.
Moreover, their habitat will be more prone to disappear or be fragmented, and therefore
these species will have major constraints for arriving to their suitable thermal conditions
6
(Wrona et al., 2006). These species should belong to categories 4 or 5 (potential area
reduction) and are probably threatened and very likely to disappear.
METHODS
Study area, species occurrences and data specifications
Total species richness in the studied area was 91 (Torralba-Burrial & Ocharan, 2007;
Boudot et al., 2009). Species occurrence was compiled using four datasets: (1)
Oxygastra, The Catalan odonatologist group; (2) ROLA’s project of AEA “El Bosque
Animado”, an environmental education association from Andalucia; (3) published
database from Aragon and Cantabria (BOS Arthropod Collection of University of Oviedo,
Spain; Torralba-Burrial & Ocharan, 2013), and (4) African data gently offered by
researcher Mohamed El Haissoufi from Université Abdelmalek Essaadi.
Before running the models, the occurrences list (species-by-site) was filtered following a
series of rules: only taxonomical identification at the species level were considered,
which implied the removal of larvae data; all occurrences were considered since 2003
when Oxygastra started a regular and systematic data collection; occurrences non-
georeferenced were discarded; and localities where site description was available but
lacked of precise location were georeferenced using the official website of “Institut
Cartogràfic de Catalunya” (ICC, www.icc.cat/vissir3/). Since the main fuse zone in the
Iberian Peninsula is 30N UTM, all geographic data were transformed and standardized.
Climatic models, bioclimatic variables and future species distribution
To assess how the potential distribution areas would change in the future, Species
Distribution Models (SDM) based on occurrences and current environmental variables
as predictors were performed. Climatic conditions were described by 19 bioclimatic
variables based on temperature and rainfall values. These bioclimatic variables are
standard, commonly used in analyses of SDMs and were created in order to generate
more biologically meaningful variables than traditionally environmental ones (Hijmans et
al., 2005; O’Donnell & Ignizio, 2012). The 19 current bioclimatic variables used to run
the models were downloaded from Worldclim.org. To reduce statistic complexity of the
models, the number of bioclimatic variables were reduced to capture the entire
environmental conditions using the lowest number of bioclimatic variables. Firstly, to
reduce collinearity, the correlated variables were identified by Spearman’s correlation
test and Variance Inflation Factors (VIF) analyses. After, the biological and ecological
meaning of each variable were considered to remove one of each pair of correlated
variables. All these analyses resulted in 5 selected predictor variables out of 19: (1) Bio
3 (Isothermality) quantifies the monthly mean diurnal range (Bio 2) relative to the annual
7
temperature oscillation (Bio 7), and then multiplying it by 100. If this value is close to 100,
the daily diurnal temperature range is equivalent to the annual diurnal temperature range
(small level of temperature oscillation compared to annual variability), while a smaller
value indicates a large temperature variability. (2) Bio 4 (Temperature Seasonality) is a
measure of annual temperature variability calculated as the standard deviation of
monthly temperature averages multiplied by 100. The larger values indicate a greater
variability of monthly mean temperature. (3) Bio 8 (Mean Temperature of Wettest
Quarter) is calculated as the average temperature of the three consecutive months with
the highest cumulative precipitation. (4) Bio 9 (Mean Temperature of Driest Quarter) is
the average temperature of the three consecutive months with the lowest cumulative
precipitation. Finally, (5) Bio 15 (Precipitation Seasonality) is a measure of monthly total
precipitation variability estimated as the ratio of the standard deviation of the monthly
total precipitation to the mean monthly total precipitation, expressed as a percentage
(i.e., coefficient of variation). Larger values of Bio 15 indicate a greater variability of
precipitation. For predicting changes in future potential habitat distribution, seasonality
of both temperature and precipitation are important to be captured because species
distribution should be strongly influenced by precipitation variability and droughts events,
especially in freshwater ecosystems (Woodward et al., 2010). In general, all selected
variables are commonly utilized for examining how temperature and precipitation
variability may affect species seasonal distributions (O’Donnell & Ignizio, 2012).
For future predictions in 2050 and 2080, two climate global models that focused in the
land component were selected: HadGEM2 ES (Met Office Hadley Centre, MOHC, and
Instituto Nacional de Pesquisas Espaciais) and MPI ESM-MR (Max Planck Institute for
Meteorology, MPI-M). For both models, the two climatic scenarios RCP 2.6 and RCP 8.5
were considered in order to capture the lowest (optimist model) and the highest
(pessimist model) anthropogenic emissions, respectively. Future climatic conditions of
the 5 selected bioclimatic variables were downloaded from ccafs-climate.org in ASCII
format and 30 arc-second resolution.
In order to determine the current species distribution, the widely used Generalized Linear
Models (GLM), Generalized Additive Models (GAM) and Boosted Regression Trees
(BRT) were selected to running Species Distribution Models (SDM) for each species
separately. These three models uses different methods to infer species potential
distribution (Guisan et al., 2002; Franklin, 2009; Elith et al., 2008; Elith & Leathwick,
2009; Kienast et al., 2012): (1) GLM represents a flexible extension of linear models that
allows for response variables that have non-normal distribution error; (2) GAM is a non-
parametric extension of GLM. GAM models are very flexible because the linear predictor
8
is the sum of smoothing functions that are selected locally along the gradients of
predictor variables to find the best solution for the data; (3) BRT combines regression
trees and boosting algorithms. Regression trees results from classifications and decision
tree, while boosting builds and combines many simple models to give the best prediction.
All of these models were performed using “Stats” (Hastie & Pregibon, 1992; Venables &
Ripley, 2002), “gam” version 1.14 (Hastie and Tibshirani, 1990) and “gbm” version 2.1.1
(Ridgeway, 1999) packages of R 2 (R Core Development Team, 2013).
Since the original data included ”presence-only”, data processing were simplified
including background values ("random-absence") in each species database for GLM and
GAM, and pseudo-absences for BRT, thereby obtaining a matrix with pseudo-absences
(Franklin, 2009). Background and pseudo-absences values were generated with the R
function “RandomPoints” of Dismo package version 1.1-1 (Hijmans et al., 2016). Pseudo-
absences values differ of background because the spatial points with a present-data
point are excluded in the former.
To create an input SDM matrix, the current bioclimatic information was extracted for each
spatial point, i.e., for species occurrence and random-absences locations, and then the
obtained data-environment matrix was divided and independently created for each
species. To run the models, occurrence data were split into 70% as training set and 30%
as testing set by random partition. Models were evaluated by means of Area Under
Curve (AUC) statistics from a receiver-operating characteristic analysis, which is
threshold-independent evaluation of model discrimination (Fielding & Bell, 1997). AUC
values ranged from 0.5 to 1: 0.5 to 0.7 represents poor model performance, 0.7 to 0.9
represents moderate performance and values higher than 0.9 represents high
performance of model. For each species, the three models of SDMs were calibrated with
current environmental conditions and posteriorly compared by the AUC values. The
model with the highest AUC value and with current predicted distribution overlapping the
actual species distribution (Dijkstra et al., 2013) was chosen and used for predicting
future potential habitat. The probability distribution maps of current and future projections
were transformed into binary presence–absence maps to compare between them by
applying a cut-off value that minimises the difference between sensitivity (true-positive
predictions) and specificity (true-negative predictions, Lobo et al., 2007). The resulting
comparisons between current and future scenarios for each model and species were
used for sorting species in the five hypothesized categories (fig. 1).
Compilation of DNA sequences and Phylogenetic analyses
DNA sequences of all Iberian and Moroccan species were searched and compiled from
GenBank for the genes that were more frequently sequenced: the mitochondrial
9
cytochrome c oxidase subunit I gene (COI; 587 bps) and the ribosomal 12S RNA (12S
rRNA; 1774 bp), 16S RNA (16S rRNA; 542 bp), 18S RNA (18S rRNA; 1813 bp) and 28S
RNA (28S rRNA; 3933 bp). 4 Ephemeroptera species were used as outgroup (Baetis
harrisoni, Callibaetis ferrugineus, Ephemera danica and E. orientalis).
The alignment procedure was executed in MAFFT 7 (Katoh & Standley, 2013) using the
E-INS-i strategy (Very slow; recommended for <200 sequences with multiple conserved
domains and long gaps). The best-fit partitioning scheme and individual models of
molecular evolution for phylogenetic analyses were specified for gene partition, the best
model of substitution was determined using the AIC (Akaike Information Criterion) in
Partition Finder (Lanfear et al., 2012). Moreover, for the protein-coding gene COI
independent model of nucleotide substitution were performed for each of the three codon
positions that were treated as one partition. Gene partition was combined in a single data
supermatrix using MEGA 5.0 (Tamura et al., 2011). Two methods of phylogenetic
inference were used to reconstruct phylogenetic relationships. The maximum likelihood
was implemented with RAxML (Randomized Axelerated Maximum Likelihood)
(Stamatakis et al., 2008) under the GTR + Γ + I model with default number of Γ -
categories implemented independently for each codon position. The best trees were
selected from 100 multiple inferences, and clade support was assessed by means of
1000 nonparametric bootstrap resampling replicates of the original matrix. Bayesian
inference was conducted using MrBayes 3.2.5 (Ronquist & Huelsenbeck, 2003). Two
independent runs with four simultaneous Markov chain Monte Carlo (MCMC) chains (one
cold and three heated), each with random starting trees, were carried out simultaneously,
sampling 1000 generations until the standard deviation of the split frequencies of these
two runs dropped below 0.01 (10 million generations). Tracer 1.4
(http://evolve.zoo.ox.ac.uk/) was used to ensure that the MCMC chains had reached
stationarity by examining the effective sample size (ESS) values and to determine the
correct number of generations to discard as burn-in. The two phylogenetic analyses were
run remotely at the CIPRES Science Gateway (Miller et al., 2010). As conservative
measures of node support, a value of bootstrap of 80% or greater might indicate
substantial confidence for the maximum likelihood tree. In the Bayesian inference,
posterior probabilities should only be considered reliable if were greater than 0.95.
Species-specific habitat preferences and trait conservatism
Individual species information of adult habitat preferences were extracted from Dijkstra
(2006). This information was used to delimit ecological trait space for each species. Trait
information was quantified using a fuzzy coding approach (Chevene et al., 1994) and
compiled in a matrix including 11 traits and 34 categories (table S1): 3 morphological
10
traits (body length, abdomen size and posterior wing longitude), 3 distributional
characteristics (local abundance, distribution width and geographical distribution range),
4 traits of habitat preferences (lotic or lentic waters, seasonality of water, vegetation type,
water chemistry) and period of adult flight. All scores within each trait were standardised
so that the sums for a given species and a given trait were 1. In order to visualize how
traits and families were distributed based on their trait composition, trait categories
variance was measured conducting a Fuzzy Principal Component Analysis (FPCA,
Chevene et al., 1994).
To assess how taxonomy influence species ecological distribution or, in other words, if
families groups were ecologically differentiated or not, a Between-Classes Analyses
(BCA) were performed based in FPCA results. BCA can be considered as a particular
case of a Principal Component Analysis (because it is also a dimension reduction
technique) but aims to discriminate groups maximizing the differences between them.
BCA applied to PCA decomposes and orders only variance between groups, with the
idea of obtaining a few dimensions in preserving the maximum variance between the
centroids of the groups, not between individual observations (Dolédec & Chessel, 1987).
Next, a permutation test was executed in order to estimate a p-value and hence detect
differences between families distribution across the ecological space.
In order to assess character evolution and trait conservatism, the correlation of biological
and ecological trait variation on the phylogenetic tree was tested. Two indices were
calculated to infer possible patterns between traits and phylogeny: Pagel’s λ parameter
(Pagel, 1999; Freckleton et al., 2002) and Blomberg’s K-statistic (Blomberg et al., 2003).
Pagel’s λ parameter is based on maximum likelihood and gives a value between 0 and
1. If λ goes towards 1, the internal branches retain their original length indicating that
there is a strong correlation between trait and the phylogenetic tree, so niche is
preserved. Similarly, when the estimate of λ is close to 0, means that trait evolution has
not followed the tree topology, therefore, there is no trait conservatism. On the other
hand and for comparison purposes, the Blomberg’s K-statistic was also measured to test
the existence of a phylogenetic signal. The higher the K statistic, the more phylogenetic
signal in a trait. K values around 1 indicates that trait disparification follows the topology
of the tree, which implies some degree of phylogenetic signal or conservatism of traits,
whereas a little K means trait variation is independent to the phylogeny, which
corresponds to a random pattern of evolution. The principal difference between the two
methods is that Pagel’s λ compares all branches together while Blomberg’s K compares
pairs of branches. The test of significance these indexes also differ: Pagel’s λ uses
likelihood ratio tests against simpler models but Blomberg’s K makes randomizations of
11
the original trait data, comparing signal in a trait to the signal under a null model of trait
evolution on a phylogeny, concretely, the Brownian motion. These analyses were carried
out using two libraries of the R package: “geiger” (Harmon et al., 2008) and “picante”
(Kembel et al., 2010).
RESULTS
Final dataset included 85 Odonata species and 53845 individual records of “presence-
only” (fig. 2). 10 species occurred exclusively in Morocco and 26 were found limited in
the Iberian Peninsula. According to the Spain Red List of Invertebrates (Verdú et al.
(Eds), 2011), 12 species included in the dataset were classified as vulnerable, 3 as
endangered and another 3 as critical. Species distribution was strongly variable across
species (fig. 3; Supplementary Material). For instance, in one of the extremes,
Onychogomphus costae, Coenagrion scitulum and another 25 species were rare and
had a small potential distribution area; these are some of the most vulnerable species.
On the other extreme, Cordulegaster boltonii and Sympetrum fonscolombii were
abundant and widely distributed across the studied area. In an intermediate situation,
Aeshna cyanea and Calopteryx haemorrhoidalis were distributed mainly in the
Fig. 2. Map showing the 53845 Odonata species
records (presence-only) distributed across the
Iberian Peninsula and N-Africa. The four groups
conforming the database are indicated.
12
Mediterranean and Atlantic coast with patched occurrence on the Meseta Central. The
comprehension of changes in the bioclimatic data is important to understand future
prediction results (fig. 4). Despite seasonality will be greater in the future, the coast and
inland regions in the Iberian Peninsula presented strong differences with reference to
future isothermality (Bio 3) and temperature seasonality (Bio 4). Coast areas presented
higher isothermality and lower temperature seasonality, which means less annual
variability than inlands regions likely due to the sea effect, noteworthy wider in Atlantic
than in Mediterranean coasts. Regarding to the mean temperature of the wettest quarter
(Bio 8), the entire eastern half of the Peninsula presented the highest values, while Plana
de Vic in northeast of Peninsula and highlands showed the lowest value of the mean
temperature of the driest quarter (Bio 9). It means that the wettest months will be colder
than in present-day and that the driest months will be hotter, except in Plana de Vic and
highlands. Precipitation seasonality (Bio 15) had the highest values in south-west of the
Iberian Peninsula and along the Moroccan coast, but this seasonality will decrease and,
therefore, precipitations will vary less during the year. Also, bioclimatic variables will
change over time. For instance, isothermality will decrease and, therefore, temperature
seasonality will increase from 2050 to 2080. Moreover, the mean temperature of the
Fig. 3. Potential distribution area represented as a probability (0-1) for a couple of abundant
and wide-distributed species (Cordulegaster boltonii and Boyeria irene), two Mediterranean
(Sympetrum fonscolombii Calopteryx haemorrhoidalis) and a pair of rare and with a constrained
distribution (Coenagrion scitulum and Onychogomphus costae).
13
wettest quarter will tend to decrease as a result of a shift in the rainfall season to either
a delay in autumn or an advanced in spring. In contrast, the mean temperature of the
driest quarter will increase. Finally, precipitation seasonality is expected to decrease in
future scenarios owing to more drought events.
BRT models provided contribution percentages of variables fitted in the model. Bio4 and
Bio15 were the predictors that commonly influenced the most to SDMs, whereas Bio9
had the lowest influence. Among species, in general, both the seasonality (i.e., annual
variability) in temperature and precipitation influenced more than predictors based
exclusively in temperature range per se.
Geographical occurrence allowed to run SDMs for 64 species out of 85 included in the
dataset, the discarded species were rare and locally distributed (e.g., Cordulegaster
bidentata that was recorded in 30 nearby sites), many classified as “vulnerable” in
ecological traits. From those 64, 41 species were modelled using GAM, another 12
species were modelled using GLM and, finally, 11 were modelled using BRT (table S2).
Remarkably, species modelled under BRT were confined in small areas at either
Morocco, Morocco plus south of the Iberian Peninsula or northern Spain, whereas many
species modelled using GLM or GAM showed a larger geographical range. The analysed
64 species fell into one of the predicted categories, however high discrepancies were
found between pessimist (high emissions, RCP 8.5) and optimist (low emissions, RCP
2.6) future scenarios under both future models, as expected (fig. 5, table 1 & S3).
Predictions of MPI in 2050th for both RCP scenarios showed a half of species as
“expansive” (category 2), while HadGEM2 showed a 40.6% of species as “displaced”
(category 1) for RCP 2.6 scenario, whereas a 43.8% of species were assigned as
“reduced” (category 4) for RCP 8.5 scenario. In 2080th, predictions for RCP 2.6 scenario
of MPI showed a 42.2% of species as “non-change” (category 3), while predictions of
HadGEM2 showed a 42% of species as “displaced”. Predictions of both future models
for RCP 8.5 showed half of species as “reduced” (category 4). Notably, RCP 8.5
scenarios showed larger percentages of species classified as “extinct”, and the highest
value was detected in 2080 for HadGEM2 under RCP 8.5, as predicted.
14
Fig. 4. Evolution of bioclimatic variables selected using HadGEM2 model under a RCP 8.5
scenario. Bio3: Isothermality (%), Bio4: Temperature seasonality (stdev x100), Bio8: Mean
temperature of wettest quarter (ºC x10), Bio9: Mean temperature of driest quarter (ºC x10),
Bio15: Precipitation seasonality (CV x100).
15
2050 2080
MPI HadGEM2 MPI HadGEM2
RCP 2.6 RCP 8.5 RCP 2.6 RCP 8.5 RCP 2.6 RCP 8.5 RCP 2.6 RCP 8.5
Ca
teg
ori
es 1 29,6875 34,375 40,625 29,6875 15,625 40,625 42,1875 20,3125
2 57,8125 46,875 37,5 21,875 34,375 10,9375 4,6875 6,25
3 7,8125 3,125 1,5625 1,5625 42,1875 1,5625 26,5625 10,9375
4 4,6875 12,5 18,75 43,75 7,8125 42,1875 26,5625 54,6875
5 0 3,125 1,5625 3,125 0 4,6875 0 7,8125
Table 1. Percentages of species potential distribution categories for each model (MPI and
HadGEM2 ES) and scenario (RCP 2.6 and RCP 8.5). Values >40% in bold. Categories: 1:
potential area is latitudinally or longitudinally displaced, 2: potential area is being expanded, 3: no
significant differences between current and future potential areas distribution, 4: habitat lost, 5:
significant reduction of potential area.
RAxML phylogenetic tree (fig. 6) included 2 suborders, 9 families and 73 species (80.2%
of the species list was covered), and the final matrix contained 8649 bp. The species
coverage was 67.12%, 58.90%, 72.60%, 72.60%, and 57.53%, for COI, 12S, 16S, 18S
and 28S genes, respectively (table S4). The two suborders Anisoptera and Zygoptera
were monophyletic, in agreement with other previous phylogenetic analyses (Dumont et
al., 2010; Dijkstra & Kalkman, 2012; Suhling et al., 2015). Zygoptera was divided in two
supported clades “Lestomorphs” and “non-Lestomorphs”. In the Iberian Peninsula,
Lestomorphs were composed by three genera: Chalcolestes, Lestes and Sympecma,
whereas the remaining genera Platycnemis, Ceriagrion, Pyrrhosoma, Ischnura,
Enallagma, Coenagrion, Erythromma and Calopteryx were clustered together in “non-
Lestomorphs”. Anisoptera order was separated into two clades: (1) families
Cordulegasteridae, Gomphidae, and Aeshnidae, and (2) families Corduliidae and
Libellulidae as indicated in Dumont et al. (2010). The first clade was well-resolved, and
all genera were placed as previously showed in Ware et al. (2007) and Dumont et al.
(2010). However, the monophyletic family Libellulidae did not have much supported
nodes, but genera position corresponds to Dijkstra & Kalkman (2012), except for the
genera Zygonyx and Crocothemis that were placed in the same clade. Finally,
Corduliidae was placed paraphyletic as Dijkstra et al. (2013) stated, but in contrast of
Dumont et al. (2009) that found this family as monophyletic.
Traits conservatism was rejected because Blomberg’s K and Pagel’s λ values in both
axes were close to 0 (K=0.000192 and λ=0.50 for the first axis, and K=0.000694 and
λ=0.37 for the second axis; p-value>0.05; table 2). In general, individual traits were
neither preserved in the phylogeny, except large body, wings and abdomen sizes that
were phylogenetically conserved in the two indexes (large body: K=3.15, λ=1; large
abdomen: K=1.27, λ=1; large wings: K=1.11, λ=1). All in all, these results indicated no
trait correlation in the phylogenetic tree and no phylogenetic signal of traits evolution. In
16
other words, the ecological similarity of closely related species was decoupled to their
phylogenetic relations.
Odonata species commonly showed a high habitat specificity (Hassall & Thompson,
2008; Dijkstra, 2006; Suhling et al., 2015). It means that each species inhabited one of
the categories for each defined habitat preference. For instance, species that preferred
lentic water never were found in lotic habitats, or species located in riparian forest were
exclusive for this habitat. For example, Calopteryx virgo was only found in permanent
forested streams, Sympetrum meridionale preferred permanent vegetated ponds and
Cordulegaster boltonii was found in pristine highland streams. We also found few
generalist species, for instance Anax imperator inhabited in all types of lotic habitats, or
all Trithemis species inhabited permanent but also temporary lotic waters.
Distribution of trait categories across niche space explained a 23.59% of the total
variability on axis 1, whereas axis 2 and 3 explained a 19.73% and a 14.99%,
respectively (fig. 7). The morphological small sizes and also abundant and continuous
distributed species were the main contributors in the trait variance on the positive sides
on axis 1, whereas morphological medium sizes and also rare and fragmented
distributed species contributed negatively. In contrast, on axis 2 medium body size and
abundant and continuous distributed species were the main contributors on the positive
sides, while small body size and fragmented and rare distributed species contributed
negatively. Large morphological sizes contributed positively on axis 3, while medium
morphological sizes were negatively (fig. 7a).
The FPCA of traits disparity within species revealed how ecological characters are
distributed among families, indicating that each family was characterised by a certain
biological and ecological preferences and identity. In fact, permutation test of variance
between groups indicated significant differences between families in relation to traits
disparity (p-value = 0.01). Given the high specificity for preferred habitat, the families of
Odonata showed high variability of how individual species were distributed in relation to
traits disparity. Some families had high trait disparity (i.e., large circle area in fig. 7b) such
as Lestidae and Libellulidae, whereas other families showed low trait disparity such as
Platycnemididae.
In fig. 7c species were grouped as currently vulnerable because they inhabit high
elevations (blue labels) or their habitat distribution was in regression (red labels) against
species that will expand geographical range (yellow label) or have the capacity to do it
(green labels). None of these groups were plotted nearby on the FPCA axes. Similarly,
we also labelled species according to their category in the hypotheses (fig. 7d for 2050
17
and fig. 7e for 2080): blue labels showed species grouped as “favoured” because
potential distribution will expand, in contrast to vulnerable or critical species that their
potential distribution will be reduced (orange labels), and in some cases their potential
habitats will disappear (red labels). Some species were classified as “stable” distribution,
since potential distribution maybe shift, but the total area is equal (green label if shifts
and yellow label if is the same region). It did not show any pattern, i.e., species distributed
across the trait space did not reveal if a species is going to expand or reduce their
potential distribution.
18
Fig. 5. Examples of each hypothesized category. Square colours correspond to the code
indicated in fig. 1. The colour of the areas showed in current, 2050th and 2080th maps have
their legend below (colours key).
19
Fig. 6. Maximum-likelihood phylogenetic tree of 5 genes of Order Odonata including 31 genera
belonging to 9 families. Comprises 77 taxa including 4 outgroup genera of Ephemeroptera. Branch
support percentage is indicated: maximum likelihood bootstrap (>80) / Bayesian inference (>0.95),
*refers to branches with discordances between the two phylogenetic inferences. Suborder
classification is given by coloured branches: Zygoptera and Anisoptera.
20
Fig. 7. Fuzzy Principal Component Analysis (FPCA) on trait data. (a) Distribution of functional traits on FPCA axes. See Table S1 for traits labels code. (b)
Distribution of the families on FPCA axes. Family names code is: Ca, Calopterygidae; Ce, Coenagrionidae; Le, Lestidae; P, Platycnemidae; A, Aeshnidae;
Cg, Cordulegasteridae; Cl, Corduliidae; G, Gomphidae; Li, Libellulidae. See Table S4 for species code. The two suborder groups are represented in the
same colours as phylogenetic tree in figure 5. (c) Current species status. Red labels: vulnerable, blue: high altitude habitats, green: potentially expansive,
yellow: expanding. (d-e) Species labelled as the category which they belong in 2050 (d) or 2080 (e). Colours correspond to the code indicated in fig. 1.
Axes 1 and 2 explained 23.59 and 19.73% of the total variability, respectively.
21
Category Traits Code K PIC.variance.P λ M
orp
ho
log
y
Body Size
<40 BS 0.00 0.23 0.87
>40 - <60 BM 0.00 0.30 0.89
>60 BL 3.15 0.00 1
Abdomen Size
<30 AS 0.44 0.00 0.86
>30 - <45 AM 0.30 0.00 0.88
>45 AL 1.27 0.00 1
Wing (posterior) Size
<30 WS 0.00 0.31 0.81
>25 - <38 WM 0.00 0.25 0.88
>38 WL 1.11 0.00 1
Dis
trib
uti
on
Density Abundant AB 0.00 0.60 0
Rare RA 0.00 0.63 0
Distribution
Continuum CO 0.09 0.13 0
Fragmented FR 0.09 0.14 0
Vulnerable VU 0.13 0.51 0
Range
North-Africa NA 0.00 0.81 0
Iberian Peninsula PI 0.00 0.68 0.54
Europe EU 0.00 0.12 0
Africa AF 0.44 0.00 1
Hab
ita
t p
refe
ren
ces
Habitat
River RV 0.15 0.08 0.64
Streams ST 0.25 0.00 0.81
Ponds PD 0.00 0.41 0.55
Lakes LK 0.00 0.40 0
Seasonality Temporal TP 0.00 0.83 0
Permanent PR 0.00 0.84 0
Vegetation
Submergent SB 0.16 0.09 0
Floating-leaf FL 0.09 0.23 0.14
Ruderal RO 0.12 0.03 0.55
Forest, riparian FO 0.26 0.02 0.84
Water Chemistry
Acid AC 0.08 0.57 0
Saline SA 0.00 0.97 0
Ad
ult
s
Flight
Winter: I-II WI 0.00 0.99 0
Spring: III-V SP 0.00 0.77 0
Summer: VI-IX SU 0.00 0.79 0.21
Autumn: X-XII AU 0.00 0.77 0.43
Axis 1 of the FPCA 0.000192 0.60 0.50
Axis 2 of the FPCA 0.000694 0.22 0.37
Table 2. Niche conservatism results expressed as Blomberg’s K and Pagel’s λ indices. The higher
the K statistic, the more phylogenetic signal in a trait: K = 0 means a random pattern of evolution,
trait variation has not followed the phylogeny; K = 1 indicates some degree of conservatism; when
K > 1 there is a strong phylogenetic signal and conservatism of traits (bold numbers). Traits with
PIC.variance.P < 0.05 have non-random phylogenetic signal. In the same way, when λ is close
to 1, internal branches retain their original length indicating a strong trait correlation with the tree;
whereas when λ go towards 0, trait evolution has not followed the tree topology.
22
DISCUSSION
Future climate predictions under the four scenarios evidenced an increase in seasonality
of temperature and a decrease in rainfall variability. In general, the predicted
environmental characteristics of Mediterranean basin and inland of Iberian Peninsula will
be harsher than the Atlantic coast due to the increase of drought events, which will be
buffered by change in the atmospheric circulation (Giorgi & Lionello, 2008). As expected,
the predicted future climate conditions will directly impact on future species distribution
of odonates, but these responses will be highly variable across species and predictions.
For instance, several species will lose their potential distribution area in Iberian Peninsula
(e.g., Sympetrum flaveolum), in opposition, other species will shift or increase their
potential distribution areas (e.g., Anax ephippiger). Across predictions, differences were
evident between the most pessimist model HadGEM2 versus MPI global climate models
and between the highest RCP 8.5 versus the lowest RCP 2.6 emissions scenarios. In
general, if only the environmental conditions are considered, the predictions for 2080 for
the most severe conditions showed that the majority of species (54.69%) will reduce their
potential distribution (category 4) and around 7.81% will go extinct (category 5). In
contrast, the most favourable conditions revealed a subtle dominance of species that
expand their potential distribution (category 2) or species that will displace their
geographical distribution (category 1) to habitats where the ecological conditions will be
more suitable Hence, the climatic models used to predict the future distribution of
odonates indicate that human measures for reducing anthropogenic emissions are
critical for ensuring habitat conservation and preserving the present-day diversity of
species.
Despite I revealed how important is to model species-specific effects of global warming
on abiotic conditions for multiple species, the high variability across species disallow me
to establish generalizations. Since the majority of species will reduce their distribution
range in similar or higher emissions than currently, the species-specific biotic responses
of odonates facing climate change will be critical for occupying or not the new potential
distribution. In general, species could respond to climate change (1) moving in space or
time to remain in a constant environmental niche, or by (2) evolutionary adaptation or (3)
phenotypic acclimation (plasticity) (Parmesan, 2006; Buckley & Kingsolver, 2016).
Odonata is one of the freshwater lineage that shows more abilities to face climate change
for their multiple pathways for adaptive thermoregulation, such as the production of body
pigment (Hassall & Thompson, 2008), but also for their aerial long-dispersal capacity
that favours northward expansion (Heino et al., 2009; Markovic et al., 2014). Empirically,
northward expansions as response of climate change are common for many taxa of high-
23
dispersive insects (Parmesan et al., 1999; Parmesan & Yohe, 2003; Hitch & Leberg,
2007) and it is also well documented in odonates from North-Africa to South-Europe
(Cano-Villegas & Conesa-García, 2009) and from South-Europe to North-Europe
(Hickling et al., 2006; Ott, 2010). However, my results did not support a massive
northward expansion of any specific linage or trait characteristics of odonates following
habitat change, in contrast, obtained patterns of future potential species distribution
indicated idiosyncratic responses when the entire diversity of odonates form the
westernmost Mediterranean regions is considered.
The Iberian Peninsula harbours a high diversity of odonates that differ in distribution
ranges and habitat specificity, but I did not find a clear pattern between analysed
biological traits, current habitat preferences and future suitable habitats. For many
species, Iberian Peninsula is either the southernmost region of their European
distribution (e.g., Brachytron pratense), the northernmost region of their African
distribution (e.g., Diplacodes lefebvrii) or the centre (e.g., Oxygastra curtisii) of their
Mediterranean distribution (Dijkstra, 2006). It is expected varying ecological and
functional traits across species with disparate distribution range, which suggests the high
diversity of differences in ecological requirements among these species. Biological and
ecological traits determine niche occupancy. For instance, freshwater species and also
odonates could be classified by their preferred water temperature conditions in warm-,
cool- and cold-water types. Climate warming will favour warm-water species opposite to
cold-water species because temperature will tend to increase (Heino et al., 2009). In fact,
theoretically species can be classified as vulnerable (e.g., bad flyers, low tolerance to
eutrophication) or favourable (e.g., mechanism of resistance to drought events or high
temperatures) depending on their response facing climate change (Ott, 2010), but I found
here that analysed biological traits and future SDMs are decoupled. It means that I
cannot predict future direction of the species responses facing climate change and
therefore there are not favourable traits to face global warming expanding to new
habitats. For instance, species with favourable traits in front of drought events (e.g.,
preference for temporary habitats such as Sympetrum flaveolum) will not expand if their
future potential distribution area is not expected to increase in parallel. This study reveals
that ecological and life-history traits are not good predictors of species shifts for
odonates, which was also previously suggested for American songbird species (Auer
and King, 2014).
The evolutionary signature of trait conservatism is not preserved, except size-related
traits such as abdomen, body and wings length. The only conserved traits are associated
also to wing patterns that can easily distinguish the two monophyletic suborders of
24
Odonata (Anisoptera and Zygoptera) (Dumont et al., 2010; Dijkstra & Kalkman, 2012).
The non-preserved ecological traits can be explained by several mechanisms. Odonata
have tropical origins and are one of the oldest winged insects that still inhabit the earth.
These taxa belong to Odonatoptera, first appearing back minimum to the Upper
Carboniferous period (~300Ma) (Suhling et al., 2015). Such old lineages must have
faced multiple climate changes and adverse environmental conditions over millions of
years that could promote repeated and independent extinction and speciation across
clades, which likely explain some of the lack of phylogenetic signal. In contrast, young
insect orders such as Trichoptera have been appeared more recently and the
phylogenetic signal is still preserved in current species (Garcia-Raventós et al., in
preparation). Moreover, Odonata diversification is more related to sexual morphology,
reproductive behaviour and interactions between species than adaptive ecological
divergence (Wellenreuther et al., 2012), which in turn can affect the loss of a
phylogenetic signal of niche conservatism. As a result, the ecological space delimited by
species traits and the high habitat-specificity of odonates did not show significant
patterns across lineages. In other words, habitat preference such as temporary pond
non-vegetated can be occupied by different species across lineages indistinctively. Thus,
the phylogeny of Odonata is not useful to elucidate which traits are characteristic in each
lineage and if this given trait can predict species vulnerability facing climate warming.
Consequently, I cannot attribute to a lineage the probability of neither extinction,
northward range expansion nor shift in its distribution range. However I did not test
species plasticity and local adaptation, therefore, there are multiple unexplored biotic
factors like species interaction, trophic networks, hot-tolerance, resistance to drought
events, etc. (Hassall & Thompson, 2008) that must influence species colonization and
establishment to new habitats. Further studies modelling multi-species distribution
considering intraspecific traits and genetic variability are needed to infer future species-
specific distribution and extinction risk in order to do a correct management of freshwater
biodiversity under climate change.
25
CONCLUSIONS
As a main result of this project, I found odonates future potential distribution will be
affected by Climate Change. Many species will shift their potential distribution to new
suitable habitat, but the rates and directions at which species will achieve the new
localities are species-specific. Other species will reduce their potential distribution and
their locally adaptability or phenotypically acclimatise will allow them avoid extinction risk.
Despite each Odonata family showed their own ecological space, which was
differentiated between families, the niche conservatism was rejected because traits were
not preserved in phylogenetic tree of Iberian and Moroccan odonates. These results
indicated that the ecological distance between species including also closely related
species was decoupled to their phylogenetic divergence. Therefore, phylogeny cannot
predict the ecological requirements of species.
Although species have high habitat-specificity, none clear pattern was found between
traits (ecological and life-history), current habitat occupancy and future potential
distribution under several models of Climate Change.
Understanding how Odonata will response in front of a Climate Change is critical to carry
out a correct management in order to protect vulnerable species and maintain freshwater
biodiversity.
ACNOWLEDGEMENTS
I would like to acknowledge the help of Oxygastra and AEA El Bosque Animado groups
who gently provided the occurrences data. I am especially gratefull to Xavier Maynou
and Ricard Martín from Oxygastra for field training and their sound advices, and also to
Florent Prunier from AEA El Bosque Animado for his recommendations. I am very
grateful to María Ángeles Pérez for her great help with SDM and also to Dr. Dani Sol and
his group for share their server-resources. Dr. Núria Bonada and Tony Herrera are
thanked for their valuable comments and contacts, and also the research group FEM
(Freshwater Ecology Management) for providing the space, equipment and
unconditional encouragement for the investigation. Finally, I really appreciate the harsh
job, collaboration and nearby support of my lab team (Macro&EvoLAB), my advisor Dr.
Cesc Múrria and Aina Garcia-Raventós.
26
REFERENCES
Auer, S. K., & King, D. I. (2014). Ecological and life-history traits explain recent boundary shifts in elevation and latitude of western North American songbirds. Global Ecology and Biogeography, 23(8), 867–875.
Blomberg, S. P., Garland Jr, T., & Ives, A. R. (2003). Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57(4): 717-745.
Boudot J.-P., Kalkman V.J., Azpilicueta Amorín M., Bogdanović T., Cordero Rivera A., Degabriele G., Dommanget J.-L., Ferreira S., Garrigós B., Jović M., Kotarac M., Lopau W., Marinov M., Mihoković N., Riservato E., Samraoui B. & Schneider W. (2009). Atlas of the Odonata of the Mediterranean and North Africa. Libellula Supplement 9: 1-256
Buckley, L. B., & Kingsolver, J. G. (2016). Functional and Phylogenetic Approaches to Forecasting Species’ Responses to Climate Change. Annual Review of Ecology, Evolution and Systematics, 43, 205–26.
Burgmer, T., H. Hillebrand, and M. Pfenninger. (2007). Effects of climatedriven temperature changes on the diversity of freshwater macroinvertebrates. Oecologia 151:93–103. doi:10.1007/s00442-006-0542-9
Cano-Villegas, F.J. & Conesa-García, M.A. (2009). Expansión de Trithemis kirbyi Sélys, 1891 (Odonata: Libellulidae) en la provincia de Málaga (sur de la Península Ibérica). Bol. S.E.A., 44: 569-572.
Chelmick, D.G. & Pickess, B.P. (2008). Trithemis kirbyi Selys in southern Spain. Notulae Odonatologicae, 7: 4-5.
Chevene, F., Dolédec, S., & Chessel, D. (1994). A fuzzy coding approach for the analysis of long‐term ecological data. Freshwater Biology 31(3): 295-309.
Compte Sart, A. (1962). Un odonato nuevo para España, Brachythemis leucosticta (Burm.). Vie Milieu 13: 604-607.
Conesa Garcia, M. Á. (1985). Aportaciones a la biolog¡a de Diplacodes lefebvrei (Rambur, 1842) (Odon., Anisop., Libellulidae) en la Península Ibérica. Boletín de La Asociación Española de Entomología, 9, 321–330.
Daufresne, M., Bady, P., & Fruget, J. F. (2007). Impacts of global changes and extreme hydroclimatic events on macroinvertebrate community structures in the French Rhône River. Oecologia, 151(3), 544–559.
Davies TJ, Purvis A, Gittleman JL. (2009). Quaternary climate change and the geographic ranges of mammals. Am. Nat. 174:297–307.
Dijkstra, K. D. B., (2006). Field guide to the dragonflies of Britain and Europe including western Turkey and north-western Africa. British Wildlife Publishing. Dorset, UK. 320 pp.
Dijkstra, K. D. B., Bechly, G., Bybee, S. M., Dow, R. a., Dumont, H. J., Fleck, G., Ware, J. (2013). The classification and diversity of dragonflies and damselflies (Odonata). Zootaxa, 3703(1), 36–45.
Dijkstra, K.-D.B. & Kalkman, V.J. (2012) Phylogeny, taxonomy and classification of European dragonflies and damselflies (Odonata): a review. Organisms, Diversity and Evolution, 12, 209–227.
Dolédec, S. and Chessel, D. (1987) Rythmes saisonniers et composantes stationnelles en milieu aquatique I- Description d'un plan d'observations complet par projection de variables. Acta Oecologica, Oecologia Generalis, 8, 3, 403–426.
Dumont, H. J., Vierstraete, A., & Vanfleteren, J. R. (2010). A molecular phylogeny of the Odonata (Insecta). Systematic Entomology, 35(1), 6–18.
Elith, J., Leathwick, J. R., & Hastie, T. (2008). A working guide to boosted regression trees. Journal of Animal Ecology, 77(4), 802–813.
Elith, J., & Leathwick, J. (2009). Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics, 40, 677–697.
Fielding A.H. & Bell J.F. (1997) A review of methods for the assessment of prediction errors in conservation presence ⁄ absence models. Environmental Conservation, 24, 38–49.
Franklin, J. (2009) Mapping species distributions: spatial inference and prediction. Cambridge University Press, Cambridge, UK.
27
Freckleton, R. P., Harvey, P. H., & Pagel, M. (2002). Phylogenetic analysis and comparative data: a test and review of evidence. The American Naturalist 160(6): 712-726.
Garcia-Raventós, A., Viza, A. & Múrria, C. (In preparation). On the origin and assembly of communities: Diversity patterns of Trichoptera across a latitudinal gradient in Europe.
Giorgi, F., & Lionello, P. (2008). Climate change projections for the Mediterranean region. Global and Planetary Change, 63(2–3), 90–104.
Guisan, T.C. Edwards, T. Hastie Jr. (2002). Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecol. Model. 157, pp. 89–100
Harmon, L. J., Weir, J. T., Brock, C. D., Glor, R. E., & Challenger, W. (2008). GEIGER: investigating evolutionary radiations. Bioinformatics 24(1): 129-131.
Hartung, M. 1985. Orthetrum trinacria (Sel.) as a part of the dragonfly fauna in southern Spain (Anisoptera: Libellulidae). Notulae Odonatologicae, 2(6): 101–102.
Hassall, C., & Thompson, D. J. (2008). The effects of environmental warming on Odonata: a review. International Journal of Odonatology, 11(2), 131–153.
Hastie, T., Tibshirani, R.J., (1990). Generalized Additive Models. Monographs on Statistics and Applied Probability, vol. 43. Chapman and Hall, London, 335 pp
Hastie, T. J. and Pregibon, D. (1992) Generalized linear models. Chapter 6 of Statistical Models in S eds J. M. Chambers and T. J. Hastie, Wadsworth & Brooks/Cole.
Heino, J., R. Virkkala, and H. Toivonen. (2009). Climate change and freshwater biodiversity: detected patterns, future trends and adaptations in northern regions. Biological Reviews of the Cambridge Philosophical Society 84:39–54.
Herrera-Grao, T., Bonada, N., Gavira, Ó. & Blanco-Garrido, F. (2012). First record of Trithemis kirbyi Sélys, 1891 in Catalonia (Odonata , Libellulidae ). Boln. Asoc. esp. Ent., 36 (3-4): 457-459.
Hickling, R., Roy, D. B., Hill, J. K., Fox, R., & Thomas, C. D. (2006). The distributions of a wide range of taxonomic groups are expanding polewards. Global Change Biology, 12(3), 450–455.
Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, P. G., & Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 25(15), 1965–1978.
Hitch AT, Leberg PL (2007) Breeding distributions of North American birds moving north as a result of climate change. Conserv Biol 21:534–539
Hutchinson, G.E. (1957). Concluding remarks. Cold Spring Harb. Symp. Quant. Biol., 22, 415-427.
Kienast, F., Bolliger, J., & Zimmermann, N. E. (2012). Species Distribution Modeling (SDM) with GLM, GAM and CART. Lecture Notes, (1966), 1–16.
Jaquemin G. & Boudot, J.P. (1999). Les libellules (Odonates) du Maroc. Société Française d’odonatologie.
Jeschke, J.M. & Strayer, D.L. (2006) Determinants of vertebrate invasion success in Europe and North America. Global Change Biology, 12, 1608–1619.
Katoh, K., & Standley, D. M. (2013). MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution 30(4): 772-780.
Kembel, S. W., Cowan, P. D., Helmus, M. R., Cornwell, W. K., Morlon, H., Ackerly, D. D., Blomberg, S. P., & Webb, C. O. (2010). Picante: R tools for integrating phylogenies and ecology. Bioinformatics 26(11): 1463-1464.
Lanfear, R., Calcott, B., Ho, S. Y., & Guindon, S. (2012). PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution 29(6): 1695-1701.
Lobo, J. M., Jiménez-valverde, A., & Real, R. (2008). AUC: A misleading measure of the performance of predictive distribution models. Global Ecology and Biogeography, 17(2), 145–151.
López, R. (1983). Odonatos de la Sierra de Grados. Actas del Congreso Ibérico de Entomología. 1983.
28
Markovic, D., Carrizo, S., Freyhof, J., Cid, N., Lengyel, S., Scholz, M., … Darwall, W. (2014). Europe’s freshwater biodiversity under climate change: Distribution shifts and conservation needs. Diversity and Distributions, 20(9), 1097–1107. http://doi.org/10.1111/ddi.12232
Miller, M. A., Pfeiffer, W., and Schwartz, T. (2010). Creating the CIPRES science gateway for inference of large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop (GCE), New Orleans, LA, 1–8.
O’Donnell, M. S., & Ignizio, D. A. (2012). Bioclimatic Predictors for Supporting Ecological Applications in the Conterminous United States. U.S Geological Survey Data Series 691, 10.
Ott, J. (2010). Dragonflies and climatic change - recent trends in Germany and Europe. BIORISK – Biodiversity and Ecosystem Risk Assessment, 5(0), 253–286.
Pagel, M. (1999). Inferring the historical patterns of biological evolution. Nature 401: 877-884.
Parmesan, C., Ryrholm, N., Stefanescu, C., Hill, J.K., Thomas,C.D., Descimon, H., Huntley, B.,, Kaila, L., Kullberg, J., Tammaru, T., Tennent, W.J., Thomas, J.A. & Warren, M. (1999). Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399, 579-583.
Parmesan, C. & Yohe, G. (2003). A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37-42.
Parmesan, C. (2006). Ecological and Evolutionary Responses to Recent Climate Change. Annual Review of Ecology, Evolution, and Systematics, 37, 637–669.
Perry A, Low P, Ellis J, Reynolds J. (2005). Climate change and distribution shifts in marine fishes. Science; 308: 1912–15.
R Core Development Team (2013) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
Ridgeway, G. (1999). “The state of boosting,” Computing Science and Statistics 31:172-181
Ronquist, F., & Huelsenbeck, J. P. (2003). MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19(12): 1572-1574.
Sala, O. E., Chapin, F. S., Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo, R., … Wall, D. H. (2000). Global Biodiversity Scenarios for the Year 2100. Science, 287(March), 1770–1774.
Schloss, C.A., Nunez, T.A. & Lawler, J.J. (2012) Dispersal will limit ability of mammals to track climate change in the Western Hemisphere. Proceedings of the National Academy of Sciences USA, 109, 8606–8611.
Stamatakis, A., Hoover, P., & Rougemont, J. (2008). A rapid bootstrap algorithm for the RaxML web servers. Systematic Biology 57(5): 758-771.
Stoks, R., A. N. Geerts, and L. De Meester. 2014. Evolutionary and plastic responses of freshwater invertebrates to climate change: realized patterns and future potential. Evolutionary Applications 7: 42–55.
Suhling, F., Sahlen, G., Gorb, S., Kalkman, V., Dijkstra, K.-D.B. & van Tol, J. (2015a) Order Odonata. Ecology and General Biology (ed. by J. Thorp and D. C. Rogers), pp. 893–932. Academic Press, London, U.K.
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., & Kumar, S. (2011). MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28(10): 2731-2739.
Testard D, P. (1975). Note sur l'émergence, le sexratio et l'activité des adultes de Mesogomphus genei Selys, dans le Sud de l'Espagne (Anisoptera: Gomphidae). Odonatologica, 4: 1126.
Travis, J.M.J. (2003) Climate change and habitat destruction: a deadly anthropogenic cocktail. Proceedings of the Royal Society of London Series B-Biological Sciences, 270, 467–473.
Torralba-Burrial, A. & Ocharan, F.J. (2007). Composición biogeográfica de la fauna de libélulas (Odonata) de la Península Ibérica, con especial referencia a la aragonesa. Bol. Soc. Entomol.Aragon, 41: 179-188.
Torralba-Burrial A, Ocharan FJ (2013 -). BOS Arthropod Collection Dataset: Iberian Odonata distribution. 16604 data records. Contributed by Torralba-Burrial A, Ocharan FJ, Outomuro D, Ocharan R, Salona MI, Benitez-Donoso A, Martinez JA, Martinez SR, and Brotons-Padilla M. Online at http://www.gbif.es:8080/ipt/archive.do?r=Bos-Odo and http://www.unioviedo.es/BOS/Zoologia/artropodos/odonata, version 1.0 (last updated on 2013-04-08), GBIF key: http://gbrds.gbif.org/browse/agent?uuid=7e31baf8-f762-11e1-a439-00145eb45e9a. Data
29
Venables, W. N. and Ripley, B. D. (2002) Modern Applied Statistics with S. New York: Springer (4th ed).
Verdú, j. r., Numa, C. y Galante, E. (Eds) 2011. Atlas y Libro Rojo de los Invertebrados amenazados de España (Especies Vulnerables). Dirección General de Medio Natural y Política Forestal, Ministerio de Medio Ambiente, Medio rural y Marino, Madrid, 1.318 pp
Walther, G. R., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T. J. C., … Bairlein, F. (2002). Ecological responses to recent climate change. Nature, 416(6879), 389–395.
Ware, J., May, M., & Kjer, K. (2007). Phylogeny of the higher Libelluloidea (Anisoptera: Odonata): An exploration of the most speciose superfamily of dragonflies. Molecular Phylogenetics and Evolution, 45(1), 289–310.
Webb, C.O., Ackerly, D.D., McPeek, M.A. & Donoghue, M.J. (2002). Phylogenies and community ecology. Annu. Rev. Ecol. Syst., 33, 475–505.
Wellenreuther, M., Larson, K. W., & Svensson, E. I. (2012). Climatic niche divergence or conservatism? Environmental niches and range limits in ecologically similar damselflies. Ecology, 93(6), 1353–1366.
Wiens, J. J., & Graham, C. H. (2005). Niche Conservatism: Integrating Evolution, Ecology, and Conservation Biology. Annual Review of Ecology, Evolution, and Systematics, 36(1), 519–539.
Woodward, G., Perkins, D. M., & Brown, L. E. (2010). Climate change and freshwater ecosystems: impacts across multiple levels of organization. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 365(1549), 2093–2106.
Wrona, F. J., Prowse, T. D., Reist, J. D., Hobbie, J. E., Levesque, L. M. J. & Vincent, W. F. (2006). Climate change effects on aquatic biota, ecosystem structure and function. Ambio 35, 359–369.
Su
bo
rder
Fa
mily
Species Code
Body size Ab. Distrib. Range Habitat Seas. Vegetation Chem. Adult
TO
AB
AP
AB
RA
CO
FR
VU
NA
PI
EU
AF
RV
ST
PD
LK
TP
PR
SB
FL
RO
FO
AC
SA
Flight
Zygopte
ra
Calo
pte
rygid
ae Calopteryx exul Ca-Cex 45-50 34-36 27-29 0 1 0 1 1 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 V-VIII
Calopteryx haemorrhoidalis
Ca-Cha 45-48 30-43 23-37 1 0 1 0 0 1 1 0 0 1 1 0 0 0 1 0 0 0 1 0 0 IV-IX
Calopteryx virgo Ca-Cvi 45-49 31-42 24-36 1 0 0 1 0 0 1 1 0 0 1 0 0 0 1 0 0 0 1 0 0 IV-IX
Calopteryx xanthostoma Ca-Cxa 45-48 35-37 28-31 1 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 IV-IX
Coenagrio
nid
ae
Ceriagrion tenellum Ce-Cte 25-35 22-30 15-21 1 0 1 0 0 1 1 1 0 0 1 1 0 0 1 1 1 1 0 0 0 IV-IX
Coenagrion caerulescens Ce-Cca 30-33 18-27 14-21 0 1 0 1 0 1 1 0 0 1 1 0 0 0 1 0 0 1 0 0 0 IV-IX
Coenagrion hastulatum Ce-Cha 31-33 22-26 16-22 1 0 1 0 0 0 0 1 0 0 0 1 1 0 1 0 0 1 0 1 0 V-VI
Coenagrion mercuriale Ce-Cme 27-31 19-27 12-21 1 0 0 1 0 1 1 0 0 0 1 0 0 0 1 1 1 1 0 0 0 III-IX
Coenagrion puella Ce-Cpu 33-35 22-31 15-24 1 0 1 0 0 1 1 1 0 1 1 1 1 0 1 1 1 1 1 0 0 IV-IX
Coenagrion scitulum Ce-Csc 30-33 20-27 14-20 0 1 0 1 0 1 1 0 0 0 0 1 1 0 1 1 1 1 0 0 0 IV-IX
Enallagma cyathigerum Ce-Ecy 29-36 22-28 15-21 1 0 1 0 0 1 1 1 0 0 0 1 1 0 1 0 0 1 0 1 0 IV-X
Enallagma deserti Ce-Ede 32-37 24-29 19-23 0 1 0 1 0 1 0 0 0 0 0 1 1 0 1 1 1 0 0 0 0 IV-IX
Erythromma (Cercion) lindenii
Ce-Eli 30-36 24-28 19-21 1 0 1 0 0 0 0 0 0 0 0 0 1 0 1 1 1 0 0 0 0 III-X
Erythromma viridulum Ce-Evi 26-32 22-25 16-20 1 0 1 0 0 1 1 1 0 0 0 1 1 0 1 0 1 0 0 0 0 IV-IX
Ischnura elegans Ce-Iel 30-34 22-29 14-21 1 0 1 0 0 0 1 1 0 1 1 1 1 0 1 0 0 1 1 0 1 IV-IX
Ischnura fountaineae Ce-Ifo 27-34 21-25 19-24 1 0 1 0 0 1 0 0 0 0 1 1 0 1 0 0 0 0 0 0 1 III-IX
Ischnura graellsii Ce-Igr 26-31 20-25 13-19 1 0 1 0 0 1 1 0 0 1 1 1 1 0 1 0 0 1 1 0 0 III-IX
Ischnura pumilio Ce-Ipu 26-31 22-25 14-18 1 0 1 0 0 1 1 1 1 0 1 1 0 1 0 0 0 1 0 1 0 III-IX
Ischnura saharensis Ce-Isa 26-31 19-25 12-17 1 0 1 0 0 1 0 0 0 1 1 1 1 0 1 0 0 1 1 0 0 II-XI
Pseudagrion sublacteum Ce-Psu 32-39 25-33 17-23 0 1 0 1 0 1 0 0 1 1 0 1 0 1 0 0 0 0 0 0 0 V-VIII
Pyrrhosoma nymphula Ce-Pny 33-36 25-30 19-24 0 1 0 1 0 1 1 1 0 1 1 0 0 0 1 0 0 1 1 0 0 III-VIII
Lestid
ae
Lestes barbarus Le-Lba 40-45 26-35 20-27 1 0 0 1 0 1 1 1 0 0 1 1 0 1 0 0 0 1 1 0 0 III-X
Lestes dryas Le-Ldr 35-40 26-33 20-25 1 0 1 0 0 0 0 1 0 0 0 1 1 1 0 0 0 1 0 0 0 III-X
Lestes macrostigma Le-Lma 39-48 31-38 24-27 0 1 0 1 0 0 1 1 0 0 0 1 0 0 1 0 0 1 0 0 1 II-IX
Lestes sponsa Le-Lsp 35-39 25-33 17-24 1 0 1 0 0 0 1 1 0 0 0 1 1 0 1 0 0 1 0 1 0 V-X
Lestes virens (+ L. numidicus)
Le-Lnu 30-39 25-32 19-23 1 0 0 1 0 1 1 1 0 0 0 1 1 1 0 0 0 1 0 0 0 III-XI
Lestes viridis Le-Lvi 39-48 29-39 23-28 1 0 1 0 0 1 1 1 0 0 0 1 0 0 1 0 0 0 1 0 0 V-XI
Sympecma fusca Le-Sfu 34-39 25-30 18-23 1 0 1 0 0 1 1 1 0 0 0 1 1 1 1 0 1 1 0 0 0 I-XII
Zyg
opte
ra
Pla
tycn
em
idid
ae Platycnemis acutipennis P-Pac 34-37 24-28 18-19 1 0 1 0 0 0 1 0 0 1 1 0 0 0 1 0 1 0 0 0 0 V-VIII
Platycnemis latipes P-Pla 33-37 25-30 18-22 1 0 1 0 0 0 1 0 0 1 1 0 1 0 1 0 0 1 0 0 0 VI-IX
Platycnemis pennipes P-Ppe 35-37 27-31 19-23 1 0 0 1 0 0 1 1 0 1 1 1 1 0 1 1 1 1 0 0 0 V-IX
Platycnemis subdilatata P-Psu 33-36 22-28 17-21 1 0 1 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 1 0 0 IV-IX
An
iso
pte
ra
Ae
sh
nid
ae
Aeshna affinis A-Aaf 57-66 39-49 37-42 0 1 0 1 0 1 1 1 0 0 0 1 0 1 0 0 0 1 0 0 0 V-VIII
Aeshna cyanea A-Acy 67-76 51-61 43-53 1 0 1 0 0 0 1 1 0 0 0 1 0 0 0 0 0 1 1 0 0 IV-X
Aeshna (Anaciaeschna) isoceles (isosceles)
A-Ais 62-66 47-54 39-45 1 0 0 1 0 1 1 1 0 0 0 1 1 0 1 0 1 1 0 0 0 V-VIII
Aeshna juncea A-Aju 65-80 50-59 40-48 0 1 0 1 0 0 1 1 0 1 0 0 1 0 1 0 0 1 0 1 0 VI-XI
Aeshna mixta A-Ami 56-64 43-54 37-42 1 0 1 0 0 1 1 1 1 0 0 1 1 0 1 0 1 1 0 0 1 V-XII
Anax (Hemianax) ephippiger A-Aep 61-70 43-56 43-48 0 1 1 0 0 1 1 0 0 1 1 1 0 1 0 0 0 1 0 0 0 I-XII
Anax imperator A-Aim 66-84 50-61 45-52 1 0 1 0 0 1 1 1 1 1 0 1 1 1 1 0 1 1 0 0 0 II-X
Anax parthenope A-Apa 62-75 46-53 44-51 1 0 1 0 0 1 1 1 1 0 0 0 1 0 1 1 1 1 0 0 0 II-XI
Boyeria irene A-Bir 63-71 44-48 39-45 1 0 1 0 0 1 1 0 0 1 1 0 0 0 1 0 0 0 1 0 0 V-X
Brachytron pratense A-Bpr 54-63 37-46 34-37 0 1 0 1 0 0 1 1 0 0 0 1 1 0 1 1 1 1 0 0 0 III-VIII
Co
rdu
lega
str
ida
e
Cordulegaster bidentata Cg-Cbi 69-78 52-60 41-46 0 1 1 0 0 0 1 1 0 0 1 0 0 0 1 0 0 0 1 0 0 V-VIII
Cordulegaster boltonii Cg-Cbo 74-80 52-64 40-47 1 0 1 0 0 1 1 1 0 0 1 0 0 0 1 0 0 0 1 0 0 V-VIII
Cordulegaster princeps Cg-Cpr 75-86 56-65 45-49 0 1 0 1 1 1 0 0 0 0 1 0 0 0 1 0 0 0 1 0 0 V-IX
Co
rdu
liida
e Cordulia aenea Cl-Cae 47-55 30-39 29-35 1 0 0 1 0 0 1 1 0 0 0 0 1 0 1 0 0 1 1 0 0 IV-VII
Macromia splendens Cl-Msp 70-75 48-55 42-49 0 1 0 1 1 0 1 0 0 1 0 0 0 0 1 0 0 0 1 0 0 V-VIII
Oxygastra curtisii Cl-Ocu 47-54 33-39 33-36 1 0 0 1 0 1 1 0 0 1 0 0 0 0 1 0 0 0 1 0 0 V-VIII
Somatochlora metallica Cl-Sme 50-55 37-44 34-38 0 1 0 1 0 0 1 1 0 0 0 1 1 0 1 0 0 1 1 0 0 V-IX
Gom
ph
ida
e
Gomphus graslinii G-Ggr 47-50 33-38 27-30 0 1 0 1 0 0 1 0 0 1 0 0 0 0 1 0 0 0 1 0 0 V-VIII
Gomphus pulchellus G-Gpu 47-50 34-38 27-31 1 0 1 0 0 0 1 0 0 0 0 1 1 0 1 0 0 0 0 0 0 III-VIII
Gomphus simillimus G-Gsi 45-50 33-36 29-33 1 0 1 0 0 1 1 0 0 1 1 0 0 0 1 0 0 0 1 0 0 V-VII
Gomphus vulgatissimus G-Gvu 45-50 33-37 28-33 1 0 1 0 0 0 1 1 0 1 1 0 1 0 1 0 0 0 1 0 0 IV-VI
Onychogomphus costae G-Oco 43-46 30-34 22-27 0 1 0 1 1 1 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 V-VIII
Onychogomphus forcipatus G-Ofo 46-50 31-37 25-30 1 0 1 0 0 1 1 1 0 1 1 0 0 0 1 0 0 0 0 0 0 V-IX
Onychogomphus uncatus G-Oun 50-53 34-42 29-33 1 0 1 0 0 1 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 V-IX
Paragomphus genei G-Pge 37-50 30-36 21-26 0 1 1 0 0 1 1 0 0 0 0 1 1 1 0 0 0 0 0 0 0 IV-X
An
iso
pte
ra
Lib
ellu
lidae
Acisoma panorpoides Li-AAr 24-31 16-22 19-25 0 1 0 1 0 1 0 0 1 0 0 1 1 0 1 0 0 1 0 0 0 V-IX
Brachythemis leucosticta Li-Ble 25-34 16-21 20-26 1 0 1 0 0 1 1 0 1 1 0 1 1 1 1 0 0 0 0 0 0 IV-X
Crocothemis erythraea Li-Cer 36-45 18-33 23-33 1 0 1 0 0 1 1 1 1 1 1 1 1 1 1 0 0 1 0 0 1 II-XI
Diplacodes lefebvrii (lefebvri)
Li-Dle 25-34 15-25 19-29 1 0 0 1 0 1 1 0 1 0 0 1 0 1 0 0 0 1 0 0 0 IV-XI
Leucorrhinia dubia Li-Ldu 31-36 21-27 23-28 1 0 0 1 0 0 1 1 0 0 0 1 1 0 1 0 0 1 1 1 0 IV-IX
Libellula depressa Li-Lde 39-48 22-31 32-38 1 0 1 0 0 0 1 1 0 0 0 1 1 0 1 1 1 1 0 0 0 IV-IX
Libellula (Ladona) fulva Li-Lfu 42-45 25-29 32-38 1 0 0 1 0 0 1 1 0 1 0 1 1 0 1 0 0 1 0 0 0 IV-VIII
Libellula (Ladona) quadrimaculata
Li-Lqu 40-48 27-32 32-40 1 0 1 0 0 1 1 1 0 0 0 1 1 0 1 1 1 1 0 0 0 V-IX
Orthetrum brunneum Li-Obr 41-49 25-32 33-37 1 0 1 0 0 1 1 1 0 0 1 1 0 0 1 0 0 0 0 0 0 IV-IX
Orthetrum cancellatum Li-Oca 44-50 29-35 35-41 1 0 1 0 0 1 1 1 0 1 0 0 1 0 1 0 0 0 0 0 0 IV-IX
Orthetrum chrysostigma Li-Och 39-48 26-33 27-32 1 0 1 0 0 1 1 0 1 1 0 1 0 0 1 0 0 1 0 0 0 IV-X
Orthetrum coerulescens Li-Oco 36-45 23-38 28-33 1 0 1 0 0 1 1 1 1 0 1 1 0 0 1 1 0 1 0 0 0 IV-XI
Orthetrum nitidinerve Li-Oni 46-50 28-33 31-38 0 1 0 1 0 1 1 0 0 0 1 1 0 1 0 0 0 1 0 0 0 IV-XI
Orthetrum sabina Li-OAr 43-50 31-36 28-33 0 1 0 1 0 1 0 0 0 0 0 1 1 0 1 0 0 1 0 0 0 IV-X
Orthetrum trinacria Li-Otr 51-67 38-44 34-38 0 1 0 1 0 1 1 0 1 0 0 1 1 0 1 0 0 1 0 0 0 III-X
Pantala flavescens Li-Pfl 45-55 26-37 38-42 0 1 1 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 VI-IX
Selysiothemis nigra Li-Sni 30-38 21-26 24-27 0 1 0 1 0 1 1 0 1 0 0 1 0 1 0 0 0 1 0 0 0 V-IX
Sympetrum flaveolum Li-Sfl 32-37 19-27 23-32 1 0 0 1 0 0 1 1 0 0 0 1 0 1 0 1 0 1 0 0 0 V-X
Sympetrum fonscolombii (fonscolombei)
Li-Sfo 33-40 22-29 26-31 1 0 1 0 0 1 1 1 1 0 0 1 0 1 0 0 0 0 0 0 0 I-XII
Sympetrum meridionale Li-Sme 35-40 22-28 25-30 1 0 1 0 0 1 1 1 0 0 0 1 0 0 1 0 0 1 0 0 0 V-X
Sympetrum pedemontanum Li-Spe 28-35 18-24 21-28 0 1 0 1 0 0 1 1 0 0 1 1 0 1 0 0 1 1 0 0 0 VII-X
Sympetrum sanguineum Li-Ssa 34-39 20-26 23-31 1 0 0 1 0 1 1 1 0 0 0 1 1 1 1 1 1 1 0 0 0 V-VIII
Sympetrum sinaiticum Li-Ssi 34-37 21-26 24-29 1 0 1 0 0 1 1 0 1 0 1 1 1 1 0 0 0 1 0 0 0 VI-III
Sympetrum striolatum Li-Sst 35-44 20-30 24-30 1 0 1 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 0 1 I-XII
Sympetrum vulgatum (decoloratum)
Li-Svu 35-40 23-28 24-29 1 0 1 0 0 0 1 1 0 0 0 1 1 0 1 1 0 1 0 0 0 VI-XI
Trithemis annulata Li-Tan 32-38 17-29 20-35 1 0 1 0 0 1 1 0 1 0 0 1 1 1 1 0 0 1 0 0 0 III-X
Trithemis arteriosa Li-Tar 32-38 20-26 23-30 1 0 1 0 0 1 0 0 1 0 0 1 1 1 1 0 0 1 0 0 0 V-X
Trithemis kirbyi Li-Tki 30-34 19-23 23-29 0 1 0 1 0 1 1 0 1 0 1 1 0 1 0 0 0 0 0 0 0 V-XI
Zygonyx torridus Li-Zto 50-60 35-43 45-50 0 1 1 0 0 0 1 0 1 1 0 0 0 0 1 0 0 0 0 0 0 IV-VIII
Table S1. Quantification of traits categories using a fuzzy code approach.
Su
bo
rde
r
Fa
mily
Sp
ecie
s
Ibe
rian
Pen
insu
la
Data
set
Mo
rocco
D
ata
set AUC
Occu
rren
ce
GLM GAM BRT
Anisoptera Aeshnidae Aeshna affinis 1 1 0.87 0.93 0.89 78
Anisoptera Aeshnidae Aeshna cyanea 1 0 0.91 0.94 0.99 716
Anisoptera Aeshnidae Aeshna juncea 1 0 - - - 266
Anisoptera Aeshnidae Aeshna mixta 1 1 0.90 0.93 0.98 708
Anisoptera Aeshnidae Anaciaeschna isosceles 1 1 0.90 - 0.99 283
Anisoptera Aeshnidae Anax ephippiger 1 1 0.89 0.94 0.96 248
Anisoptera Aeshnidae Anax imperator 1 1 0.91 0.92 - 2862
Anisoptera Aeshnidae Anax parthenope 1 1 0.91 0.94 - 1019
Anisoptera Aeshnidae Boyeria irene 1 1 0.81 0.90 - 1806
Anisoptera Aeshnidae Brachytron pratense 1 0 - - - 2
Anisoptera Cordulegastridae Cordulegaster bidentata 1 0 - - - 30
Anisoptera Cordulegastridae Cordulegaster boltonii 1 1 0.82 0.86 0.97 1173
Anisoptera Cordulegastridae Cordulegaster princeps 0 1 - - - 48
Anisoptera Corduliidae Cordulia aenea 1 0 - - - 23
Anisoptera Corduliidae Macromia splendens 1 0 0.98 0.99 - 137
Anisoptera Corduliidae Oxygastra curtisii 1 1 0.87 0.96 1.00 706
Anisoptera Corduliidae Somatochlora metallica 1 0 - - - 28
Anisoptera Gomphidae Gomphus graslini 1 0 0.97 0.98 - 162
Anisoptera Gomphidae Gomphus lucasii 0 1 - - - 0
Anisoptera Gomphidae Gomphus pulchellus 1 0 0.91 0.93 0.98 242
Anisoptera Gomphidae Gomphus simillimus 1 1 0.90 0.94 0.96 214
Anisoptera Gomphidae Gomphus vulgatissimus 1 0 - - - 2
Anisoptera Gomphidae Onychogomphus boudoti 0 0 - - - 0
Anisoptera Gomphidae Onychogomphus costae 1 1 0.97 0.97 0.99 227
Anisoptera Gomphidae Onychogomphus forcipatus 1 1 0.82 0.91 0.98 1344
Anisoptera Gomphidae Onychogomphus uncatus 1 1 0.80 0.93 - 1450
Anisoptera Gomphidae Paragomphus genei 1 1 - - 1.00 60
Anisoptera Libellulidae Acisoma panorpoides 0 0 - - - 0
Anisoptera Libellulidae Brachythemis impartita 1 1 0.97 0.99 0.97 253
Anisoptera Libellulidae Crocothemis erythraea 1 1 0.91 0.93 0.98 2810
Anisoptera Libellulidae Diplacodes lefebvrii 1 1 - - 1.00 187
Anisoptera Libellulidae Leucorrhinia dubia 1 0 - - - 83
Anisoptera Libellulidae Libellula depressa 1 0 0.90 0.93 - 583
Anisoptera Libellulidae Libellula fulva 1 0 - - - 264
Anisoptera Libellulidae Libellula quadrimaculata 1 1 - 0.94 0.98 299
Anisoptera Libellulidae Orthetrum brunneum 1 1 0.89 0.93 0.97 978
Anisoptera Libellulidae Orthetrum cancellatum 1 1 0.93 0.95 0.98 1232
Anisoptera Libellulidae Orthetrum chrysostigma 1 1 0.93 0.95 0.99 1032
Anisoptera Libellulidae Orthetrum coerulescens 1 1 0.90 0.93 0.97 1319
Anisoptera Libellulidae Orthetrum nitidinerve 1 1 0.95 0.98 - 117
Anisoptera Libellulidae Orthetrum ransonnetii 0 1 - - - 4
Anisoptera Libellulidae Orthetrum sabina 0 0 - - - 0
Anisoptera Libellulidae Orthetrum trinacria 1 1 - - 1.00 296
Anisoptera Libellulidae Pantala flavescens 0 1 - - - 1
Anisoptera Libellulidae Selysiothemis nigra 1 1 0.97 0.98 0.98 218
Anisoptera Libellulidae Sympetrum flaveolum 1 0 - - 1.00 154
Anisoptera Libellulidae Sympetrum fonscolombii 1 1 0.88 0.91 0.97 3267
Anisoptera Libellulidae Sympetrum meridionale 1 1 0.87 0.93 0.97 141
Anisoptera Libellulidae Sympetrum pedemontanum 1 0 - - - 43
Anisoptera Libellulidae Sympetrum sanguineum 1 1 - - 1.00 132
Anisoptera Libellulidae Sympetrum sinaiticum 1 1 0.97 0.99 0.98 175
Anisoptera Libellulidae Sympetrum striolatum 1 1 0.90 0.94 0.98 1573
Anisoptera Libellulidae Sympetrum vulgatum 1 0 - - - 53
Su
bo
rde
r
Fa
mily
Sp
ecie
s
Ibe
rian
Pen
insu
la
Data
set
Mo
rocco
D
ata
set AUC
Occu
rren
ce
GLM GAM BRT
Anisoptera Libellulidae Trithemis annulata 1 1 0.91 0.93 0.98 1873
Anisoptera Libellulidae Trithemis arteriosa 0 1 - - - 22
Anisoptera Libellulidae Trithemis kirbyi 1 1 0.96 0.97 0.98 900
Anisoptera Libellulidae Urothemis edwardsii 0 0 - - - 0
Anisoptera Libellulidae Zygonyx torridus 1 1 - 1.00 - 172
Zygoptera Calopterygidae Calopteryx exul 0 1 - - - 8
Zygoptera Calopterygidae Calopteryx haemorrhoidalis 1 1 0.89 0.91 0.98 2263
Zygoptera Calopterygidae Calopteryx virgo 1 1 0.87 0.90 0.99 1367
Zygoptera Calopterygidae Calopteryx xanthostoma 1 0 0.90 0.92 0.98 1108
Zygoptera Coenagrionidae Ceriagrion tenellum 1 1 0.92 0.95 0.98 511
Zygoptera Coenagrionidae Coenagrion caerulescens 1 1 0.90 0.90 0.97 409
Zygoptera Coenagrionidae Coenagrion hastulatum 1 0 - - - 24
Zygoptera Coenagrionidae Coenagrion mercuriale 1 1 0.89 0.91 - 390
Zygoptera Coenagrionidae Coenagrion puella 1 1 - - 0.99 695
Zygoptera Coenagrionidae Coenagrion scitulum 1 1 0.89 0.94 0.97 135
Zygoptera Coenagrionidae Enallagma cyathigerum 1 1 0.82 0.87 0.99 745
Zygoptera Coenagrionidae Enallagma deserti 0 1 - - - 26
Zygoptera Coenagrionidae Erythromma lindenii 1 1 0.91 0.94 0.99 1703
Zygoptera Coenagrionidae Erythromma viridulum 1 1 0.89 0.93 0.98 288
Zygoptera Coenagrionidae Ischnura elegans 1 0 0.99 - 1.00 1093
Zygoptera Coenagrionidae Ischnura fountaineae 0 1 - - - 6
Zygoptera Coenagrionidae Ischnura graellsii 1 1 0.90 0.94 0.97 2548
Zygoptera Coenagrionidae Ischnura pumilio 1 0 0.92 0.95 0.98 359
Zygoptera Coenagrionidae Ischnura saharensis 0 1 0.77 - 0.98 76
Zygoptera Coenagrionidae Pseudagrion sublacteum 0 1 - - - 16
Zygoptera Coenagrionidae Pyrrhosoma nymphula 1 1 0.89 0.92 0.99 1037
Zygoptera Lestidae Lestes barbarus 1 1 0.86 0.94 0.98 300
Zygoptera Lestidae Lestes dryas 1 1 0.86 0.92 0.99 174
Zygoptera Lestidae Lestes macrostigma 1 0 - - - 68
Zygoptera Lestidae Lestes numidicus 0 0 - - - 0
Zygoptera Lestidae Lestes sponsa 1 0 0.89 - - 114
Zygoptera Lestidae Lestes virens 1 1 0.86 0.92 0.97 366
Zygoptera Lestidae Chalcolestes viridis 1 1 0.92 0.95 0.99 1817
Zygoptera Lestidae Sympecma fusca 1 1 0.90 0.93 0.99 616
Zygoptera Platycnemididae Platycnemis acutipennis 1 0 0.93 0.96 0.99 540
Zygoptera Platycnemididae Platycnemis latipes 1 0 0.93 0.96 0.98 2080
Zygoptera Platycnemididae Platycnemis pennipes 1 0 - - - 4
Zygoptera Platycnemididae Platycnemis subdilatata 0 1 0.95 0.96 1.00 90
Table S2. AUC values of each species and model selected. Dataset origin is also indicated.
Species MPI RCP 2.6 MPI RCP 8.5 HadGEM2 RCP 2.6 HadGEM2 RCP 8.5 Aeshna affinis 2/2 1*/1 2/1 1/1 Aeshna cyanea 1/3 4/4 4/3 4/4 Aeshna mixta 2/3 2/1 2/3 1/3 Anaciaeschna isosceles 2/3 2/1 3/1 3/2 Anax ephippiger 2/3 2/2 2/2 2/2 Anax imperator 2/2 2/1 2/3 1/3 Anax parthenope 2/3 2/4 1/4 3/1 Boyeria irene 1/2 1/1 4/2 4/4 Brachythemis impartita 2/3 2/1 2/3 2/4 Calopteryx haemorrhoidalis 2/2 1/1 1/1 4/4* Calopteryx virgo 1*/2 1/4 1/1 4/5 Calopteryx xanthostoma 1*/2 1/4 1/1 4/4 Ceriagrion tenellum 2/3 1*/4 1/4* 4/4 Coenagrion caerulescens 3/1 4/1 4/1 4/4 Coenagrion mercuriale 1/3 1/4 1/4 4/4 Coenagrion puella 4/4 5/5 4/4 4/5 Coenagrion scitulum 1/3 1/4 1/4* 4/4* Cordulegaster boltonii 3/2 1/1 4/1 4/4 Crocothemis erythraea 2/3 2/1 2/1 3/1 Diplacodes lefebvrii 2/3 2/1 2/3 2/1 Enallagma cyathigerum 3/2 1/4 4/1 4/4 Erythromma lindenii 2/3 2/1 2/1 3/1 Erythromma viridulum 2/3 2/1 1/1 4/4 Gomphus graslini 1/4* 4/4 1/4 4/4 Gomphus pulchellus 2/1* 2/2 2/1 1/1 Gomphus simillimus 1/3 1/4* 1/4* 1/4 Ischnura elegans 2/3 1*/4 1/4 5/1 Ischnura graellsii 2/1* 2/1 1/1 1/4* Ischnura pumilio 1/1* 4*/4 4*/4* 4/4 Ischnura saharensis 2/3 2/3 2/4* 2/3 Lestes barbarus 1/3 1/4* 1/4* 4/4 Lestes dryas 3/2 4*/4 4/1 4/4 Lestes sponsa 4/4* 4/5 4/2 4/5 Lestes virens 1/2 1/4 1/1 4/4 Chalcolestes viridis 2/1* 2/1 1/1 3/1 Libellula depressa 2/2 1/4 1/1 4/4 Libellula quadrimaculata 1/3 1/4 1/1 1/4 Macromia splendens 1/3 1/1 1/4 4/4 Onychogomphus costae 2/3 2/1 2/3 2/3 Onychogomphus forcipatus 1/2 1/4 1/1 4/4 Onychogomphus uncatus 4/2* 4/4* 4/1 4/4 Orthetrum brunneum 1/2 1/4 1/1 4/4 Orthetrum cancellatum 2/2 2/1 2/1 1/4 Orthetrum chrysostigma 2/2* 2/1 2/3 2/3 Orthetrum coerulescens 2/1 2/1 1/4* 4/4* Orthetrum nitidinerve 1/3 1/1 2/3 1/4 Orthetrum trinacria 2/3 2/4 2/4 2/4 Oxygastra curtisii 2/4* 2/4 1/4 1/4* Paragomphus genei 2/3 3/1 2/3 2/4 Platycnemis acutipennis 2/2* 2/2 2/1 2/2* Platycnemis latipes 2/1* 2/1 1/1 1/1 Platycnemis subdilatata 3/3 3/4 2/3 2/4 Pyrrhosoma nymphula 2/2 1/4* 4*/1 4/4 Selysiothemis nigra 2/2 2/2 2/1 3/1 Sympecma fusca 2/2* 2*/1 1*/4* 1/4* Sympetrum flaveolum 1/1 5/5 5/3 5/5 Sympetrum fonscolombii 2/1* 1/1 1/3 4/1 Sympetrum meridionale 1/4 2/4 2/4* 1/4 Sympetrum sanguineum 1/1 4/4 4/3 4/5 Sympetrum sinaiticum 2/2* 2/2 2/3 1/3 Sympetrum striolatum 2/2* 2/1 1/1 1/3 Trithemis annulata 2/3 2/2 2/3 2/3 Trithemis kirbyi 2/3 2/4 2/3 2/1 Zygonyx torridus 2/3 2/2 2/3 2/2*
Table S3. Species potential distribution categories for each model (MPI and HadGEM2) and
scenario (RCP 2.6 and RCP 8.5). Left number in 2050, right number in 2080. Categories: 1:
potential area is latitudinally or longitudinally displaced, 2: potential area is being expanded, 3:
no significant differences between current and future potential areas distribution, 4: habitat lost,
5: significant reduction of potential area. 1*: there is a displacement of the potential area but
some regions maintain the same environmental conditions, 2*: expansion but with area lost, 4*:
area reduction but with new habitat gain.
top related