Science of the Total Environment 631-632 (2018) 429-438
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Increased anthropogenic disturbance and aridity reduce phylogenetic
and functional diversity of ant communities in Caatinga dry forest
Xavier Arnan a,b '*, Gabriela B. Arcoverde c , Marcio R. Pie d , Jose D. Ribeiro-Neto a,e , Inara R. Leal 1
a Programa de Pos-Cradua(ao em Biologia Vegetal, Universidade Federal de Pernambuco, Av. Prof. Moraes Rego s/no, Recife, PE 50670-901, Brazil
b CREAF, Cerdanyola del Valles, ES-08193, Catalunya, Spain
c Research School of Environment and Livelihoods, Charles Darwin University, Darwin, NT 0909, Australia
d Departamento de Zoologia, Universidade Federal do Parana, Caixa Postal 19020, Curitiba, PR 81531-980, Brazil
e Departamento de Fitotecnia e Ciencias Ambientais, Centro de Ciencias Agrarias, Universidade Federal da Paraiba, Rodovia PB-079,58397-000 Areia, PB, Brazil
f Departamento de Botanica, Universidade Federal de Pernambuco, Av. Prof. Moraes Rego s/no, Recife, PE 50670-901, Brazil
Ant biodiversity patterns under global
change drivers are assessed in Caatinga.
Functional and phylogenetic diversity
decrease with aridity and human distur¬
Human disturbance and aridity interact
in complex ways to endanger biodiver¬
Aridity can intensify the negative effects
of disturbance on biodiversity.
Concerns about the future of biodiver¬
sity in neotropical semi-arid regions.
Miscellaneous resource use Miscellaneous resource use
Miscellaneous resource use
Received 16 November 2017
Received in revised form 3 March 2018
Accepted 4 March 2018
Available online xxxx
Editor: Yolanda Pico
Anthropogenic disturbance and climate change are major threats to biodiversity. The Brazilian Caatinga is the
world's largest and most diverse type of seasonally dry tropical forest. It is also one of the most threatened, but re¬
mains poorly studied. Here, we analyzed the individual and combined effects of anthropogenic disturbance (three
types: livestock grazing, wood extraction, and miscellaneous use of forest resources) and increasing aridity on tax¬
onomic, phylogenetic and functional ant diversity in the Caatinga. We found no aridity and disturbance effects on
taxonomic diversity. In spite of this, functional diversity, and to a lesser extent phylogenetic diversity, decreased
with increased levels of disturbance and aridity. These effects depended on disturbance type: livestock grazing
and miscellaneous resource use, but not wood extraction, deterministically filtered both components of diversity.
Interestingly, disturbance and aridity interacted to shape biodiversity responses. While aridity sometimes intensified
the negative effects of disturbance, the greatest declines in biodiversity were in the wettest areas. Our results imply
that anthropogenic disturbance and aridity interact in complex ways to endanger biodiversity in seasonally dry trop¬
ical forests. Given global climate change, neotropical semi-arid areas are habitats of concern, and our findings sug¬
gest Caatinga conservation policies must prioritize protection of the wettest areas, where biodiversity loss stands
to be the greatest. Given the major ecological relevance of ants, declines in both ant phylogenetic and functional di¬
versity might have downstream effects on ecosystem processes, insect populations, and plant populations.
© 2018 Elsevier B.V. All rights reserved.
* Corresponding author at: CREAF, Campus UAB, 08193 Cerdanyola del Valles, Spain.
E-mail address: email@example.com (X. Arnan).
https://d 0 i. 0 rg/l 0.1016/j.scitotenv.2018.03.037
0048-9697/© 2018 Elsevier B.V. All rights reserved.
X. Arrian et al. / Science of the Total Environment 631-632 (2018) 429-438
Anthropogenic disturbance and global climate change are key
threats to biodiversity (Bellard et al., 2012) because they have signifi¬
cant impacts on biological populations and community organization.
This is especially true in seasonally diy tropical forests (SDTFs), which
are experiencing increased rates of both acute and chronic disturbance.
Major acute disturbances include habitat loss and fragmentation (Miles
et al., 2006). Major chronic disturbances (hereafter referred to as CADs—
chronic anthropogenic disturbances [sensu Singh, 1998]) include live¬
stock grazing, wood extraction, and the exploitation of miscellaneous
forest resources. These activities all result in the removal of significant
amounts of biomass. In STDFs, few efforts have been made to under¬
stand the impacts of CADs on biodiversity, but negative impacts have
been described in communities of both plants (Sagar et al., 2003;
Ribeiro et al., 2015, 2016; Rito et al., 2017) and animals (Ribeiro-Neto
et al., 2016; Oliveira et al„ 2017). Climate change also threatens SDTFs
(Dirzo et al., 2011) and might even exacerbate the effects of anthropo¬
genic disturbances (Hirota et al., 2011; Ponce-Reyes et al., 2013; Gibb
et al., 2015a; Frishkoff et al., 2016). Hot and arid environments are likely
at the greatest risk (Anderson-Teixeira et al., 2013; Gibb et al., 2015a).
Studies of biological communities have generally focused on pat¬
terns of species diversity, which are often quantified using species rich¬
ness and/or composition (Pavoine and Bonsall, 2011). However, new
diversity metrics that incorporate information about phylogenetic di¬
versity (PD) and functional diversity (FD) can reveal more information
about community organization in different contexts (Faith, 1992;
Webb et al., 2002; Petchey and Gaston, 2006; Swenson, 2014), including
those marked by anthropogenic disturbance and climate change
(Mouillot et al., 2013). While phylogenetic diversity reflects the accu¬
mulated evolutionary history of a community (Webb et al., 2002), func¬
tional diversity reflects the diversity of morphological, physiological,
and ecological traits found therein (Petchey and Gaston, 2006). It is gen¬
erally accepted that PD and FD can increase with taxonomic diversity by
chance, since the presence of more species should mean that more lin¬
eages and functions are represented. However, these relationships are
not always linear. Two communities with equal taxonomic diversity
might greatly differ in PD and FD (Petchey and Gaston, 2006; Safi
et al., 2011 ; Arnan et al„ 2015,2017) due to different levels of functional
redundancy, different evolutionary histories, and/or contrasting envi¬
ronmental conditions. However, a strong correlation between FD and
PD would be expected if the functional traits that allow species to per¬
sist in the environment are evolutionarily conserved, that is to say,
they display phylogenetic signals (Webb et al., 2002; Cavender-Bares
et al., 2009). Remarkably, very little is known about how PD and the
functional composition of animal communities change in response to
disturbance and environmental conditions, especially in SDTFs.
When examining biodiversity patterns, ants are a good study group
—they are among the most diverse and abundant terrestrial organisms
on earth and they are highly sensitive to environmental change
(Holldobler and Wilson, 1990). Moreover, ants play an important role
in many basic ecosystem services (Bihn et al., 2010, Del Toro et al.,
2012). In particular, ants are crucial contributors to soil cycling and aer¬
ation, organic matter decomposition, seed dispersal, and plant protec¬
tion (Del Toro et al., 2012). Ants are extremely phylogenetically
diverse, especially in the tropics (Holldobler and Wilson, 1990), and
ant morphological traits have frequently been used to infer ecosystem
services (Weiser and Kaspari, 2006; Gibb et al., 2015b; Parr et al.,
2017; Salas-Lopez, 2017).
In this study, we analyzed the effects of CADs and climate change,
notably increasing aridity, on the phylogenetic and functional diversity
of ants in the Brazilian Caatinga, the largest and most diverse of the
world's SDTFs (Leal et al., 2005). The Caatinga is the third most-
threatened Brazilian ecosystem and yet is the most poorly studied and
understood (Overbeck et al., 2015; Oliveira and Bernard, 2017). The
27 million people living in the Caatinga are highly dependent on its
natural resources for their livelihoods, which has resulted in its slow
degradation over time (Leal et al., 2005; Ribeiro et al., 2015). Moreover,
the Caatinga is one of the six ecosystems with the greatest intrinsic vul¬
nerability to climate variability (Seddon et al., 2016); climate models
consistently predict a reduction in rainfall levels (22%) and an increase
in temperature (3-6 °C) (Magrin et al., 2014).
In this context, the Caatinga is a good model system with which to
investigate the effects of anthropogenic disturbance and climate change
(i.e., increased aridity) on the biological communities of dry forests. It
can also be used to characterize changes in community organization
arising from transformations in SDTFs. Previous studies in the Caatinga
found no or small differences in ant species richness along CAD gradi¬
ents; however, large changes in species composition were observed
(Ribeiro-Neto et al., 2016; Oliveira et al., 2017). This finding suggests
shifts in phylogenetic and functional diversity along CAD gradients.
We therefore first corroborated that species diversity is not modulated
by CAD and aridity gradients, and hypothesize the following: (a) PD
and FD will decrease as anthropogenic disturbance and aridity increase;
(b) PD and FD will decrease even more sharply in areas that are both
highly disturbed and arid; and (c) PD and FD patterns along gradients
of disturbance and aridity will be driven by deterministic processes
rather than by stochasticity.
2. Materials and methods
2.1. Study area
This study was conducted in Catimbau National Park (8°24'00" and
8°36'35" S; 37°0'30" and 37°1'40" W, state of Pernambuco, Brazil),
which cover an area of 607 km 2 of Caatinga vegetation (Sociedade
Nordestina de Ecologia, 2002). The climate is hot. Mean annual temper¬
ature is 25 °C, and mean annual rainfall ranges between 1100 mm in the
southeast to 480 mm in the northwest (Rito et al., 2017). However, the
park experiences substantial interannual and spatial variability in con¬
ditions (Sociedade Nordestina de Ecologia, 2002). Most of the park has
quartzolic sandy soils (70%), but planosols (15%) and lithosols (15%)
are also present (Sociedade Nordestina de Ecologia, 2002). The domi¬
nant families of woody plants are Fabaceae, Euphorbiaceae, and
Boraginaceae; on the surface of the forest floor, Cactaceae,
Bromeliaceae, Malvaceae, Asteraceae, and Fabaceae dominate (Rito
et al., 2017).
The park was established in 2002 (Sociedade Nordestina de
Ecologia, 2002), but its original human inhabitants remain; they
continue to hunt, graze livestock, extract timber, collect firewood,
and harvest other plant resources (Rito et al., 2017). Their historical
presence has resulted in an extensive mosaic of differential land use
and anthropogenic pressure on biota. This fact means Catimbau rep¬
resents an excellent opportunity for examining how anthropogenic
disturbance (e.g., farming, livestock grazing, extraction of timber,
firewood gathering, and hunting) affects the biota of the Caatinga.
Also, the considerable variation in precipitation within the park
(100%) can help reveal whether high levels of aridity can intensify
the negative effects of human disturbance.
2.2. Characterization of disturbance and aridity gradients
We sampled 20 0.1-ha plots (20 x 50 m; separated by at least
2 km) located within areas dominated by old-growth vegetation;
the plots occurred along an aridity gradient and experienced varying
degrees of CAD (Fig. 1). Thanks to aerial photographs and
preliminary interviews with locals, we could confirm that the plots
had not experienced any acute disturbances over the past 80 years.
All plots were located in areas with the same soil type (sandy soil),
slope (flat terrain), and vegetation type (dry forest with short-
stature trees) (Rito et al., 2017).
X.Aman et al. / Science of the Total Environment 631-632 (2018) 429-438
We characterized disturbance intensity by calculating three different
indices that corresponded to the main CADs affecting the Caatinga in
general and Catimbau in particular: (1) livestock grazing (LG)—con¬
sumption of vegetation, trampling, and other physical damage caused
by cattle and goats; (2) wood extraction (WE)—the extraction of dead
and live wood for fuel, fence construction, and artisanal production;
and (3) miscellaneous resource use (MU)—use of non-wood resources
by humans (e.g., food and medicinal plants, hunting). Index values
were calculated using the following formula: / = Sfeji 3 ' 1 x
100, where / is disturbance intensity; y,- is the observed value for a given
disturbance metric in plot i; y min is the minimum observed value for the
disturbance metric across all plots; y max is the maximum observed value
for the disturbance metric across all plots; and n is the number of indi¬
vidual disturbance metrics incorporated in the index. This formula thus
standardizes the metrics (sometimes of different units) to take on a
value between 0 and 1, allowing them to be combined in the same
index. Index values ranged from 0 to 100 (from no disturbance to
maximum-intensity disturbance). Both the LG and WE indices quanti¬
fied disturbances that were directly measured in the field. For the LG
index, we estimated grazing levels by measuring the length of goat trails
and the frequency of cattle and goat dung (see Appendix SI for details).
Then, we combined the two estimates of goat grazing (trail length and
dung frequency) by means of principal component analysis (PCA).
Both measures were highly positively correlated (r > 0.90) with the
first PCA axis, which explained 88% of variance. We therefore used its
coordinates to obtain a single measure of goat grazing. The LG index
was then calculated by inputting measures of goat grazing and cattle
dung frequency into the formula above. For the WE index, we estimated
the extraction of live wood and the collection of firewood (Appendix
SI) and plugged them directly into the formula above. Finally, the MU
index was determined using three indirect variables that are proxies
for local anthropogenic pressure and habitat accessibility. More specifi¬
cally, we estimated two relevant geographic distances—plot proximity
to the nearest house and plot proximity to the nearest road (using sat¬
ellite imagery and ArcGis f 0.1 software). We also used a socioecological
variable—the number of people living in the area that influence the plot
(Appendix SI). Then, the values of these metrics were inputted into the
formula above to obtain the MU index. The three disturbance indices
displayed a wide range of values (min-max for LG: 0-60, WE: 0-100,
and MU: 5-63; Fig. 1 and Appendix S2) and were not highly correlated
(LG vs. WE: r = 0.05, LG vs. MU: r = 0.61, and MU vs. WE: r = -0.10).
This result underscores that the indices are quite independent and mea¬
sure different forms of anthropogenic disturbance.
Field-based research into temporal evolutionary change typically re¬
quires long-term data that are unavailable for most systems.
Fig. 1 . Location of the study region (in dark gray; within Brazil) (A), Catimbau National Park (white box; within Pernambuco) (B), and the study plots (bar graphs; within Catimbau) (C).
The bar graphs depict the intensity of each anthropogenic disturbance (wood extraction, WE; livestock grazing, LG; miscellaneous resource use, MU) on each plot. The color scale depicts
aridity (i.e„ climatic water deficit).
X. A man et al. / Science of the Total Environment 631-632 (2018) 429-438
Furthermore, carrying out climate change studies in the laboratory is
difficult, especially if the aim is to address issues at the community
level. At present, the only effective means of exploring changes in biodi¬
versity due to rapid climatic shifts is the space-for-time approach,
where space acts as a substitute for time (Blois et al„ 2013). However,
historical and evolutionary processes might not act similarly along tem¬
poral versus spatial gradients. Consequently, inferences based on cer¬
tain climatic conditions can, but do not always, reflect the responses
expected over the timeline of future climate change (Bellard et al.,
2012). Here, we explored the potential effects of declining precipitation
and increasing temperature in the Caatinga by analyzing changes in
aridity along a spatial gradient. Aridity was estimated using mean an¬
nual climatic water deficit, which is the difference between potential
evapotranspiration (PET) and actual evapotranspiration (AET; based
on biologically usable energy and water) (Lutz et al., 2010). Climatic
water deficit was calculated using 30-arc-second (1-km) resolution
maps of long-term mean annual PET and AET (CGIAR-CSl's Global Arid¬
ity and PET Database and Global High-Resolution Soil-Water Balance
Database; www.csi.cgiar.org/(). These maps were generated
using temperature and precipitation data from the WorldClim global
climate data repository (www.worldclim.org). For each plot, the differ¬
ence between annual PET and AET was calculated to obtain a climatic
water deficit value. All calculations were performed using ArcGIS 10.1
software. Climatic water deficit values ranged from 658 mm (minimum
aridity) to 1086 mm (maximum aridity), and was not correlated to any
of the disturbance indices (r= —0.14, —0.18 and —0.06 for LG, WE and
2.3. Sampling ant communities
Each plot contained 20 pitfall traps (4x5 grid; separated by 5 m).
The traps were 4.5-cm diameter plastic containers partially filled with
a mixture of alcohol, ethylene glycol, and soap. Traps were left open
for a single 48-h period in March 2015, at the beginning of the rainy sea¬
son. All the ants collected were sorted to morphospecies following
Baccaro et al. (2015). They were identified when possible; unidentifi¬
able species were assigned a code. Identifications were verified by R.
Feitosa (Laboratorio de Sistematica e Biologia de Formigas,
Universidade Federal do Parana). Vouchers of all species are available
at the Universidade Federal Pernambuco in Recife and the Universidade
Federal do Parana in Curitiba.
2.4. Characterization of ant phylogenetic relationships
At present, there is no complete, species-level ant phylogeny. We
therefore used an approach that incorporated as much information as
possible given our current understanding of ant relationships while
simultaneously accounting for existing phylogenetic uncertainty. We
began by using a backbone tree derived from a time-calibrated, genus-
level phylogeny (Moreau and Bell, 2013) ; however, the phylogenetic re¬
lationships within Myrmicinae were taken from Ward et al. (2015). This
phylogeny was then pruned to keep a single species per genus and thus
generate a genus-level phylogeny. We subsequently used the list of spe¬
cies in our dataset (Appendix S3) to simulate 1000 species-level phylog-
enies. Species relationships within genera were obtained from a Yule
(pure-birth) process using the genus.to.species.ti'ee function in the
phytools package (Revell, 2012) in R (R Development Core Team,
2016). However, given that one genus ( Mycetophylax ) was missing,
we randomly added this lineage to our genus-level tree as a sister
genus of Kalathomyrmex (Klingenberg and Brandao, 2009) prior to the
addition of species in each iteration. The entire process was repeated
1000 times to account for phylogenetic uncertainty in later analyses.
2.5. Characterization of ant functional traits
We quantified functional diversity using a suite of morphological
traits that reflect body size, foraging capacity, foraging period, and re¬
source acquisition mode (Bihn et al., 2010; Parr et al., 201 7). These traits
serve as proxies for the impact a species might have on ecosystem pro¬
cesses related to resource use. By focusing on morphological traits, we
obviously ignored some aspects of the ants' ecology; however, direct
links between morphological traits and functional roles have been ob¬
served in ants (Table 1 ). For each ant species, we determined body
size (Weber's length), relative eye length, relative scape length, relative
mandible length, relative clypeus length, and relative leg length
(Table 1). We standardized all trait measurements (except Weber’s
length) by dividing each by the Weber's length to limit correlations
with body size. The trait values were then log-transformed to achieve
normality. We measured approximately six randomly selected workers,
and the mean measurements were used as the species-specific values
for monomorphic and polymorphic species. In species with distinct
minor and major worker castes, only minor workers were employed.
In total, 503 workers representing 69 ground-foraging ant species
were measured (number of individuals measured per species: mean
± SE = 7.4 ± 0.16, median = 6.4, min = 1 and max = 70).
2.6. Estimating taxonomic, phylogenetic and functional diversity
For each plot, taxonomic diversity was characterized as species rich¬
ness (S, the number of ant species in each plot) and species diversity
(Shannon diversity index, H, which accounts for both species richness
We measured five complementary metrics of phylogenetic diversity
and functional diversity (Swenson, 2014). Phylogenetic diversity was
Morphometric traits of workers used to characterize the functional diversity of Caatinga ant communities (Catimbau National Park, Pernambuco state, NE Brazil).
Trait Functional significance
Strongly correlated with many physiological, ecological, and life-history
traits, including resource use
Likely correlated with main foraging period (day vs. night)
Possibly correlated with ability to receive chemosensory information. Ants
with long scapes may be more sensitive to pheromone trails.
Possible indicator of predatory lifestyle and thus types of resources consumed
Correlated with sucking ability and liquid-feeding behavior
Possibly correlated with resource acquisition mode and foraging efficiency, as
well as with the ability to cope with the foraging surface temperature
Kaspari and Weiser,
1999, Bihn et al.,
Bihn et al., 2010
Weiser and Kaspari,
Weiser and Kaspari,
Davidson et al., 2004
Kaspari and Weiser,
1999, Bihn et al.,
Maximum longitudinal length from the most anterior
part of the clypeus to the occipital margin, in full face
Ratio of eye length to mesosoma length
Ratio of scape length to mesosoma length
Ratio of mandible length to mesosoma length
Ratio of clypeus length to mesosoma length
Ratio of leg length (combined length of femur and tibia)
to mesosoma length
X.Aman et al. / Science of the Total Environment 631-632 (2018) 429-438
estimated using the following indices (Table 2): (a) Faith's phylogenetic
diversity (Faith's PD); (b) mean pairwise distance (MPD); (c) mean
nearest-taxon distance (MNTD); (d) the net relatedness index (NRI);
and (e) the nearest taxon index (NTI). Faith's PD is widely used in con¬
servation research (Forest et al., 2007; Morion et al., 2011) and, here,
was the total branch length (divergence time) of the phylogenetic tree
linking all the species represented in the community (Faith, 1992).
MPD was the mean distance (in millions of years) between two ran¬
domly selected individuals within a specific plot (considering conspe-
cifics), while MNTD was the mean distance separating each individual
in the community from its closest heterospecific relative (Webb et al.,
2002). Thus, when MNTD is more strongly correlated with environmen¬
tal gradients than is MPD, it indicates that the environment has a stron¬
ger effect on terminal than basal community phylogenetic composition.
Since both metrics might depend on species richness, we also measured
the standardized effect size (SES) for MPD and MNTD (i.e., NRI and NTI,
respectively) by comparing observed phylogenetic relatedness to ex¬
pected phylogenetic relatedness in null communities generated at ran¬
dom. Random communities were generated by randomizing the
community data matrix using the independent swap algorithm 1000
times. Then, we computed the SES of MPD and the SES of MNTD by tak¬
ing the difference between the mean phylogenetic distances in the ob¬
served communities versus in the null communities, standardized by
the standard deviation of the phylogenetic distances in the null data
(SES = (mean ob s — mean nu ii)/sd n uii) (Webb et al., 2002; Swenson,
2014). We then multiplied the SES of MPD and the SES of MNTD by —
1, obtaining NRI and NTI, respectively. The NRI and NTI values indicated
whether taxa in the community were more closely related (positive
values) or less closely related (negative values) than expected by
chance. However, the two indices differ in phylogenetic scale. NRI re¬
flects information about whole phytogenies, while NTI reflects informa¬
tion about branch tips.
The phylogenetic indices were calculated using the pd. (PD), ses.mpd
(MPD and NRI), and ses.mntd (MNTD and NTI) functions in the picante
package in R. For each index, 1000 phylogenetic trees were simulated,
and the mean value was retained for use in further analyses.
To characterize functional diversity, we calculated Petchey and
Gaston's FD (hereafter PG-FD; Petchey and Gaston, 2002), as well as
the functional equivalents of MPD, MNTD, NRI, and NTI (hereafter, FD-
MPD, FD-MNTD, FD-NRI, and FD-NTI) (Table 2). PG-FD is the total
branch length of the functional dendrogram that results when species
are clustered in trait space (Petchey and Gaston, 2002). Here, the func¬
tional dendrogram was created by generating a Euclidean distance ma¬
trix from the z-standardized trait values of the different species and by
clustering the species using the unweighted pair group method with ar¬
ithmetic mean (UPGMA). PG-FD was calculated using the alpha function
in the BAT package in R. FD-MPD, FD-NRI, FD-MNTD, and FD-NTI were
calculated as described above, except that trait-based Euclidean dis¬
tances rather than phylogenetic distances (i.e., divergence times) were
used. Note that while Faith's PD and PG-FD are not abundance-
weighted indices, all the remaining indices are abundance-weighted,
i.e. they reflect trends in both abundance and evenness.
2.7. Statistical analyses
We used general linear models (GLMs) with a Gaussian distribution
error and “identity" link to analyze the relationships between the taxo¬
nomic, phylogenetic and functional diversity indices and the distur¬
bance and aridity gradients. We used a separate model for each
response variable (S, H, Faith's PD, MPD, MNTD, NRI, NTI, PG-FD, FD-
MPD, FD-MNTD, FD-NRI, and FD-NTI); the explanatory variables were
climatic water deficit, the LG index, the WE index, and the MU index. In¬
teractions between the disturbance indices and water deficit were also
included. We employed Akaike's information criterion with a correction
for finite sample sizes (AICc) to select the best-supported models; this
approach reduces the problems associated with multiple testing, co¬
linearity of explanatory variables, and small sample sizes (Burnham
and Anderson, 2002). All the initial models were full models. The best-
supported models were selected based on their AICc weights, which re¬
veal the relative likelihood of a given model—based on the data and the
fit—scaled to one; thus, models with a delta (AICc difference) of <2 were
selected (Burnham and Anderson, 2002). The relevant variables were
those that were retained in the best-supported models (except, obvi¬
ously, when the best-supported model consisted only of the intercept).
Model selection was carried out using the dredge function in the MuMIn
package in R.
2.8. Phylogenetic signals in ant functional traits
To test for phylogenetic signals (i.e., the degree of phylogenetic con¬
straint in species resemblance) in the six morphological traits, we com¬
puted Pagel's \ (Pagel, 1999). This index compares the observed
List of selected phylogenetic and functional diversity metrics, their description and references, and their range of values in this study.
Total branch length (divergence time) of the phylogenetic tree linking all the species represented in the
community. Not abundance-weighted.
Mean pairwise distance
Mean phylogenetic distance (in millions of years) between two randomly selected individuals within a
specific plot (considering conspecifics). Abundance-weighted.
Webb et al., 2002,
Mean phylogenetic distance (in millions of years) separating each individual in the community from its
closest heterospecific relative. Abundance-weighted.
Webb et al., 2002,
Net relatedness (NRI)
Quantifies the structure of a sample phylogeny derived from the mean phylogenetic distance, consequently
capturing the degree of clustering of the phylogeny from root to terminal leaves. Abundance-weighted.
Webb et al., 2002,
Nearest taxon index
Quantifies the terminal structure of the sample phylogeny, hence only captures the clustering of the terminal
nodes in the tree. Abundance-weighted.
Webb et al., 2002,
Petchey and Gaston's FD
Total branch length of the functional dendrogram that results when species are clustered in trait space. Not
Petchey and Gaston,
FD-MPD (MPD with
Mean functional distance between two randomly selected individuals within a specific plot (considering
conspecifics). Abundance weighted.
Webb et al., 2002,
FD-MNTD (MNTD with
Mean functional distance separating each individual in the community from its closest co-occurring relative.
Webb et al., 2002,
FD-NRI (NRI with
Quantifies the structure of a sample functional dendrogram derived from the mean functional distance,
consequently capturing the degree of clustering of the functional dendrogram from root to terminal leaves.
Webb et al., 2002,
FD-NTI (NTI with
Quantifies the terminal structure of the sample functional dendrogram, hence only captures the clustering of
the terminal nodes in the functional dendrogram. Abundance weighted.
Webb et al., 2002,
X. A man et al. / Science of the Total Environment 631 -632 (2018) 429-438
distribution of traits with the expected distribution of traits based on a
Brownian motion model of evolution. Values of 0 and 1 indicate the ab¬
sence and presence, respectively, of a phylogenetic signal under such
conditions. We computed \ values for each trait and for each of the
1000 simulated trees using the phylosig function in the phytools pack¬
age (Revell, 2012) in R. To conservatively test signal significance, we
used a likelihood ratio test based on the minimum values to estimate
the probability that the observed \ differed from the null \ value of 0.
It is important to note that, given that the simulated tree might lead to
an underestimation of phylogenetic signal, our estimates of lambda
are probably conservative.
3.1. Ant communities
Our traps captured representatives of 71 ant species belonging to 23
genera and 7 subfamilies (Appendix S4). Myrmicinae was by far the
most species-rich subfamily (39 species), followed by Dolichoderinae
and Formicinae (12 and 9 species, respectively). The most species-rich
genus was Pheidole (14 species), followed by Dorymyrmex (9 species)
and Solenopsis (8 species); Camponotus, Cephalotes, and Pseudomyrmex
were represented by 6 species each. The most frequently occurring spe¬
cies were Ectatomma muticum (61.1% of all traps), Solenopsis virulens
(44.8% of all traps), and Dinoponera quadriceps (41.2% of all traps).
Dinoponera quadriceps occurred on all 20 plots, while Ectatomma
muticum and Solenopsis virulens occurred on 19 plots. Nineteen and 18
species occurred on one and two plots, respectively. We found between
15 and 26 species per plot. Species richness and species diversity were
not modulated by the aridity and disturbance gradients (Appendix S5).
3.2. Phylogenetic and functional diversity along disturbance and aridity
The best-supported models retained climatic water deficit (called
aridity hereafter), the LG index, and the MU index, but not the WE
index (Table 3, Appendix S5). Faith's PD decreased along the LG gradient
(Table 3, Fig. 2a, R 2 = 0.31). The NRI decreased along the aridity, LG
(Fig. 2b, R 2 = 0.02), and MU gradients (Table 3): communities
transitioned from being more closely related to less closely related.
This result indicates that ants coexisting under the same aridity or dis¬
turbance conditions are non-randomly phylogenetically clustered. Fur¬
thermore, the NRI was influenced by an interaction between aridity
and MU (Fig. 3a, R 2 = 0.64). More specifically, the decrease in the NRI
along the aridity gradient was much stronger in areas with lower MU,
and the negative relationship between the NRI and MU became positive
when aridity was greater. None of the variables considered here helped
explain MPD, MNTD, or NTI values.
PG-FD decreased as aridity and LG increased (Table 3, Fig. 2c,d, R 2 =
0.11 and R 2 = 0.22, respectively). FD-MPD decreased along the aridity
gradient (Table 3, Fig. 2e, R 2 = 0.16). The FD-NRI was not associated
with any of the explanatory variables (Table 3). FD-MNTD generally de¬
creased along the aridity, LG, and MU gradients (Table 3), but there
were also complex aridity-by-disturbance interactions. More specifi¬
cally, when aridity was lower, FD-MNTD increased faster as MU de¬
clined, and FD-MNTD only decreased along the aridity gradient when
MU was lower (Fig. 3b, R 2 = 0.20). At the same time, FD-MNTD de¬
creased and increased with LG at low and high levels of aridity, respec¬
tively. It also increased and decreased with aridity at low and high levels
of LG (Fig. 3c, R 2 = 0.22), respectively. The FD-NTI generally increased
along the aridity and MU gradients (Table 3), meaning that communi¬
ties transitioned from being functionally overdispersed to functionally
clustered. However, there was once again an interaction between arid¬
ity and disturbance. The positive relationship between disturbance and
the FD-NTI became negative in the most arid areas, and the positive re¬
lationship between aridity and the FD-NTI became negative in the most
disturbed areas (Table 3, Fig. 3d, R 2 = 0.56).
3.3. Phylogenetic signals in ant functional traits
Strong, significant (p < 0.0001) phylogenetic signals were present in
all traits, indicating that they were phylogenetically conserved. The only
exception was relative clypeus length (p = 0.334) (Appendix S6).
Our results highlight how ant phylogenetic diversity and functional
diversity changed, despite unmodified taxonomic diversity, along CAD
and aridity gradients in the Caatinga, a type of SDTF found exclusively
in Brazil. In the 21st century, increased CAD and aridity are the main
threats faced by biota in dry tropical regions (Dirzo et al., 2011 ). Overall,
we found that functional diversity and, to a lesser extent, phylogenetic
diversity, decreased with increasing aridity. Both types of diversity
also decreased as anthropogenic disturbance increased; livestock graz¬
ing and miscellaneous resource use had an influence while wood ex¬
traction did not. Remarkably, we also found that anthropogenic
disturbance and aridity interacted to shape biodiversity responses. We
discovered clear evidence that the main mechanism involved was hab¬
We found support for our first hypothesis: in general, increased
levels of disturbance and aridity decreased ant phylogenetic diversity
and functional diversity. However, these relationships were somewhat
complex because different patterns were observed in different diversity
metrics. Also, environmental gradients interacted in different ways. The
patterns were especially complex in the case of phylogenetic diversity.
For example, increased levels of livestock grazing reduced the amount
Variables retained in the best-supported models analyzing how ant phylogenetic diversity and functional diversity change along aridity (climatic water deficit) and anthropogenic distur¬
bance gradients in the Caatinga (Catimbau National Park, Pernambuco, NE Brazil). The positive and negative signs depict the direction of the relationship; an “X” signals that an interaction
was present. The absence of any sign means there was no relationship between the response variable and the explanatory variable.
Response variable Water deficit (WD)
Miscellaneous resource use (MU)
Livestock grazing (LG) Wood extraction (WE) MU x WD
Interactions WE x WD
X. Arrian et al./Science of the Total Environment 631-632 (2018) 429-438
Fig. 2. Linear relationships between ant phylogenetic diversity and functional diversity metrics and the explanatory variables retained in the best-supported models: (a) Faith's PD and
livestock grazing; (b) NR1 and livestock grazing; (c) PG-FD and climatic water deficit; (d) PG-FD and livestock grazing; and (e) FD-MPD and climatic water deficit.
of evolutionary history shared by ants in the community (decrease in
Faith's PD), but increased (although weekly) the non-random mean
phylogenetic distance between ant species (decrease in the NRI).
These results suggest that livestock grazing might act as an important
filter, removing phylogenetically distant groups. Indeed, previous stud¬
ies in the same area (Arcoverde et al., unpublished data; Oliveira et al„
unpublished data) and in other Caatinga areas (Oliveira et al., 2017)
found that populations of Dinoponera quadriceps (subfamily Ponerinae)
decreased as disturbance increased. Alternatively, increased livestock
grazing might non-randomly remove species from the most species-
rich lineages or add species from relatively distant lineages (although
not as distant as Ponerinae). Meanwhile, aridity and miscellaneous re¬
source use appeared to have positive effects on phylogenetic diversity:
as both pressures increased, the NRI decreased, which meant there
was more phylogenetic overdispersion. Thus, these environmental gra¬
dients are important abiotic filters that may structure ant communities
by removing species from different lineages in the most humid or least
disturbed areas. Conversely, they may supply species from different lin¬
eages when disturbance or aridity increase. However, in the most arid
areas, the NRI increased as disturbance increased, meaning that ant
communities in areas experiencing high levels of both disturbance and
aridity were more phylogenetically clustered. These results provide
support for our second hypothesis—that disturbance and aridity can in¬
tensify each other's negative impacts on diversity. They also draw atten¬
tion to the synergistic effects that anthropogenic disturbance and
climate change may have on biodiversity (Hirota et al., 2011; Ponce-
Reyes et al., 2013; Gibb et al., 2015a; Frishkoff et al., 2016).
Disturbance and aridity play strong and consistent roles in reducing
ant functional diversity, as evidenced by the functional responses that
we observed along the disturbance and aridity gradients. Interestingly,
these patterns cannot be explained by chance alone, since the effects
were significant for the FD-NTI. Thus, aridity and certain forms of an¬
thropogenic disturbance (i.e., livestock grazing and miscellaneous re¬
source use) act as important abiotic filters of ant functional diversity
in the Caatinga. More specifically, greater aridity and disturbance pro¬
duced a major drop in the diversity of functions related to food acquisi¬
tion and foraging habits. Interestingly, however, they did so in an
interactive way. While miscellaneous resource use had very weak ef¬
fects on FD-MNTD and the FD-NTI in the most arid areas of the park,
its effects were quite strong in the less arid areas, which contradicts
our second hypothesis. Given that functional diversity is already
impoverished in highly arid areas, there might not be much room for
further loss. In the less arid and more diverse areas of the park, however,
disturbance can exert a much stronger influence. The effects of livestock
grazing on FD-MNTD were also modulated by aridity: positive effects
were found in the most arid areas, while negative effects were found
in the least arid areas. Taken together, these findings present new evi¬
dence that anthropogenic disturbance and climate change, acting in
tandem, can have complex effects on biodiversity (Travis, 2003;
Ponce-Reyes et al., 2013; Garcia-Valdes et al., 2015; Rito et al., 2017).
X. Arnan et al. / Science of the Total Environment 631-632 (2018) 429-438
n i i r
10 20 30 40 50 60
Miscellaneous resource use
10 20 30 40 50 60
Miscellaneous resource use
10 20 30 40 50
£ 700 -
20 30 40 50 60
Miscellaneous resource use
Fig. 3. Contour plots showing model results for the interactive effects of (a) climatic water deficit and miscellaneous resource use on the NRI; (b) climatic water deficit and miscellaneous
resource use on FD-MNTD; (c) climatic water deficit and livestock grazing on FD-MNTD; and (d) climatic water deficit and miscellaneous resource use on the FD-NTI.
Our results do not match those of previous studies that analyzed the
effects of aridity or precipitation (here, precipitation and water deficit
were highly correlated: r = 0.98) on ant phylogenetic diversity and
functional diversity. For instance, Arnan et al. (2015) examined geo¬
graphical gradients in central and western Europe, Machac and collabo¬
rators (2011) examined three altitudinal gradients in the USA, and
Smith (2015) examined several altitudinal gradients worldwide. All
three found that mean precipitation had a weak to non-existent influ¬
ence on ant phylogenetic diversity. Instead, patterns of phylogenetic di¬
versity were primarily driven by mean temperature. In the case of
functional diversity, no effects of mean precipitation were found along
elevational gradients in northwestern Patagonia (Argentina)
(Werenkraut et al., 2015) or across different vegetation types in the cen¬
tral North Kimberley region of Australia's seasonal tropics (Cross et al.,
2016). In western and central European ant communities, functional di¬
versity was found to be shaped by mean precipitation but was lowest in
the wettest areas (Arnan et al., 2015). This pattern was attributed to re¬
laxed local competition in areas with high levels of primary productivity
and resource availability (Pavoine and Bonsall, 2011). Although mean
precipitation does not generally seem to be an important driver of ant
community structure at large spatial scales (Dunn et al., 2009;
Andersen et al., 2015), our results suggest that water availability
might significantly shape community structure in some regions or bi-
omes, such as semi-arid areas where water is scarce (Parr et al., 2004).
In contrast, our discovery that disturbance reduced ant phylogenetic
diversity and functional diversity corresponds to what has been found
in many other studies (e.g., Bihn et al., 2010; Arnan et al., 2013, 2015;
Liu et al., 2016). Interestingly, these studies explored the effects of
acute disturbances—mainly land-use changes—which significantly
drive down biodiversity (Sala et al., 2000). Other studies have also re¬
ported that intense grazing has negative effects on ant functional diver¬
sity in semi-arid areas (e.g., Chillo et al., 2017; Oliveira et al., 2017), but
effects on invertebrate phylogenetic diversity have never before been
reported. Our study is the first to highlight that small but constant bio¬
mass removal (i.e„ chronic disturbance) can have similar deleterious ef¬
fects on phylogenetic diversity and functional diversity as large, sudden
biomass removal (i.e., acute disturbance). These findings are notewor¬
thy given concerns about the functional consequences of current biodi¬
versity losses (Bellard et al., 2012) that result from acute and chronic
disturbances alike (Barlow et al., 2016).
We found strong phylogenetic signals in all the functional traits we
measured (with the exception of relative clypeus length), indicating
that more closely related ant species share more similar functional
traits. These results agree with those of other studies that found signif¬
icant and, frequently, strong phylogenetic signals in ant morphological
traits (Machac et al., 2011; Donoso, 2014; Arnan et al., 2017). They
also suggest that the functional morphological traits of ant species in
the Caatinga are evolutionarily conserved, and consequently, a strong
correlation between PD and FD patterns is to be expected (Webb
et al., 2002; Cavender-Bares et al., 2009). However, we found that the
phylogenetic diversity and functional diversity indices did not respond
in the same way to disturbance and aridity gradients (Table 1 ). Other
studies across very different taxonomic groups have observed similar
mismatches, even when strong phylogenetic signals exist
(e.g., Purschke et al., 2013 for plants; Devictor et al., 2010 for birds;
Safi et al., 2011 for mammals; Arnan et al., 2015, 2017 and Liu et al.,
2016 for ants). Our results thus lend further support to the idea that
the environment may strongly condition covariation between different
diversity components via differential filtering (Safi et al., 2011; Arnan
et al., 2015, 2017).
From a conservation perspective, our results echo recent work dem¬
onstrating that CADs and aridity are immediate threats to biodiversity in
SDTFs (e.g., Ribeiro et al., 2015, 2016; Ribeiro-Neto et al., 2016; Oliveira
et al., 2017; Rito et al., 2017). However, this study is the first to describe
these detrimental effects in an animal taxon in a species-rich SDTF such
as the Caatinga. Ants provide a variety of key ecosystem services and
disservices in most terrestrial ecosystems (Del Del Toro et al., 2012);
these services are largely related to species dietaiy ecology. It is thus
likely that a decline in both ant phylogenetic diversity and functional di¬
versity (but especially in the latter) could have downstream effects on
ecosystem processes, plant populations, and non-ant insect
X. Arrian et al. / Science of the Total Environment 631-632 (2018) 429-438
We need a clear understanding of the main factors threatening bio¬
diversity in SDTFs. Conducted in Catimbau National Park, our study pro¬
vides evidence that ant phylogenetic diversity and functional diversity
can be deterministically impoverished due to increased anthropogenic
disturbance and aridity, even if absolute levels of ant species diversity
remain unchanged. More alarmingly, aridity can intensify the negative
effects of disturbance. Taken together, our results underscore concerns
about what will happen under future global change scenarios in neo¬
tropical semi-arid regions. These regions are already facing major de¬
clines in precipitation and constant-to-increasing anthropogenic
exploitation of forest resources (Magrin et al., 2014). However, we ob¬
served that anthropogenic disturbance had most negative impacts in
the wettest areas, which contain the highest levels of phylogenetic di¬
versity and functional diversity. Caatinga conservation policies must
thus give special priority to the wettest areas, where biodiversity loss
could be the highest. Finally, our findings strongly suggest that studies
in the Caatinga must address functional and phylogenetic diversity in
addition to species richness if they wish to uncover how ant communi¬
ties are reorganized after disturbance and how climate change modu¬
lates this process.
We are very grateful to Rodrigo Feitosa for helping to identify the
ants, to Davi Jamelli for providing Fig. 1, and to Jessica Pearce-Duvet
for editing the manuscript's English. This study was funded by the Foun¬
dation for Science and Technology Support of the State of Pernambuco
(FACEPE; APQ 06012.05/15, APQ 0738-2.05/12, and PRONEX 0138-
2.05/14), the Brazilian National Council for Scientific and Technological
Development (CNPq; PELD 403770/2012-2, Universal 470480/2013-0),
and the Rufford Small Grants Foundation (RSG 17372-1). CNPq receives
thanks from XA for his postdoctoral grants (PDS-167533/2013-4 and
PDS-165623/2015-2), from GBA for her scholarship (236918/2012-5),
and from IRL for her research grants (Produtividade 305611/2014-3).
Appendices 1-6. Supplementary data
Supplementary data to this article can be found online at https://doi.
Andersen, A.N., Del Toro, I., Parr, C.L, 2015. Savanna ant species richness is maintained
along a bioclimatic gradient of increasing latitude and decreasing rainfall in northern
Australia. J. Biogeogr. 42, 2313-2322.
Anderson-Teixeira, K.J., Miller, A.D., Mohan, J.E., Hudiburg, T.W., Duval, B.D., DeLucia, E.H.,
2013. Altered dynamics of forest recovery under a changing climate. Glob. Chang.
Biol. 19, 2001-2021.
Arnan, X., Cerda, X., Rodrigo, A., Retana, J., 2013. Response of ant functional composition
to fire. Ecography 36,1182-1192.
Arnan, X., Cerda, X., Retana, J., 2015. Partitioning the impact of environment and spatial
structure on alpha and beta components of taxonomic, functional, and phylogenetic
diversity in European ants. PeerJ 3, el 241.
Aman, X., Cerda, X., Retana, J., 2017. Relationships among taxonomic, functional, and phy¬
logenetic ant diversity across the biogeographic regions of Europe. Ecography 40,
Baccaro, F.B., Feitosa, R.M., Fernandez, F., Fernandes, I.O., Izzo, T.J., Souza, J.L.P., Solar, R.,
2015. Guia para os generos de formigas do Brasil. INPA, Manaus, Brasil.
Barlow, J., Lennox, G.D., Ferreira, J., Berenguer, E., Lees, A.C., Mac Nally, R., et al., 2016. An¬
thropogenic disturbance in tropical forests can double biodiversity loss from defores¬
tation. Nature 35,144-147.
Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., Courchamp, F., 2012. Impacts of cli¬
mate change on the future of biodiversity. Ecol. Lett. 15, 365-377.
Bihn, J.H., Gebauer, G., Brandi, R., 2010. Loss of functional diversity of ant assemblages in
secondary tropical forests. Ecology 91, 782-792.
Blois, J.L., Williams, J.W., Fitzpatrick, M.C., Jackson, S.T., Ferrier, S„ 2013. Space can substi¬
tute for time in predicting climate-change effects on biodiversity. Proc. Natl. Acad. Sci.
U. S. A. 110, 9374-9379.
Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: a
Practical Information-theoretic Approach. Springer-Verlag, New York.
Cavender-Bares, J., Kozak, K.H., Fine, P.VA, Kembel, S.W., 2009. The merging of commu¬
nity ecology and phylogenetic biology. Ecol. Lett. 12, 693-715.
Chillo, V., Ojeda, R.A., Capmourteres, V., Anand, M., 2017. Functional diversity loss with in¬
creasing livestock grazing intensity in dry lands: the mechanisms and their conse¬
quences depend on the taxa. J. Appl. Ecol. 54, 986-996.
Cross, A.T., Myers, C., Mitchell, C.N.A., Cross, S.L., Jackson, C., Waina, R., et al., 2016. Ant bio¬
diversity and its environmental predictors in the North Kimberley region of
Australia's seasonal tropics. Biodivers. Conserv. 24,1727-1759.
Davidson, D.W., Cook, S.C., Snelling, R.R., 2004. Liquid-feeding performances of ants
(Formicidae): ecological and evolutionary implications. Oecologia 139,255-266.
Del Toro, I., Ribbons, R.R., Pelini, S.L., 2012. The little things that run the world revisited: a
review of ant-mediated ecosystem services and disservices (Hymenoptera:
Formicidae). Myrmecol. News 17,133-146.
Development Core Team, R., 2016. R: a language and environment for statistical comput¬
ing. R Foundation for Statistical Computing. Austria. URL, Vienna https://www.R-pro-
Devictor, V., Mouillot, D., Meynard, C., Jiguet, F., Thuiller, W., Mouquet, N., 2010. Spatial
mismatch and congruence between taxonomic, phylogenetic and functional diver¬
sity: the need for integrative conservation strategies in a changing world. Ecol. Lett.
Dirzo, R., Young, H.S., Mooney, H.A., Ceballos, G., 2011. Seasonally Dry Tropical Forests.
Ecology and Conservation. Island Press.
Donoso, D.A., 2014. Assembly mechanisms shaping tropical litter ant communities.
Ecography 37, 490-499.
Dunn, R.R., Agosti, D., Andersen, A.N., Arnan, X., Bruhl, CA, Cerda, X., et al., 2009. Climatic
drivers of hemispheric asymmetry in global patterns of ant species richness. Ecol.
Lett. 12, 324-333.
Faith, D.P., 1992. Conservation evaluation and phylogenetic diversity. Biol. Conserv. 61,
1 - 10 .
Forest, F., Grenyer, R., Rouget, M., Davies, J., Cowling, R.M., Faith, D.P., et al., 2007.
Preserving the evolutionary potential of floras in biodiversity hotspots. Nature
Frishkoff, L.O., Karp, D.S., Flanders, J.R., Zook, J., Hadly, E.A., Daily, G.C., et al., 2016. Climate
change and habitat conversion favour the same species. Ecol. Lett. 19,1081-1090.
Garda-Valdes, R., Svenning, J.C., Zavala, MA, Purves, D.W., Araujo, M.B., 2015. Evaluating
the combined effects of climate and land-use change on tree species distributions.
J. Appl. Ecol. 52, 902-912.
Gibb, H., Sanders, N.J., Dunn, R.R., Photakis, M., Andersen, A.N., Angulo, E., et al., 2015a. Cli¬
mate regulates the effects of anthropogenic disturbance on ant assemblage structure.
Proc. R. Soc. Lond. B Biol. Sci. 282, 20150418.
Gibb, H., Stoklosa, J., Warton, D., Brown, A., Andrew, N., Cunningham, S., 2015b. Does mor¬
phology predict trophic position and habitat use of ant species and assemblages?
Hirota, M., Holmgren, M., Van Nes, E.H., Scheffer, M., 2011. Global resilience of tropical
forest and savanna to critical transitions. Science 334,232-235.
Holldobler, B., Wilson, E.O., 1990. The Ants. Springer-Verlag, Berlin-Heidelberg, Germany.
Kaspari, M., Weiser, M., 1999. The size-grain hypothesis and interspecific scaling in ants.
Funct. Ecol. 13, 530-538.
Klingenberg, C., Brandao, C.R.F., 2009. Revision of the fungus-growing ant genera
Mycetophylax Emery and Paramycetophylax Kusnezov rev. stat., and description of
Kalathomyrmex n. gen (formicidae: Myrmicinae: Attini). Zootaxa 2052,1-31.
Leal, I.R., Tabarelli, M., Silva, J.M.C., Larcher, T.E., 2005. Changing the course of biodiversity
conservation in the Caatinga of northeastern Brazil. Conserv. Biol. 19, 701-706.
Liu, C., Guenard, B., Blanchard, B., Peng, Y.-Q., Economo, E.P., 2016. Reorganization of tax¬
onomic, functional, and phylogenetic ant biodiversity after conversion to rubber
plantation. Ecol. Monogr. 86, 215-227.
Lutz, J.A., Van Wagtendonk, J.W., Franklin, J.F., 2010. Climatic water deficit, tree species
ranges, and climate change in Yosemite National Park. J. Biogeogr. 37, 936-950.
Machac, A., Janda, M., Dunn, R.R., Sanders, N.J., 2011. Elevational gradients in phylogenetic
structure of ant communities reveal the interplay of biotic and abiotic constraints on
diversity. Ecography 34, 364-371.
Magrin, G.O., Marengo, J.A., Boulanger, J.P., Buckeridge, M.S., Castellanos, E., Poveda, G.,
Vicuna, S., 2014. Central and South America. In: Boulanger, J.P., Buckeridge, M.S.,
Castellanos, E., Poveda, G., Scarano, F.R., Vicuna, S. (Eds.), Climate Change 2014: Im¬
pacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Work¬
ing Group II to the Fifth Assessment Report of the Intergovernmental Panel on
Climate Change. Cambridge University Press, Cambridge.
Miles, L., Newton, A.C., DeFries, R.S., Ravilious, C., May, I., Blyth, S., et al., 2006. A global
overview of the conservation status of tropical dry forests. J. Biogeogr. 33,491-505.
Moreau, C.S., Bell, C.D., 2013. Testing the museum versus cradle biological diversity hy¬
pothesis: phylogeny, diversification, and ancestral biogeographic range evolution of
the ants. Evolution 67, 2240-2257.
Morion, H., Schwilk, D.W., Bryant, J.A., Marquet, P.A., Rebelo, A.G., Tauss, C, et al., 2011.
Spatial patterns of phylogenetic diversity. Ecol. Lett. 14,141-149.
Mouillot, D., Graham, N.A.J., Villeger, S., Mason, N.W.H., Bellwood, D.R., 2013. A functional
approach reveals community responses to disturbances. Trends Ecol. Evol. 28,
Oliveira, A.P.C., Bernard, E., 2017. The financial needs vs. the realities of in situ conserva¬
tion: an analysis of federal funding for protected areas in Brazil's Caatinga. Biotropica
Oliveira, F.M.P., Ribeiro-Neto, J.D., Andersen, A.N., Leal, I.R., 2017. Chronic anthropogenic
disturbance as secondary driver of ant community structure: interactions with soil
type in Brazilian Caatinga. Environ. Conserv. 44,115-123.
Overbeck, G.E., Velez-Martin, E., Scarano, F.R., Lewinsohn, T.M., Fonseca, C.R., Meyer, S.T.,
et al., 2015. Conservation in Brazil needs to include non-forest ecosystems. Divers.
X. Arrian et al. / Science of the Total Environment 631-632 (2018) 429-438
Pagel, M., 1999. Inferring the historical patterns of biological evolution. Nature 401,
Parr, C.L, Robertson, H.G., Biggs, H.C., Chown, S.L., 2004. Response of African savanna ants
to long-term fire regimes. J. Appl. Ecol. 41, 630-642.
Parr, C.L., Dunn, R.R., Sanders, N.J., Weiser, M.D., Photakis, M., Fitzpatrick, M.C., et al., 2017.
GLobal ants trait database (GLAD): a new database on the geography of ant traits
(hymenoptera: Formicidae). Insect Conserv. Divers. 10, 5-20.
Pavoine, S., Bonsall, M.B., 2011. Measuring biodiversity to explain community assembly: a
unified approach. Biol. Rev. 86, 792-812.
Petchey, O.L, Gaston, K.J., 2002. Functional diversity (FD), species richness and commu¬
nity composition. Ecol. Lett. 5, 402-411.
Petchey, O.L., Gaston, K.J., 2006. Functional diversity: back to basics and looking forward.
Ecol. Lett. 9, 741-758.
Ponce-Reyes, R., Nicholson, E., Baxter, P.W.J., Fuller, R.A., Possingham, H., 2013. Extinction
risk in cloud forest fragments under climate change and habitat loss. Divers. Distrib.
Purschke, 0., Schmid, B.C., Sykes, M.T., Poschlod, P., Michalski, S., Durka, W., et al., 2013.
Contrasting changes in taxonomic, phylogenetic and functional diversity during a
long-term succession: insights into assembly processes. J. Ecol. 101, 857-866.
Revell, L.J., 2012. Phytools: an R package for phylogenetic comparative biology (and other
things). Methods Ecol. Evol. 3, 217-223.
Ribeiro, E.M.S., Arroyo-Rodrfguez, V., Santos, BA., Tabarelli, M., Leal, I.R., 2015. Chronic an¬
thropogenic disturbance drives the biological impoverishment of the Brazilian
Caatinga vegetation. J. Appl. Ecol. 52, 611-620.
Ribeiro, E.M.S., Santos, B.A., Arroyo-Rodrfguez, V., Tabarelli, M., Souza, G., Leal, I.R., 2016.
Phylogenetic impoverishment of plant communities following chronic human distur¬
bances in the Brazilian Caatinga. Ecology 97,1583-1592.
Ribeiro-Neto, J.D., Arnan, X., Tabarelli, M., Leal, I.R., 2016. Chronic anthropogenic distur¬
bance causes homogenization of plant and ant communities in the Brazilian Caatinga.
Biodivers. Conserv. 25, 943-956.
Rito, K.F., Arroy-Rodrfguez, V., Queiroz, R.T., Leal, I.R., Tabarelli, M., 2017. Precipitation me¬
diates the effect of human disturbance on the Brazilian Caatinga vegetation. J. Ecol.
Safi, K., Cianciaruso, M.V., Loyola, R.D., Brito, D., Armour-Marshall, K., Diniz-Filho, J.A.F.,
2011. Understanding global patterns of mammalian functional and phylogenetic di¬
versity. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 366, 2536-2544.
Sagar, R., Raghubanshi, A.S., Singh, J.S., 2003. Tree species composition, dispersion and di¬
versity along a disturbance gradient in a dry tropical forest region of India. For. Ecol.
Manag. 186, 61-71.
Sala, O.E., Chapin, F.S., Armesto, J.J., Berlow, E„ Bloomfield, J., Dirzo, R., et al., 2000. Global
biodiversity scenarios for the year 2100. Science 287,1770-1774.
Salas-Lopez, A., 2017. Predicting resource use in ant species and entire communities by
studying their morphological traits: influence of habitat and subfamily. Ecol. Indie.
Seddon, A.W.R., Macias-Fauria, M., Long, P.R., Benz, D., Willis, K.J., 2016. Sensitivity of
global terrestrial ecosystems to climate variability. Nature 531,229-232.
Singh, S.P., 1998. Chronic disturbance, a principal cause of environmental degradation 702
in developing countries. Environ. Conserv. 25,1-2.
Smith, M.A., 2015. Ants, elevation, phylogenetic diversity and community structure. Eco-
sphere 6, 221.
Sociedade Nordestina de Ecologia, 2002. Projeto Tecnico para a Criagao do Parque
Nacional do Catimbau/PE. Secretaria de Ciencia. Tecnologia e Meio Ambiente de Per¬
nambuco - SECTMA, Recife.
Swenson, N., 2014. Functional and Phylogenetic Ecology in R. Springer, New York.
Travis, J.M.J., 2003. Climate change and habitat destruction: a deadly anthropogenic cock¬
tail. Proc. R. Soc. Lond. B Biol. Sci. 270 (711), 467-473.
Ward, P.S., Brady, S.G., Fisher, B.L, Schultz, T.R., 2015. The evolution of myrmicine ants:
phylogeny and biogeography of a hyperdiverse ant clade (Hymenoptera:
Formicidae). Syst. Entomol. 40, 61-81.
Webb, C.O., David, D., Ackerly, D.D., McPeek, MA, Donoghue, M.J., 2002. Phylogenies and
community ecology. Annu. Rev. Ecol. Evol. Syst. 33,475-505.
Weiser, M.D., Kaspari, M., 2006. Ecological morphospace of new world ants. Ecol.
Werenkraut, V., Fergnani, P.N., Ruggiero, A., 2015. Ants at the edge: a sharp forest-steppe
boundary influences the taxonomic and functional organization of ant species assem¬
blages along elevational gradients in northwestern Patagonia (Argentina). Biodivers.
Conserv. 24, 287-308.