ns
e
,c,
Institute of Experimental Ecology, University of Ulm, Albert-Einstein-Allee 11, 89069 Ulm, Germany
d Laboratorio de Ecología, Universidad Industrial de Santander, Carrera 27, Calle 9, Ciudadela Universitaria, Bucaramanga, Colombia
a r t i c l e i n f o
Article history:
Received 11 September 2014
tropical forest degradation (Hill and Curran, 2003; Turner, 1996).
Type and intensity of land degradation has an effect on habitat
quality and structure, which in turn impacts species populations,
often in a negative way. At highly altered sites, tropical mammal
ko, 2011; Sattler
homogen
ey, 2006).
unctions a
vices (Foley et al., 2007; Klein et al., 2002; Maas et al., 2013
as the second largest mammalian order provide essential e
tem functions and services including pollination, seed di
and insect control (Castro-Arellano et al., 2007; Kunz et al.,
2011). Tropical bat studies have revealed that some assemblages,
such as gleaning animalivorous bat species, are negatively affected
by deforestation and urbanization, while large frugivorous bats
profit from such changes (Williams-Guillen and Perfecto, 2010)
and have higher diversity in areas with human impact than in
areas without (Garcia-Morales et al., 2013).
⇑ Corresponding author at: Swiss Federal Research Institute WSL, Zürcherstrasse
111, CH-8903 Birmensdorf, Switzerland. Tel.: +41 62 213 81 56.
E-mail addresses: ebader@gmx.ch (E. Bader), kirsten.jung@uni-ulm.de (K. Jung),
pager@si.edu (R.A. Page), rrodriguezuis@gmail.com (R. Rodriguez), thomas.
sattler@vogelwarte.ch (T. Sattler).
Biological Conservation 186 (2015) 97–106
Contents lists availab
Biological Co
.e ldeforestation and urbanization represents one of the most impor-
tant threats to biodiversity, especially in the tropics (e.g. Barlow
et al., 2007). Over the last decades, many studies have revealed
species loss and altered community composition in response to
munity composition drastically (Jung and Kal
et al., 2010; Threlfall et al., 2011) and may lead to
of urban species assemblages worldwide (McKinn
biodiversity often results in a loss of ecosystem fhttp://dx.doi.org/10.1016/j.biocon.2015.02.028
0006-3207/ 2015 Elsevier Ltd. All rights reserved.ization
Loss of
nd ser-
). Bats
cosys-
spersalinformation on occurrence and distribution is limited.
 2015 Elsevier Ltd. All rights reserved.
1. Introduction
Deterioration and fragmentation of original forest through
communities are characterized by decreased species richness and
diversity as well as by higher dominance compared to less dis-
turbed habitats (Ahumada et al., 2011). Urbanization affects com-Received in revised form 19 February 2015
Accepted 23 February 2015
Keywords:
Panama
Vulnerability
Wing morphology
Acoustic monitoring
Occupancy
Detectabilitya b s t r a c t
Habitat loss due to anthropogenic activities is a critical threat to biodiversity. Understanding which fac-
tors determine species persistence in degraded environments is essential for conservation management.
Here, we investigated how aerial insectivorous bats in the Neotropics, an ecologically important, but
understudied group of vertebrates, are affected by deforestation and urbanization. We conducted a stan-
dardized acoustic survey in four habitat types in 14 areas across Panama (500 km  260 km), a scale hith-
erto unprecedented for the Neotropics, and assessed occupancy in 13 aerial insectivorous bat species,
accounting for species- and habitat-specific detectability. In addition, we used wing measures to derive
a proxy for flight efficiency and flight speed, as an indication of species-specific mobility. Results show
that detectability does not only vary strongly among species but also, within a species, depending on
the habitat: for nine species intra-specific detectability varied >10%. Model estimates for occupancy
revealed that aerial insectivorous bats are highly heterogeneous in their response to deforestation and
urbanization. Canopy height and canopy cover predicted occupancy best for most species. Relating occu-
pancy to mobility we found that less mobile species with broad wings (high wing loading and low aspect
ratio) decrease occupancy in deforested and urban areas, while more mobile species with narrow wings
increase in these habitats. We suggest that this pattern applies also to species not analyzed in this study.
Our study exemplifies how morphological information can help define conservation priorities whena Institute of Evolutionary Biology and Environmental Studies, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland
b Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Ancón, Panama
cMobility explains the response of aerial i
to anthropogenic habitat change in the N
Elias Bader a,b,⇑, Kirsten Jung c, Elisabeth K.V. Kalko b
Thomas Sattler b,c
journal homepage: wwwectivorous bats
otropics
Rachel A. Page b, Raul Rodriguez b,d,
le at ScienceDirect
nservation
sevier .com/ locate /biocon
species- and site-specific detectability of AIBs. We tested the
hypothesis that more mobile bat species are more common in
nserAerial insectivorous bats (AIB) emit ultrasonic calls for orienta-
tion, foraging and communication. This species group has been lar-
gely omitted from Neotropical community analysis because the
sophisticated echolocation skills of these bats allow them to detect
and avoid mist nets, and thus evade capture, identification and
quantification (Kalko et al., 2008). As a result, knowledge on the
distribution and susceptibility to deforestation and urbanization
in AIB is scarce (but see Estrada and Coates-Estrada, 2002; Jung
and Kalko, 2011), resulting in weak baseline information for con-
servation threat status. The description of new, usually cryptic,
species (Baird et al., 2012; Baker et al., 2009; Clare et al., 2013) fur-
ther complicates the identification of AIB threat status (Sattler
et al., 2007). AIB constitute 30–50% of local bat species assem-
blages (Kalko et al., 2008), so omitting this group from bat commu-
nity analysis is a major drawback of many studies. Additionally,
AIB are of great value in multiple aspects of conservation: AIB
are good bioindicators as their position on a high trophic level
makes them sensitive to pesticides as well as to general changes
in prey abundance (Jones et al., 2009), and they are known to be
of important economic value, consuming pest insects (Boyles
et al., 2011).
Most AIB can be recognized by their species-specific echoloca-
tion calls, i.e., ultrasonic vocalizations emitted for orientation,
communication and foraging (Parsons and Szewczak, 2009).
Recent advances in acoustic survey techniques and continuously
growing libraries of reference calls for acoustic identification to
the level of species (e.g. Jung et al., 2007, 2014; Obrist et al.,
2004; Rydell et al., 2002; Walters et al., 2012) now allow the inclu-
sion of this group in bat community studies (Avila-Flores and
Fenton, 2005; Estrada-Villegas et al., 2012; Jung and Kalko,
2011). Acoustic surveys have helped to increase knowledge on
the distribution and ecology of AIB in temperate and tropical areas.
In temperate areas, AIB activity was shown to increase with struc-
tural connectivity in farmlands, but optimum connectivity varied
for short- and long-range echolocating bats (Frey-Ehrenbold
et al., 2013). Ethier and Fahrig (2011) found positive effects of for-
est fragmentation on relative abundance of various AIB species.
Jung et al. (2012) found that bat occurrence and activity was corre-
lated to structural forest heterogeneity. Studies on the isthmus of
Panama have shown that AIB react in more diverse ways to habitat
patch isolation than other bat trophic guilds (Estrada-Villegas
et al., 2010). While some AIB species disappear with urbanization
(urban avoiders), others seem rather unaffected (urban adapters)
or might even profit from human settlements, where they find high
numbers of roosting sites (urban exploiters, Jung and Kalko, 2011).
Most studies on AIB in the Neotropics, however, have focused on
geographically restricted areas like national parks or cenotes
(Delgado-Jaramillo et al., 2011; MacSwiney et al., 2008; Rydell
et al., 2002) or the canal area in Panama (Estrada-Villegas et al.,
2012; Jung and Kalko, 2010, 2011). Working across broader geo-
graphical scales is important to reveal broad scale patterns and
to test the validity of ecological hypotheses in multiple habitat
contexts. Here we investigated whether and how aerial insectivo-
rous bats respond to different land use types on a large scale in
Panama by sampling nationwide in different locations and multi-
ple landscape contexts.
Detectability, the likelihood of detecting a given species in a
survey with a given method, may vary significantly depending on
species, habitat, weather conditions, field method, observer, and
survey effort (e.g. Gorresen et al., 2008; Royle et al., 2005).
Ignoring the odds of detectability often leads to an underestima-
tion of site occupancy, which is the fraction of an area that is
inhabited by a given species. This can lead to erroneous estima-
98 E. Bader et al. / Biological Cotions of species richness, extinction rates, or modeled species dis-
tributions (Kéry, 2011; MacKenzie et al., 2006). AIB use different,
species-specific call frequencies, call structures and call intensitiesanthropogenically altered habitats than in primary habitats (for-
ests), and that less mobile species are more common in primary
than in anthropogenically altered habitats. By measuring the
relationship between occupancy and wing indices across different
habitats, we quantified the degree to which mobility explains sus-
ceptibility to deforestation and urbanization in AIBs. To test the
general validity of this hypothesis, we collected data across the
entire country of Panama (500 km  260 km), a scale hitherto
unprecedented for the Neotropics.
2. Material and methods
2.1. Study area and study design
Panama is dominated by a tropical climate with average tem-
peratures around 27 C and annual precipitation from 1600 mm
to over 3020 mm per year (Condit et al., 2000). The meteorological
divide follows the central cordillera, splitting the country into a
wetter Caribbean and a dryer Pacific region.
To study the effects of anthropogenic habitat change on AIBs,
we chose a space-for-time substitution (Pickett, 1989). Studies
using space-for-time substitutions assume that, in species diver-
sity, what drives spatial variation also drives temporal changes.
Thus variation in one can be used as a proxy for variation in the
other. For site selection in the field, we first chose four habitat
types that span a gradient of anthropogenic disturbance: mature
forest, disturbed forest, pasture, and human settlement. Then, we
measured and analyzed the same environmental covariates in all
habitat types to know the effects on a continuous scale, rather than
in discrete units (see 2.3. and 2.6 for details). To study AIBs on a
broad scale, we replicated the sampling in the four habitat types
in 14 areas in a spatially stratified design throughout Panama
(Fig. 1). Site selection was influenced by availability of appropriate
habitat type, permissions for land use, and accessibility with field
equipment. Each habitat type in a given area contained one sam-(Kalko et al., 2008), all of which affect the ability of ultrasound
microphones to record them, which in turn affects detectability
in acoustic surveys (Adams et al., 2012). Hence, to calculate spe-
cies-specific detectability and occupancy, repeated surveys are
necessary (MacKenzie et al., 2006). The only acoustic sampling
study to date to calculate detectability for Neotropical AIBs found
an average detectability of 0.71 (i.e., given its presence, a species
is detected in 71% of all sampling events), a value considerably
higher than the detectability ranging from 0.25 to 0.58 calculated
for other bat trophic guilds through mist netting (Meyer et al.,
2011). Thus including detectability can be crucial when estimating
the effects of anthropogenic processes on species.
Mobility is a key trait affecting susceptibility to habitat change
in many species (Henle et al., 2004), including bats (Meyer et al.,
2008). Typically bat mobility is assessed using wing shape,
described as a function of aspect ratio (wing length to width)
and wing loading (body mass per wing area, Norberg and Rayner,
1987). More mobile bat species with small-surfaced and narrow
wings are able to fly fast, cover long distances between roosts
and foraging sites, and predominantly hunt in open space
(Norberg and Rayner, 1987). Less mobile species with large-sur-
faced and broad wings fly more slowly but have greater maneuver-
ability and rather hunt in habitats of high structural complexity
such as forests (Estrada-Villegas et al., 2010; Kalko, 1998).
In this study, we estimated occupancy by explicitly considering
vation 186 (2015) 97–106pling point, summing to a total of 14  4 = 56 points. As small alti-
tudinal differences can change the composition of tropical
vertebrate species communities (McCain, 2009), we restricted
nserE. Bader et al. / Biological Cosampling to lowlands <700 m above sea level. Sampling points
were positioned within habitat patches such that they met our dis-
tance criteria: we required that sampling points be a minimum of
400 m from the edge of the habitat type in mature forest, disturbed
forest, and pasture, and 100 m from the edge in settlement.
Additionally, sampling points were separated by a minimum dis-
tance of four kilometers to minimize the possibility of recording
an individual bat at multiple sampling points in one night, avoiding
spatial dependence of sampling points.
2.2. Field sampling
From 3 June to 14 September 2012, we sampled the four habitat
types of each area simultaneously (168 sample nights).
Simultaneous sampling in one area avoids possibly strong effects
of climatic variation on bat activity between sampling nights
(Fischer et al., 2009). Points were sampled in three non-consecutive
nights for three hours each, starting at sunset. This sampling regime
has been shown to yield >90% of all local AIB species (Estrada-
Villegas et al., 2010; Meyer et al., 2011). We restricted sampling to
nights with no or very limited rainfall (<20 mm). Bat calls were
recorded with Batloggers (Elekon AG, Lucerne, Switzerland), which
allow autonomous real-time recording. An electret condenser
microphone was connected to the Batlogger, positioned 120 cm
above ground at an angle of 45 above the horizontal, and protected
against humidity by a layer of tightly sealed cellophane. The single
microphone was faced in the direction where we expected AIB to
pass with highest probability. Usually, this was the direction least
obscuredwith obstacles in the forest (trees andbranches) and in set-
tlements (buildings) or along a fence or close to a tree in pastures. All
Fig. 1. Map of the study area in Panama. Each square represents a sampling area of th
Minimum distance between sampling points was 4 km.vation 186 (2015) 97–106 99fieldwork was conducted according to regulations of the
Smithsonian Tropical Research Institute and Panama’s Autoridad
Nacional del Ambiente (permit SE/A-54-12).
2.3. Habitat types and environmental covariates
For each sampling point in all habitat types, we recorded three
structural covariates and one climatic covariate, all known to affect
detectability and occupancy (Patriquin et al., 2003). In addition to
relevant habitat covariates, the structural covariates included veg-
etation clutter, which is known to attenuate bat calls before they
reach the microphone and thus affect species detectability
(Kalcounis et al., 1999). (1) We calculated canopy cover from 180
fisheye lens photos, taken 50 cm above ground, aimed toward
the sky (Canon EOS Digital Rebel with Sigma 4.5 mm F 2.8 EX DC
HSM Circular Fisheye Lens), and analyzed with software
ImageJ64 1.46r (Rasband, 2012). (2) We assessed canopy height
by visual estimation, taking the average of estimates of three
observers. (3) We calculated tree distance as the average distance
from our sampling point to the four nearest trees (more than
10 cm diameter at breast height) in four opposite directions, using
the same method of visual estimation. (4) As proxy for climate we
included orientation to the ocean (categorizing each sampling point
as either Atlantic or Pacific, with the central cordillera as the
dividing line), as Estrada-Villegas et al. (2012) found precipitation-
dependent distribution patterns in AIB in the Panama Canal area.
These measurements helped us to characterize the four habitat
types used for site selection in the field: (1) Mature forest, the habi-
tat type with least anthropogenic influence, was characterized by
extensive canopy cover, a high canopy and reduced distance
e spatially stratified design (n = 14). Each dot represents a sampling point (n = 56).
nserbetween trees (see Table A1 and Fig. A1 in the online appendix for
examples and values). We did not distinguish between pristine and
old-grown secondary forest, as distinction is difficult, habitat struc-
ture (e.g. tree height, canopy cover) has a greater impact on bat
diversity than other characteristics typical for pristine forests
(Jung et al., 2012; Mueller et al., 2012), and animal species richness
resembles that of mature forests only 20–40 years after farmed
land is abandoned (Dunn, 2004). Here we selected sampling points
in relatively open areas such as tree-fall gaps or forests with scanty
understory, so that bat calls were less attenuated by foliage, result-
ing in higher detectability. (2) Disturbed forest was characterized
by intermediate canopy coverage, low canopy and intermediate
tree distance. Tall trees were scarce. Disturbed forest may have
been impacted by substantial timber extraction or was regrowing
after land abandonment. (3) Pasture, usually grazed by cattle,
was habitat that had been cleared, did not have a forest canopy
and thus showed very low crown coverage. (4) Settlement was
the habitat of highest anthropogenic influence, where vegetation
was greatly reduced. Canopy coverage contained buildings and
was therefore intermediate. We avoided small settlements
<200 m in diameter and centers of large cities.
2.4. Call analysis
We selected ten species (Centronycteris centralis, Cormura brevir-
ostris, Peropteryx macrotis, Saccopteryx bilineata, Saccopteryx leptura,
Noctilio albiventris, Pteronotus parnellii, Pteronotus gymnonotus,
Myotis nigricans andMolossusmolossus) and three sonotypes (groups
of species with similar calls) for call and occupancy analysis.
Sonotypes were defined as follows: Vespertilionid > 56 (frequency
modulated – constant–frequent calls of the Vespertilionidae family
with an end frequency >56 kHz, stemming from Myotis riparius or
Myotis keaysi); Molossus < 35 (the large Molossus species M. bondae,
M. rufus and M. sinaloae, whose calls predominantly are below
35 kHz); and Cynomops sp. (in Panama probably C. cf. mexicanus
andC. planirostris). To simplify the text, the term ‘species’ henceforth
includes sonotypes. All focal species were common enough for sta-
tistical analysis and could be identified to their respective taxo-
nomic level with high confidence.
Bat calls were analyzed with Raven Pro 1.4 (Bioacoustics
Research Program, 2011) and Batscope 3.1.6 (Boesch and Obrist,
2013), which is both a database of reference calls and a software
program used to automatically measure echolocation call charac-
teristics. To ensure consistent species determination, criteria for
the different species were established based on call characteristics
(e.g. shape, duration, maximum, minimum and peak frequency)
taken from existing literature (Jung et al., 2007, 2009, 2014;
Rydell et al., 2002) and from our own reference calls. Species-speci-
fic calls are shown in Fig. A2. To obtain presence or absence of a
species at a given location during a specific sampling night, we
manually screened all sequences potentially corresponding to the
species of interest until either the species was found or all
sequences in question within the range of the possible call charac-
teristics for this species had been checked.
2.5. Mobility estimation
Wing morphology was measured with ImageJ64 (Rasband,
2012) from photos of the opened wing, positioned on 1 mm2 graph
paper. Photos for morphological data were obtained from our own
captures, as well as from the University of Ulm (Elisabeth Kalko
and Hans Dietrich von Staaden, unpublished data). For C. centralis,
no photos were available, so measurements were taken from
100 E. Bader et al. / Biological CoSanborn (1937). For specimens for which no body mass measure-
ments were available, we used average body masses from Reid
(2009). The measurements of two species, M. riparius and M. keaysi,were averaged for the sonotype Vespertilionid > 56, while M. bondae
was used for Molossus < 35 and Cynomops cf. mexicanus for
Cynomops sp. Due to the lack of data, we could not include all pos-
sible species for the latter two sonotypes. However, we consider
our approximation acceptable because there is little variation in
wing shape among Molossids.
Aspect ratio and wing loading, metrics derived from wing mor-
phology, are widely used calculations for estimating bat mobility
(Norberg and Rayner, 1987). Despite their frequent application,
cross-author comparisons are problematic because researchers vary
considerably in theirmeasurements, due to individual differences in
the position inwhich they position thewingwhenmeasuring. In our
study, in addition to taking our own measurements, we used pic-
tures from different authorswith differentwing positions, andwere
thus confronted with the challenge of integrating potentially differ-
ingmeasurement techniques. To circumvent this difficulty, we used
the following bone-based measurements – instead of measures of
wing area – as proxies for mobility.
Aspect ratio is based on wing length, which depends on how
much the wing is stretched open for measurement (Norberg and
Rayner, 1987). To circumvent inconsistencies associated with stan-
dard measurements for aspect ratios, Findley et al. (1972) intro-
duced a method that relied on bone measurements: Aspect Ratio
Index ARI = ((d3 + FA)/d5), where FA = forearm, d3 = third digit
and d5 = fifth digit.
Wing loading is defined as body weight divided by wing area
(Norberg and Rayner, 1987; Winkler and Leisler, 1992). The values
for wing loading depend not only on the position of thewing (which
varies by researcher) but also on a researcher’s definition of ‘wing’.
Researchers vary in their delineations of the membranes that com-
pose thewing, for example, differing in the amountof tailmembrane
that is included. To circumvent these inconsistencies, we replaced
the traditional Wing Loading with a more robust index: we defined
the Wing Loading Index WLI = m/((FA + d3)  d5  2), with
m = body mass. A Pearson Correlation between WLI and Wing
Loading based on our own pictures (angle between humerus and
forearm kept at 90) was 0.972 with p < 0.001 (n = 102). We thus
conclude that WLI, which is based on absolute bone-based mea-
sures, is well suited to replace the original wing loading
measurements.
2.6. Statistical analysis
Presence–absence data derived from call analysis and summed
as detection histories (detections and non-detections for the three
sampling nights per site) served as baseline data to calculate
detectability and occupancy. Data was analyzed for each species
with the software PRESENCE 5.5 (Hines, 2006). We assumed closed
populations, meaning that a site remained either occupied or unoc-
cupied for the entire study duration, a justifiable assumption since
we sampled over a rather short period (three months during wet
season). We further assumed detectability to be constant for a spe-
cies at a sampling point during the three survey repetitions. We
then assessed how environmental covariates affect occupancy of
different bat species by using an information-theory approach
based on maximum likelihood models (Burnham and Anderson,
2002). To identify the relevant environmental covariate sets, we
chose a two-step approach: first we calculated occupancy combin-
ing all different sets of the four covariates (orientation, tree dis-
tance, canopy cover and canopy height) obtaining 16 candidate
models for each species (Table A2). Occupancy models were ranked
by the Akaike Information Criterion corrected for small sample
size, AICc (Hurvich and Tsai, 1989). In a second step and to the
vation 186 (2015) 97–106plausible models only (AICc weight >10%), we added habitat as
an additional covariate. This way, we tested the relevance of the
covariates included in the first step on occupancy of a certain
habitat type by a given bat species. If these habitat models were
selected, then additional covariates not explicitly measured but
consisting of habitat were considered relevant for occupancy. We
then obtained species-specific detectability and occupancy for
each location from the most plausible model of this second step.
Location-specific values were then summarized by habitat, which
allowed us to calculate the mean and the variation of detectability
and occupancy per habitat.
In a further analysis, we related each species’ occupancy to its
wing morphology. We used the difference in mean occupancy of
the respective habitat to the mean of all habitats, as we were inter-
ested in the relative change of occupancy among habitats and not
in the absolute occupancy per se. Each species represented one
data point (n = 13). We studied the dependency of the difference
in occupancy to median wing indices with linear and quadratic
least square regressions. This analysis was performed with SPSS
20 (IBM, 2011).
3. Results
3.1. Occupancy and detectability modeling
Field surveys at 56 locations sampled three times each (total of
the analysis, explained M. molossus occupancy in the different habi-
tats. For two species, P. gymnonotus and S. leptura, the null models
without covariates were themost plausible, suggesting indifference
to habitat and the environmental covariates included in this study.
Infive species, orientationwas included in themostplausiblemodel.
Three species showed considerable differences in occupancy on the
twoslopeswithM.nigricansbeingmore commonon thePacific slope
and Vespertilionid > 56 and S. bilineata being more common on the
Caribbean slope.
Species- and habitat-specific detectability varied greatly
(Fig. A3) from 4.2% ± SD 3.7 (Vespertilionid > 56 in pastures) to
100% ± 0.0 (M. molossus in settlement, meaning M. molossus was
detected in every survey in this habitat type). For nine species
detectability among habitat types varied more than 10%, while
for four species this value varied less than 10% (Table 1). This
means that for most species, ignoring detectability in a cross-habi-
tat study would seriously affect results. Mean detectability for all
species in a single survey was 65.2% ± 22.1 and ranged from 31%
(Vespertilionid > 56) to 91% (M. molossus).
3.2. Occupancy and mobility
The study species varied greatly in their occupancies in the dif-
data consists of the number of surveys with detections (nmax per species: 56 sites  3
= 14, n total = 56) for all 13 species. Max. change in mean detectability (%) shows the
er sp
the habitats with highest and lowest occupancy per species.
ecti
Pas
0
7
11
13
12
12
11
13
12
14
7
14
2
133
E. Bader et al. / Biological Conservation 186 (2015) 97–106 101Species Field data
No. surveys with
detections
No. sampling points with det
Mature
forest
Disturbed
forest
Centronycteris centralis 21 5 4
Cormura brevirostris 31 4 6
Peropteryx macrotis 43 2 7
Saccopteryx bilineata 107 9 12
Saccopteryx leptura 95 9 12
Pteronotus gymnonotus 70 6 8
Pteronotus parnellii 84 13 9
Cynomops sp.a 83 3 13
Molossus molossus 66 1 10
Molossus < 35b 118 3 13
Noctilio albiventris 30 0 7
Myotis nigricans 130 9 13
Vespertilionid > 56c 22 6 5
Total 900 70 119
a504 h) resulted in 54,226 call sequences, from which we obtained
900 detections and 1284 non-detections for the detection histories
of the 13 selected species (for species-specific information, see
Table 1).
Selection ofmodelswithAICcweight >10% resulted in one to four
plausiblemodels per species, which revealed species-specific differ-
ences in the importance of environmental covariates (Table 2).
Environmental covariates explained the occupancy of all species
well,with the exceptionofM.molossusand S. leptura, suggesting that
the selected covariates indeed are crucial for AIB. The covariates that
best predicted the occupancy of most species were canopy height
and canopy cover, both present in plausible models of 11 species.
Five out of the 34 plausible models included habitat as a covariate
(Table 2). For just one species,M.molossus, themost plausiblemodel
was a model including only habitat as covariate. This means that
additional factors not reflected by the covariates we included into
Table 1
Overview of field data and model results for detectability and occupancy. The field
surveys = 168) and the number of sampling points with detections (nmax per habitat
difference in detectability between the habitats with highest and lowest detectability pCynomops cf. mexicanus and C. planirostris.
b Molossus bondae, M. rufus and M. sinaloae.
c Myotis riparius and M. keaysi.max
ecies. Max. change in mean occupancy (%) shows the difference in occupancy between
Model results
ons Max. change per species in
ture Settlement Total Mean detectability
(%)
Mean occupancy
(%)
0 9 70 23
2 19 12 12
4 24 7 64
9 43 17 31
7 40 8 29
7 33 23 4
5 38 9 44
14 43 23 71
14 37 20 93
14 44 9 71
3 17 33 33
13 49 8 29
2 15 29 9
94 411 20.6 39.5ferent habitat types (Fig. 2). Within species, the difference between
the habitat types with highest and lowest occupancy, based on the
most plausible model was on average 39.5%. For mobility estima-
tion, we obtained wing measurements of 132 individuals
representing 14 species (Fig. A4).
Relating occupancy and mobility, we identified four main occu-
pancy patterns among the 13 species when summarized to the four
habitat types (Fig. 2): (1) Less mobile species such as P. parnellii
and C. centralis were most common in mature forest and decreased
in anthropogenically altered habitats. C. centralis was never
observed in pastures and settlements (Table 1), the species’ occu-
pancy estimates >0 in those habitat types (Fig. 2) are a conse-
quence of the covariates included in the most plausible model
(Table 2). (2) Mobile species such as the molossids M. molossus,
Molossus < 35 and Cynomops sp. showed lowest occupancy in
mature forest and increased in the other habitats, often with
od m
del
opy
nserTable 2
Plausible occupancy models for 13 aerial insectivorous bat species. A maximum likeliho
of a-priori models in Table A2; X indicates that this covariate was included in this mo
without any covariates); k is the number of parameters per model.
Species Model no. Orientation Tree distance Can
Centronycteris centralis 9 X X
10 X
3 X
Cormura brevirostris 4 X
5
1
Peropteryx macrotis 10 X
9 X X
Saccopteryx bilineata 7 X X
14 X X
Saccopteryx leptura 1
Noctilio albiventris 5
–
10 X
11 X
Pteronotus parnellii 15 X X
10 X
–
102 E. Bader et al. / Biological Cohighest values in settlements suggesting that these species may
profit from anthropogenically altered conditions such as artificial
roosting sites and increased food accessibility at artificial lights.
Species included in occupancy pattern categories 3 and 4 are of
intermediate mobility. (3) M. nigricans, S. bilineata, P. macrotis, N.
albiventris and C. brevirostris had lowest occupancy values in
mature forests and highest in pastures with intermediate values
in the two other habitat types. Based on this study they preferably
forage in open spaces and are able to utilize those habitats created
by deforestation. (4) S. leptura, P. gymnonotus and
Vespertilionid > 56 exhibit no or very limited variation in occupancy
among the different habitat types, at least with the covariates
included in this study.
In summary, these findings suggest that among AIB in the
Neotropics, species with an ARI of less than 2.1 or a WLI of less than
0.4 should be classified as less mobile and thus as susceptible to
habitat change reducing mature forest coverage. Species with an
ARI greater than 2.8 or aWLI greater than 1.3 are classified asmobile
and thus seem to profit from human induced habitat change.
Differences in occupancy among habitats correlated well with
wing morphology (Fig. 3). Simple linear regressions were signifi-
cant for all habitats but pasture (marginally significant for WLI in
settlements, p = 0.062) and explained around 70% of the variation
between species in ARI and between 28% and 60% in WLI (exact
Pteronotus gymnonotus 1
11 X
13 X X
2 X
Vespertilionid > 56a 7 X X
13 X X
3 X
Myotis nigricans 7 X X
8 X
–
1
Cynomops sp.b 8 X
13 X X
11 X
Molossus molossus –
Molossus < 35c 8 X
–
a Myotis riparius and M. keaysi.
b Cynomops cf. mexicanus and C. planirostris.
c Molossus bondae, M. rufus and M. sinaloae.odel was considered plausible when its AICc weight > 10%. Model no. refers to full list
(Model no. 1 has no X which indicates the selection of the null model, i.e., the model
cover Canopy height Habitat k Delta AICc AICc weight (%)
4 0 24.53
X 4 0.69 17.25
2 1.19 13.42
2 0 35.48
X 2 0.71 24.88
2 2.24 10.58
X 4 0 57.42
4 2.82 14.02
6 0 33.96
X 8 1.74 14.23
2 0 65.15
X 2 0 31.48
X 8 0.28 27.37
X 4 1.89 12.23
X 4 2.26 10.17
X 6 0 30.19
X 4 1.48 14.40
X 8 1.8 12.27
vation 186 (2015) 97–106values in Fig. 3). Occupancy increased positively with increasing
wing indices in settlements (y = 0.34x  0.77) and in disturbed for-
est (y = 0.16x  0.38), which contrasted strongly with mature for-
est where occupancy showed a strong negative correlation
(y = 0.70x + 1.52).
4. Discussion
In this study, we used an approach involving space-for-time
substitution in different habitat types to assess the impact of
anthropogenic habitat change on AIB on a broad spatial scale
across the country of Panama. By correcting occurrence for the
detectability of bats we obtained true occupancy (sensu Kéry,
2011) which we related to mobility. We found that less mobile
species with large-surfaced, broad wings are more widely dis-
tributed in forests and decrease in occupancy in more open and
anthropogenically altered habitats like pastures and settlements.
Conversely, more mobile species with small-surfaced, narrow
wings show the opposite pattern. Mobility, therefore, has a high
impact on occupancy in a given habitat type, and as such can be
used as a predictor of susceptibility to habitat change. In general,
less mobile species forage in dense vegetation clutter where broad
wings enable for high maneuverability. Broad wings have been
identified as a factor contributing to higher extinction risk and thus
2 0 32.06
X 4 1.8 13.03
X 8 1.82 12.90
4 2.22 10.56
6 0 31.99
X 8 0.37 26.59
2 1.45 15.50
6 0 34.90
X 6 0.25 30.80
X 8 1.31 18.13
2 2.08 12.34
X 6 0 46.20
X 8 1.68 19.94
X 4 1.92 17.69
X 8 0 99.95
X 6 0 48.41
X 8 0.84 31.81
ƕƒƑ
nser1.0
0.8
0.6
0.4
ŦƈƑƗƕƒƑƜƆƗƈƕƌƖŃƆƈƑƗƕƄƏƌƖ ųƗƈ
E. Bader et al. / Biological Coincreased conservation concern in temperate zone bats (Safi and
Kerth, 2004) as well as in bats worldwide (Jones et al., 2003).
4.1. Differences in detectability affect occupancy
Naive occupancy estimates, i.e. those not corrected for
detectability, often deviate substantially from the estimates cor-
rected for detectability (here between 36% underestimation and
59% overestimation), indicating that environmental covariates are
ŹƈƖƓƈƕƗƌƏƌƒƑƌƇŃšŃŃŘř
1.0
0.8
0.6
0.4
0.2
0.0
ŰƒƏƒ
ŦƒƕƐƘƕƄŃƅƕƈƙƌƕƒƖƗƕƌƖ
1.0
0.8
0.6
0.4
0.2
0.0
ŶƄƆƆƒƓ
0.2
0.0
ŰƒƏƒƖƖƘƖŃƐƒƏƒƖƖƘƖ
1.0
0.8
0.6
0.4
0.2
0.0
ųƈƕƒƓƗƈ
O
cc
u
p
a
n
cy
ŶƄƆƆƒƓƗƈƕƜƛŃƅƌƏƌƑƈƄƗƄ
1.0
0.8
0.6
0.4
0.2
0.0
Fig. 2. Boxplots of the occupancy by species summarized by habitat type. Each boxplot s
best model (Table 2); n per habitat = 14; 1 signifies a species occupies every sampling poi
of a certain habitat type. The horizontal line represents the naive occupancy estimate foƒƗƘƖŃƓƄƕƑƈƏƏƌƌŃ ŦƜƑƒƐƒƓƖŃƖƓőŃ
vation 186 (2015) 97–106 103important factors in determining whether a bat is detected or
not. This finding highlights the need to take habitat- and species-
specific detectability into account when studying occupancy or
distribution. Otherwise, the variation in detectability biases
occupancy results, possibly leading to erroneous conclusions.
Our results showed, that inter-specific detectability varied from
31% to 91% and intra-specific detectability varied up to 70% among
different habitat types (Table 1). This is very likely due to the fact
that detectability in bats is affected by several factors, including
ƖƖƘƖŃşŃŖŘ
ųƗƈƕƒƑƒƗƘƖŃƊƜƐƑƒƑƒƗƘƖƗƈƕƜƛŃƏƈƓƗƘƕƄ
ŰƜƒƗƌƖŃƑƌƊƕƌƆƄƑƖ
ƕƜƛŃƐƄƆƕƒƗƌƖ űƒƆƗƌƏƌƒŃƄƏƅƌƙƈƑƗƕƌƖ
hows median, second and third quartile and 95% – CI of the values obtained by the
nt of a certain habitat type; 0 signifies a species does not occupy any sampling point
r each species without accounting for habitat-specific detectability.
ect 
g L
3.
2.0
e be
nserAsp
∆
 O
cc
u
p
a
n
cy
Win
mature forest disturbed forest 
2.0 2.2 2.4 2.6 2.8
R2=0.73 
p < 0.001
0.4
0.2
0.0
-0.2
-0.4
-0.6
R2=0.67 
p=0.001
2.0 2.2 2.4 2.6 2.8 3.0 
R2=0.60
p=0.002
0.5 1.0 1.5
-0.6
0.5 1.0 1.5 2.0
0.4 
0.2 
0.0 
R2=0.41 
p=0.019
-0.4
-0.2
Fig. 3. Linear regression of wing indices and occupancy. D occupancy is the differenc
104 E. Bader et al. / Biological Cocharacteristics of echolocation calls such as intensity and frequency
(Patriquin et al., 2003); species’ activity, the assessment of which is
in turn affected by the overall abundance; climate and time of
recording (McCarthy et al., 2013); and habitat characteristics such
as vegetation structure (e.g. canopy height, and canopy closure),
which can absorb echolocation calls of lower frequencies that are
not already attenuated by air (Kalcounis et al., 1999; Patriquin
et al., 2003). All of these factors contribute to thedetectability of spe-
cies. Thus it is essential to standardize recording equipment, record-
ing situations and include habitat characteristics when evaluating
species- and habitat-specific occupancy patterns correctly.
Our overall detectability of 65% in a single survey agrees well
with observations from Meyer et al. (2011), who found average
detectability with acoustic sampling to be 71% for Neotropical
AIB. Both values are substantially higher than the 36–40% observed
by Duchamp et al. (2006) for temperate North America, which
could be related to lower overall bat abundance in higher latitudes
or to methodological differences, e.g. in recording devices or survey
effort. This finding underlines the importance of considering
detectability not only as species-, but also as habitat-specific.
4.2. Wing measures as proxies for mobility
Our findings underline the usefulness of wing measurements to
assess the mobility of bats, corroborating former studies (Duchamp
and Swihart, 2008; Findley, 1993; Norberg and Rayner, 1987).
Additionally, our study highlights the importance and practicality
of wing measurements that rely on bone lengths instead of wing
area and wingspan. In line with the well-established Aspect Ratio
Index ARI describing energy efficiency (Findley et al., 1972), we
hereby introduce another bone-based Wing Loading Index WLI,
which is related to flight speed. Mobility is determined by both
energy efficiency and flight speed, but the two indices do not
Each point represents a species (n = 13).Ratio Index 
oading Index 
2.0 2.2 2.4 2.6 2.8 3.00 2.0 2.2 2.4 2.6 2.8 3.0
R2=0.17 
p=0.159
0.5 1.0 1.5 2.0 
R2=0.28
p=0.062
0.5 1.0 1.5 2.0
tween the species’ value and the mean occupancy of all species in a certain habitat.pasture settlement 
R2=0.21
p=0.114
R2=0.68
p=0.001
vation 186 (2015) 97–106necessarily coincide. While the two indices showed quite consis-
tent ranking among the more mobile species, rankings varied con-
siderably among the less mobile species (Fig. A4). Thus, for species
that can clearly be classified into distinct categories, the assess-
ment of the susceptibility to habitat change is quite straightfor-
ward (less mobile, negatively affected vs. mobile, positively
affected, Section 3.2, Fig. 2). For species in between, a prediction
of their response to anthropogenic habitat change is more com-
plex. In such cases it is useful to assess mobility from several per-
spectives, e.g. with both, ARI and WLI, to better relate these traits
to susceptibility to habitat change. These indices are not indepen-
dent across species, however, because a species’ mobility is not
independent from phylogeny. This is most apparent in molossids,
where nearly all species are highly mobile (Vaughan, 1966).
However, this objection does not limit the utility of wing indices,
as used in this applied context. Also, the presence of a species or
its abundance in a certain habitat does not only depend on its
mobility. Other relevant factors may include abundance of pre-
ferred prey items (e.g. Gonsalves et al., 2013) or climatic prefer-
ences (Estrada-Villegas et al., 2012).
4.3. Conclusions and Implications for conservation
Our study underlies the importance of explicitly modeling
detectability when studying the occurrence or distribution of bats
with acoustic methods, concurring with results from previous stud-
ies (Duchamp et al., 2006; Kéry, 2011; MacKenzie et al., 2006). The
great variationwe found in detectabilitywas not limited to different
AIB species but was also found across different habitat types within
the same species. This finding is of crucial importance when com-
paring different habitat types. Conservation studies ignoring this
variation risk either over- or underestimating species’ occurrence
software (Version 1.4). The Cornell Lab of Ornithology, Ithaca, NY, USA.
Boesch, R., Obrist, M.K., 2013. BatScope – Implementation of a Bioacoustic Taxon
Practical Information-Theoretic Approach. Springer-Verlag, New York.
Castro-Arellano, I., Presleya, S.J., Saldanha, L.N., Willig, M.R., Wunderle, J.M., 2007.
nserin one habitat type relative to another, potentially leading to erro-
neous conclusions.
We found habitat-specific occupancies of AIB species. Some
species, such as C. centralis, M. molossus, Cynomops sp. and N.
albiventris, clearly preferred or avoided certain habitat types. In
general, our findings, covering a very large geographic scale, con-
firm habitat preferences identified on geographically more
restricted scales (Estrada-Villegas et al., 2012; Jung and Kalko,
2010, 2011). In order to sustain a species-rich AIB community,
which will provide crucial ecosystem functions and services
including insect control on a regional scale, it is critical to secure
a diversity of different habitat types. For AIB, important habitat
types are not limited to pristine, intact forest but also include
anthropogenically altered habitats, where we show several species
increase their occupancy.
By studying the habitat-specific occupancy values of AIB species,
we identified a general pattern linking a species’ mobility and its
susceptibility to habitat change. The broader and larger-surfaced
thewings, the less mobile the species and themore likely it will dis-
appear with deforestation and urbanization. Bats with small-sur-
faced and narrow wings are more mobile and can profit from
anthropogenic habitat changes. Based on our quantitative results
on wing shape, we can go beyond these qualitative statements and
predict the vulnerability to habitat change even for species not
included in the analysis. As long as its wing shape measures are
known, one can predict the degree to which a species will be
impacted by deforestation and urbanization. Such a prediction is
especially valuable for species for which little or nothing is known
about ecology or natural history, e.g. newly described species.
There is evidence that in addition to wing morphology (Jones et al.,
2003), characteristics such as distribution range, habitat specializa-
tion and, to a certain degree, food preferences, predict extinction
risk in bats (Boyles and Storm, 2007). For 22% of all mammals,
there is no data on these predictive factors other than morphology;
small species such as bats are overrepresented in this fraction
(Gonzalez-Suarez et al., 2012). For AIB in the tropics, the absence
of such information is the norm (Kalko et al., 2008), thus alternative
approachesestimatingvulnerability, suchas theonepresentedhere,
may be critical to guide conservation efforts. Conservation agencies
could use mobility, as predicted from wing morphology, possibly
together with additional information, such as the extent of
occurrence (IUCN, 2012), to prioritize conservation actions. Our
study contributes to a growing body of literature investigating
how trait information can be used to assess species’ susceptibility
to habitat change and extinction risk (e.g. Jennings and Pocock,
2009; Meyer et al., 2008). The inclusion of the type and extent of
expected future habitat change, as well as additional traits, would
further improve the assessment of vulnerability and thus extinction
risk for many species, including the elusive AIB in the tropics.
Role of the funding source
STRI, University of Ulm, University of Zurich: Support of
fieldwork.
Swiss Academy of Natural Sciences SCNAT: Funding of travel
expenses.
Elekon, Lucerne, and Swiss Federal Institute for Forest, Snow
and Landscape Research WSL: Material support (recording
devices).
Swiss National Science Foundation (SNSF): Research project of
last author, in which this study was included.
Acknowledgements
E. Bader et al. / Biological CoWe are deeply grateful to the late Elisabeth Kalko, who was cru-
cial for the development of this study, but sadly could not witnessEffects of reduced impact logging on bat biodiversity in terra firme forest of
lowland Amazonia. Biol. Conserv. 138, 269–285.
Clare, E.L., Adams, A.M., Maya-Simoes, A.Z., Eger, J.L., Hebert, P.D.N., Fenton, M.B.,
2013. Diversification and reproductive isolation: cryptic species in the only
New World high-duty cycle bat, Pteronotus parnellii. BMC Evol. Biol. 13 (26), 1–
18.
Condit, R., Watts, K., Bohlman, S.A., Pérez, R., Foster, R.B., Hubbell, S.P., 2000.
Quantifying the deciduousness of tropical forest canopies under varying
climates. J. Veg. Sci. 11 (5), 649–658.
Delgado-Jaramillo, M., Machado, M., Garcia, F.J., Ochoa, J., 2011. Bats (Chiroptera:
Mammalia) from Yurubi National Park, Venezuela: taxonomic list and
community study. Rev. Biol. Trop. 59, 1757–1776.
Duchamp, J.E., Swihart, R.K., 2008. Shifts in bat community structure related to
evolved traits and features of human-altered landscapes. Landscape Ecol. 23,
849–860.
Duchamp, J.E., Yates, M., Muzika, R.M., Swihart, R.K., 2006. Estimating probabilities
of detection for bat echolocation calls: An application of the double-observerIdentification Tool. Swiss Federal Research Institute WSL, Birmensdorf,
Switzerland.
Boyles, J.G., Storm, J.J., 2007. The perils of picky eating: dietary breadth is related to
extinction risk in insectivorous bats. PLoS ONE 2 (7), e672.
Boyles, J.G., Cryan, P.M., McGracken, G.F., Kunz, T.H., 2011. Economic importance of
bats in agriculture. Science 332 (6025), 41–42.
Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: Aits realization. We thank Martin K. Obrist, Swiss Federal Research
Institute WSL, as well as Hans Gysin and Marco Gumprich from
Elekon AG, Lucerne for support with field material; the
Smithsonian Tropical Research Institute and Marco Tschapka,
University of Ulm for logistic and financial support; the Swiss
National Science Foundation SCNAT for financial support; Barbara
König, Ruedi Boesch, Sergio Estrada Villegas, Francisco Delgado,
Matt Miller and his team, Rafael and Julieta Samudio, as well as
Raúl Arias for varied support of this study; two anonymous review-
ers for improving earlier versions of this manuscript; and last but
not least, the numerous landowners who allowed us to conduct
fieldwork on their private grounds.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biocon.2015.02.
028.
References
Adams, A.M., Jantzen, M.K., Hamilton, R.M., Fenton, M.B., 2012. Do you hear what I
hear? Implications of detector selection for acoustic monitoring of bats.
Methods Ecol. Evol. 3, 992–998.
Ahumada, J.A., Silva, C.E.F., Gajapersad, K., Hallam, C., Hurtado, J., Martin, E.,
McWilliam, A., Mugerwa, B., O’Brien, T., Rovero, F., Sheil, D., Spironello, W.R.,
Winarni, N., Andelman, S.J., 2011. Community structure and diversity of tropical
forest mammals: data from a global camera trap network. Philos. Trans. Roy.
Soc. B-Biol. Sci. 366, 2703–2711.
Avila-Flores, R., Fenton, M.B., 2005. Use of spatial features by foraging insectivorous
bats in a large urban landscape. J. Mammal. 86, 1193–1204.
Baird, A.B., Marchan-Rivadeneira, M.R., Perez, M.G., Baker, R.J., 2012. Morphological
analysis and description of two new species of Rhogeessa (Chiroptera:
Vespertilionidae) from the Neotropics. Occas. Pap. Mus. Texas Tech Univ. 307,
1–25.
Baker, R.J., McDonough, M.M., Swier, V.J., Larsen, P.A., Carrera, J.P., Ammerman, L.K.,
2009. New species of bonneted bat, genus Eumops (Chiroptera: Molossidae) from
the lowlands of western Ecuador and Peru. Acta Chiropterol. 11, 1–13.
Barlow, J., Gardner, T.A., Araujo, I.S., Avila-Pires, T.C., Bonaldo, A.B., Costa, J.E.,
Esposito, M.C., Ferreira, L.V., Hawes, J., Hernandez, M.M., Hoogmoed, M.S., Leite,
R.N., Lo-Man-Hung, N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., Miranda-
Santos, R., Nunes-Gutjahr, A.L., Overal, W.L., Parry, L., Peters, S.L., Ribeiro-Junior,
M.A., da Silva, M.N.F., Motta, C.d.S., Peres, C.A., 2007. Quantifying the
biodiversity value of tropical primary, secondary, and plantation forests. Proc.
Natl. Acad. Sci. U.S.A. 104, 18555–18560.
Bioacoustics Research Program, 2011. Raven Pro: interactive sound analysis
vation 186 (2015) 97–106 105method. Wildl. Soc. Bull. 34, 408–412.
Dunn, R.R., 2004. Recovery of faunal communities during tropical forest
regeneration. Conserv. Biol. 18 (2), 302–309.
Kunz, T.H., de Torrez, E.B., Bauer, D., Lobova, T., Fleming, T.H., 2011. Ecosystem
services provided by bats. Year Ecol. Conserv. Biol. 1223, 1–38.
106 E. Bader et al. / Biological Conservation 186 (2015) 97–106Estrada, A., Coates-Estrada, R., 2002. Bats in continuous forest, forest fragments and
in an agricultural mosaic habitat-island at Los Tuxtlas, Mexico. Biol. Conserv.
103, 237–245.
Estrada-Villegas, S., Meyer, C.F.J., Kalko, E.K.V., 2010. Effects of tropical forest
fragmentation on aerial insectivorous bats in a land-bridge island system. Biol.
Conserv. 143, 597–608.
Estrada-Villegas, S., McGill, B.J., Kalko, E.K.V., 2012. Climate, habitat, and species
interactions at different scales determine the structure of a Neotropical bat
community. Ecology 93, 1183–1193.
Ethier, K., Fahrig, L., 2011. Positive effects of forest fragmentation, independent of
forest amount, on bat abundance in eastern Ontario, Canada. Landscape Ecol.
26, 865–876.
Findley, J.S., 1993. Bats: A Community Perspective. Cambridge University Press,
Cambridge, UK.
Findley, J.S., Studier, E.H., Wilson, D.E., 1972. Morphologic properties of bat wings. J.
Mammal. 53, 429–444.
Fischer, J., Stott, J., Law, B.S., Adams, M.D., Forrester, R.I., 2009. Designing effective
habitat studies: quantifying multiple sources of variability in bat activity. Acta
Chiropterol. 11, 127–137.
Foley, J.A., Asner, G.P., Costa, M.H., Coe, M.T., DeFries, R., Gibbs, H.K., Howard, E.A.,
Olson, S., Patz, J., Ramankutty, N., Snyder, P., 2007. Amazonia revealed: forest
degradation and loss of ecosystem goods and services in the Amazon Basin.
Front. Ecol. Environ. 5, 25–32.
Frey-Ehrenbold, A., Bontadina, F., Arlettaz, R., Obrist, M.K., 2013. Landscape
connectivity, habitat structure and activity of bat guilds in farmland-
dominated matrices. J. Appl. Ecol. 50, 252–261.
Garcia-Morales, R., Badano, E.I., Moreno, C.E., 2013. Response of Neotropical bat
assemblages to human land use. Conserv. Biol. 27, 1096–1106.
Gonsalves, L., Law, B., Webb, C., Monamy, V., 2013. Foraging ranges of insectivorous
bats shift relative to changes in mosquito abundance. PLoS ONE 8 (5), e64081.
Gonzalez-Suarez, M., Lucas, P.M., Revilla, E., 2012. Biases in comparative analyses of
extinction risk: mind the gap. J. Anim. Ecol. 81, 1211–1222.
Gorresen, P.M., Miles, A.C., Todd, C.M., Bonaccorso, F.J., Weller, T.J., 2008. Assessing
bat detectability and occupancy with multiple automated echolocation
detectors. J. Mammal. 89, 11–17.
Henle, K., Davies, K.F., Kleyer, M., Margules, C., Settele, J., 2004. Predictors of species
sensitivity to fragmentation. Biodivers. Conserv. 13, 207–251.
Hill, J.L., Curran, P.J., 2003. Area, shape and isolation of tropical forest fragments:
effects on tree species diversity and implications for conservation. J. Biogeogr.
30, 1391–1403.
Hines, J.E., 2006. PRESENCE, Software to estimate patch occupancy and related
parameters. USGS–PWRC, http://www.mbr-pwrc.usgs.gov/software/presence.
html.
Hurvich, C.M., Tsai, C.L., 1989. Regression and time-series model selection in small
samples. Biometrika 76, 297–307.
IBM, 2011. SPSS.
IUCN, 2012. IUCN red list categories and criteria: version 3.1. Second edition.
Jennings, N., Pocock, M.J.O., 2009. Relationships between sensitivity to agricultural
intensification and ecological traits of insectivorous mammals and arthropods.
Conserv. Biol. 23, 1195–1203.
Jones, K.E., Purvis, A., Gittleman, J.L., 2003. Biological correlates of extinction risk in
bats. Am. Nat. 161, 601–614.
Jones, G., Jacobs, D.S., Kunz, T.H., Willig, M.R., Racey, P.A., 2009. Carpe noctem: the
importance of bats as bioindicators. Endangered Species Res. 8, 93–115.
Jung, K., 2009. Effects of Urbanization on Neotropical Aerial Insectivorous Bats.
University of Ulm, Germany, Ulm.
Jung, K., Kalko, E.K.V., 2010. Where forest meets urbanization: foraging plasticity of
aerial insectivorous bats in an anthropogenically altered environment. J.
Mammal. 91, 144–153.
Jung, K., Kalko, E.K.V., 2011. Adaptability and vulnerability of high flying Neotropical
aerial insectivorous bats to urbanization. Divers. Distrib. 17, 262–274.
Jung, K., Kalko, E.K.V., von Helversen, O., 2007. Echolocation calls in Central
American emballonurid bats: signal design and call frequency alternation. J.
Zool. 272, 125–137.
Jung, K., Kaiser, S., Böhm, S., Nieschulze, J., Kalko, E.K.V., 2012. Moving in three
dimensions: effects of structural complexity on occurrence and activity of
insectivorous bats in managed forest stands. J. Appl. Ecol. 49 (2), 523–531.
Jung, K., Molinari, J., Kalko, E.K.V., 2014. Driving factors for the evolution of species-
specific echolocation call design in new world free-tailed bats (Molossidae).
PLoS ONE 9 (1), e85279.
Kalcounis, M.C., Hobson, K.A., Brigham, R.M., Hecker, K.R., 1999. Bat activity in the
boreal forest: importance of stand type and vertical strata. J. Mammal. 80, 673–
682.
Kalko, E.K.V., 1998. Organisation and diversity of tropical bat communities through
space and time. Zool.-Anal. Complex Syst. 101, 281–297.
Kalko, E.K.V., Villegas, S.E., Schmidt, M., Wegmann, M., Meyer, C.F.J., 2008. Flying
high – assessing the use of the aerosphere by bats. Integr. Comp. Biol. 48, 60–73.
Kéry, M., 2011. Towards the modelling of true species distributions. J. Biogeogr. 38
(4), 617–618.
Klein, A.M., Steffan-Dewenter, I., Buchori, D., Tscharntke, T., 2002. Effects of land-
use intensity in tropical agroforestry systems on coffee flower-visiting and trap-
nesting bees and wasps. Conserv. Biol. 16, 1003–1014.Maas, B., Clough, Y., Tscharntke, T., 2013. Bats and birds increase crop yield in
tropical agroforestry landscapes. Ecol. Lett. 16 (12), 1480–1487.
MacKenzie, D.I., Nichols, J.D., Royle, J.A., Pollock, K.H., Bailey, L.L., Hines, J.E., 2006.
Occupancy Estimation and Modeling. Elsevier-Academic, San Diego, California.
MacSwiney, M.C., Clarke, F.M., Racey, P.A., 2008. What you see is not what you get:
the role of ultrasonic detectors in increasing inventory completeness in
Neotropical bat assemblages. J. Appl. Ecol. 45, 1364–1371.
McCain, C.M., 2009. Vertebrate range sizes indicate that mountains may be ‘higher’
in the tropics. Ecol. Lett. 12, 550–560.
McCarthy, M.A., Moore, J.L., Morris, W.K., Parris, K.M., Garrard, G.E., Vesk, P.A.,
Rumpff, L., Giljohann, K.M., Camac, J.S., Bau, S.S., Friend, T., Harrison, B., Yue, B.,
2013. The influence of abundance on detectability. Oikos 122, 717–726.
McKinney, M.L., 2006. Urbanization as a major cause of biotic homogenization. Biol.
Conserv. 127, 247–260.
Meyer, C.F.J., Frund, J., Lizano, W.P., Kalko, E.K.V., 2008. Ecological correlates of
vulnerability to fragmentation in Neotropical bats. J. Appl. Ecol. 45, 381–391.
Meyer, C.F.J., Aguiar, L.M.S., Aguirre, L.F., Baumgarten, J., Clarke, F.M., Cosson, J.-F.,
Estrada Villegas, S., Fahr, J., Faria, D., Furey, N., Henry, M., Hodgkison, R., Jenkins,
R.K.B., Jung, K.G., Kingston, T., Kunz, T.H., MacSwiney Gonzalez, M.C., Moya, I.,
Patterson, B.D., Pons, J.-M., Racey, P.A., Rex, K., Sampaio, E.M., Solari, S., Stoner,
K.E., Voigt, C.C., von Staden, D., Weise, C.D., Kalko, E.K.V., 2011. Accounting for
detectability improves estimates of species richness in tropical bat surveys. J.
Appl. Ecol. 48, 777–787.
Mueller, J., Mehr, M., Blässler, C., Fenton, M.B., Hothorn, T., Pretzsch, H., Klemmt, H.-
J., Brandl, R., 2012. Aggregative response in bats: prey abundance versus
habitat. Oecologia 169 (3), 673–684.
Norberg, U.M., Rayner, J.M.V., 1987. Ecological morphology and flight in bats
(Mammalia; Chiroptera): wing adaptations, flight performance, foraging
strategy and echolocation. Philos. Trans. Roy. Soc. Lond. Ser. B Biol. Sci. 316,
335–427.
Obrist, M.K., Boesch, R., Flückiger, P.F., 2004. Variability in echolocation call design
of 26 Swiss bat species: consequences, limits and options for automated field
identification with a synergetic pattern recognition approach. Mammalia 68,
307–322.
Parsons, S., Szewczak, J.M., 2009. Detecting, recording, and analyzing the
vocalizations of bats. In: Kunz, T.H., Parsons, S. (Eds.), Ecological and
Behavioral Methods for the Study of Bats. Johns Hopkins University Press,
Baltimore, MD, USA, pp. 91–111.
Patriquin, K.J., Hogberg, L.K., Chruszcz, B.J., Barclay, R.M.R., 2003. The influence of
habitat structure on the ability to detect ultrasound using bat detectors. Wildl.
Soc. Bull. 31, 475–481.
Pickett, S.T.A., 1989. Space-for-time substitution as an alternative to long-term
studies. In: Likens, G.E. (Ed.), Long-Term Studies in Ecology: Approaches and
Alternatives. 2nd Cary Conference, Millbrook, New York, USA, May 13, 1987.
Springer-Verlag, New York, Inc New York, USA, pp. 110–135.
Rasband, W., 2012. ImageJ 1.46r. National Institutes of Health, USA.
Reid, F., 2009. A Field Guide to the Mammals of Central America and Southeast
Mexico. Oxford University Press, Oxford, UK.
Royle, J.A., Nichols, J.D., Kéry, M., 2005. Modelling occurrence and abundance of
species when detection is imperfect. Oikos 110, 353–359.
Rydell, J., Arita, H.T., Santos, M., Granados, J., 2002. Acoustic identification of
insectivorous bats (order Chiroptera) of Yucatan, Mexico. J. Zool. 257, 27–36.
Safi, K., Kerth, G., 2004. A comparative analysis of specialization and extinction risk
in temperate-zone bats. Conserv. Biol. 18, 1293–1303.
Sanborn, C.C., 1937. American bats of the subfamily Emballonurinae Chicago Field
Museum of Natural History. Zool. Ser. 20 (24), 321–354.
Sattler, T., Bontadina, F., Hirzel, A.H., Arlettaz, R., 2007. Ecological niche modelling of
two cryptic bat species calls for a reassessment of their conservation status. J.
Appl. Ecol. 44, 1188–1199.
Sattler, T., Borcard, D., Arlettaz, R., Bontadina, F., Legendre, P., Obrist, M.K., Moretti,
M., 2010. Spider, bee and bird communities in cities are shaped by
environmental control and high stochasticity. Ecology 91, 3343–3353.
Threlfall, C., Law, B., Penman, T., Banks, P.B., 2011. Ecological processes in urban
landscapes: mechanisms influencing the distribution and activity of
insectivorous bats. Ecography 34, 814–826.
Turner, I.M., 1996. Species loss in fragments of tropical rain forest: a review of the
evidence. J. Appl. Ecol. 33, 200–209.
Vaughan, T.A., 1966. Morphology and flight characteristics of molossid bats. J.
Mammal. 47 (2), 249–260.
Walters, C.L., Freeman, R., Collen, A., Dietz, C., Fenton, M.B., Jones, G., Obrist, M.K.,
Puechmaille, S.J., Sattler, T., Siemers, B.M., Parsons, S., Jones, K.E., 2012. A
continental-scale tool for acoustic identification of European bats. J. Appl. Ecol.
49, 1064–1074.
Williams-Guillen, K., Perfecto, I., 2010. Effects of agricultural intensification on the
assemblage of leaf-nosed bats (Phyllostomidae) in a coffee landscape in chiapas,
Mexico. Biotropica 42, 605–613.
Winkler, H., Leisler, B., 1992. On the ecomorphology of migrants. Ibis 134, 21–28.