Content uploaded by Pascual López-López
Author content
All content in this area was uploaded by Pascual López-López on Jan 18, 2018
Content may be subject to copyright.
Short communication
Snake species richness predicts breeding distribution of
short-toed snake eagle in central Italy
JACOPO G. CECERE
1,2,*
,MICHELE PANUCCIO
3
,ANDREA GHIURGHI
4
,
FERDINANDO URBANO
5
,SIMONA IMPERIO
6
,CLAUDIO CELADA
2
and PASCUAL LÓPEZ-LÓPEZ
7
1
Institute for Environmental Protection and Research (ISPRA), Via Ca’Fornacetta 9,
Ozzano dell’Emilia (Bologna), Italy
2
LIPU –Conservation Department, Via Udine 3/A, Parma, Italy
3
MEDRAPTORS (Mediterranean Raptor Migration Network), c/o Michele Panuccio, Via
Mario Fioretti 18, Rome, Italy
4
Freelance consultant, Via Tullio Passarelli 67, Rome, Italy
5
Freelance consultant, Via Nuoro 2, Milan, Italy
6
National Research Council of Italy (CNR) –Institute of Geosciences and Earth
Resources, Via G. Moruzzi 1, Pisa, Italy
7
University of Valencia, Cavanilles Institute of Biodiversity and Evolutionary Biology,
Terrestrial Vertebrates Group, Paterna, Valencia, Spain
Received 25 January 2017, accepted 27 March 2017
Birds of prey, as top predators, play a key role in ecosystem functioning by
regulating prey populations and, by means of cascade effects, promoting biodi-
versity. This makes them adequate sentinels of ecosystem health. Here we analyse
the relationship between the occurrence of breeding short-toed snake eagle
(Circaetus gallicus) and both the richness of potential prey species and landscape
characteristics by taking into account two different spatial scales (i.e. nest-site
scale and landscape scale). The short-toed snake eagle offers an interesting case
study for investigating the relationships between top predators, prey diversity, and
habitats, because it is an extremely specialised raptor that feeds on mesopreda-
tors, mostly snakes. Additionally, short-toed snake eagles are mainly threatened by
changes in agriculture and land use in Europe, which have reduced the extent of
suitable hunting habitats, and by the decrease in snake populations. Our study
was conducted in the Latium Region (central Italy) in 2007, where most of the
Italian breeding population is concentrated. By means of habitat selection ana-
lyses using generalised linear models, our results showed that the species selected
breeding areas characterised by low elevations, rugged slopes, and high snake
species richness at the nest-site scale (1 km
2
). At the landscape scale (25 km
2
), the
*
Corresponding author: Jacopo G. Cecere, Istituto Superiore per la Protezione e la Ricerca
Ambientale (ISPRA), Via Ca’Fornacetta 9, I-40064 Ozzano dell’Emilia (Bologna), Italy (E-mail:
jacopo.cecere@isprambiente.it).
Ethology Ecology & Evolution, 2018
Vol. 30, No. 2, 178–186, https://doi.org/10.1080/03949370.2017.1323800
© 2017 Dipartimento di Biologia, Università di Firenze, Italia
best model showed that birds selected areas characterised by lower elevations for
nesting, with a tendency towards intermediate values of wood cover and high
snake species richness. Our study highlights the strong relationship between
snake species richness and the occurrence of breeding eagles at both spatial
scales, with optimal breeding sites located closer to hunting areas than expected
by chance. This study provides further support for the role of short-toed snake
eagles as sentinel species for Mediterranean habitats, and highlights the link
between the location of nesting sites and the occurrence of human-modified
landscapes characterised by high prey richness.
KEY WORDS:bird of prey, elevation, habitat selection, mesopredator, raptors, wood
cover.
INTRODUCTION
Birds of prey occupy a top position in the food web of most terrestrial habitats,
which implies a high sensitivity of raptors to ecosystem dysfunction (Newton 1979).
Their vulnerability to environmental changes combined with their occurrence range
often being linked to high habitat quality and ecosystem productivity (Sergio &
Newton 2003) lead conservationists to consider raptors sentinel and flagship species
(Sergio et al. 2006,2008). As top predators, raptors exert substantial top-down effects
on lower trophic levels including regulating prey populations and, by means of cas-
cade effects, promoting biodiversity (Sergio et al. 2014).
The short-toed snake eagle (Circaetus gallicus) offers an interesting case study for
investigating the relationships between top predators, prey diversity, and habitats. In
fact, it specialises in feeding on mesopredators, mostly snakes (Amores & Franco 1981;
Bakaloudis et al. 1998), and requires heterogeneous landscapes with both open areas
for catching prey and forests for nesting (Sánchez-Zapata & Calvo 1999). In Europe,
the short-toed snake eagle is mainly threatened by changes in agriculture and land use,
which have reduced the extent of suitable hunting habitat, and by the reduction of
snake populations due to increased cultivation of monocultures, hedge destruction,
the use of pesticides, and the abandonment of traditional farmland and subsequent
afforestation (BirdLife International 2015).
In this study, we aim to analyse the link between the occurrence of breeding short-
toed snake eagle and both the richness of potential prey species and landscape char-
acteristics by taking into account two different spatial scales (1 and 25 km
2
). The study
focused on central Italy along the European–African flyway, where most of the Italian
breeding population is concentrated (Agostini & Mellone 2008; Panuccio et al. 2012).
METHODS
Study area and breeding site survey
The study is based on a large survey of breeding short-toed snake eagle carried out in the
Latium Region (Italy) in 2007, whose methods and results are reported in Speranza and Cecere
(2008), and which is partially reported by Ceccarelli and Ricci (2007). The survey identified
and localised 40 nesting sites, defined as very small areas inside which an active nest, or adult
short-toed snake eagles showing specific behaviours (e.g. adults carrying nesting material;
Habitat selection of short-toed snake eagle 179
copulation; displaying marked territoriality), has been observed, indicating a high probability
of breeding. The Latium Region (17,203 km
2
) is located in west-central Italy. Nearly one fourth
of the territory (23%) is constituted by flat areas below 100 m above sea level (asl), while 52%
is hilly (100–600 m asl), but both are mainly characterised by Mediterranean vegetation. The
remaining area (23%) is mountainous (over 600 m asl) with more continental vegetation (Blasi
1994). Woods and forests cover 28.6% of the region, while agricultural land use represents
49.7% of the territory (Blasi 1994).
Environmental variables
In addition to the 40 points of presence, we generated 40 random points (absence)
within the borders of the Latium Region, while avoiding urban areas and bodies of water,
using ESRI-ArcGIS 9.2. Around each point, we generated two square cells with different
sizes: 1 km
2
for the nest-site scale and 25 km
2
for the landscape-level scale (López-López
et al. 2006). All 160 cells (80 for each scale) were characterised by seven variables: (1) average
“slope”and (2) average “elevation”, obtained by a digital elevation model (DEM) 20 m raster
re-classified by a grid with 60 × 60 m pixel; random plots were coerced within the optimal
elevation range known for the species (0–1600 m asl; values over this range are considered
uncommon for the species, Cramp & Simmons 1980); (3) amount of “urban”and (4) “wood”
habitats, obtained by re-classifying the CORINE Land Cover map for 2006 (class 1 and 3.1,
respectively) by a grid with 60 × 60 m pixel; (5) squared-transformed wood cover, “wood
2
”,
was entered to account for possible non-linear relationships, e.g. selection (or avoidance) of
intermediate values; (6) “wood-edge”, calculated as the sum of perimeters of wood patches
was used as a proxy for ecotonal habitats; (7) predicted number of “snake”species obtained
from the Regional Ecological Network raster (RER –Rete Ecologica Regionale, Scalisi et al.
2011)downloadedfromhttp://dati.lazio.it/geoserver/reteecologicalazio/wms?service=
WMS&version=1.1.0&request=GetCapabilities. RER is a multilevel raster including species
distribution models (SDM) for all the vertebrate species occurring in the Latium Region.
Models for all nine snake species occurring in the region were validated by means of
independent data sets of species occurrence. Since it is based on SDMs, the variable
“snake”does not provide the actual number of snake species for a given spatial scale, but
rather the predicted number of species. Even considering this limit, we have to acknowledge
that RER is currently the best source of information on snake species occurrence at a
regional scale.
The amount of open area was indirectly included in the analyses, since it was highly
negatively correlated with wood cover at both 1 km
2
(r = –0.92) and 25 km
2
scales (r = –0.73).
Data analysis
Habitat selection was assessed using logistic generalised linear models (GLM) with the
presence/absence of each cell treated as a dependent variable and the seven environmental
variables treated as independent variables (see. e.g. López-López et al. 2006 for similar
methods). Akaike’s information criterion (AIC) was used to select the best models (ΔAIC ≤
2), which were then used to perform model averaging with their corresponding Akaike
weights (Burnham & Anderson 2002). To avoid problems with parameter estimations, we
first checked for pair-wise correlations between variables (Zuur et al. 2007). All analyses were
performed in R ver. 3.3.2 (R Core Team 2016) separately for 1 and 25 km
2
scales, using the
“MuMIn”package (Barton 2013) for model averaging. Final models were validated by means
of the Area Under the Receiver Operating Characteristic Curve (AUC) using the “pROC”
package for R (Robin et al. 2011).
180 J.G. Cecere et al.
RESULTS
None of the environmental variables were highly correlated to each other at
1km
2
, allowing us to include all of them in the GLMs. At the 25 km
2
scale, slope
and elevation were highly correlated (r = –0.72); we chose to enter elevation in the
models since it has been found to predict the presence of the short-toed snake eagle at
a larger scale (Panuccio et al. 2015). The best models (ΔAIC < 2) for both spatial scales
are shown in Table 1. The final model at a nest-site scale (1 km
2
) showed that the
species selected breeding areas characterised by lower elevations, higher slopes, and
higher snake species richness (Table 2;Fig. 1). At a landscape scale (25 km
2
), the final
Table 1.
Best logistic regression models (ΔAIC ≤2) at two spatial scales comparing occupied and a random
sample of available locations.
Scale Model AIC ΔAIC w
1
1km
2
Elevation + Slope + Snake 80.92 0.00 0.69
Elevation + Slope + Wood + Wood
2
+ Wood-edge + Urban + Snake 82.56 1.64 0.31
25 km
2
Elevation + Wood + Wood
2
+ Urban + Snake 87.40 0.00 0.43
Elevation + Wood + Wood
2
+ Wood-edge + Urban + Snake 88.09 0.69 0.31
Elevation + Wood + Wood
2
+ Urban 88.45 1.05 0.26
Table 2.
Coefficient estimates of the final model, obtained by averaging the best-performing models (shown in
Table 1) with the corresponding Akaike weights separately at the two different spatial scales. Significant
P-values (P< 0.05) are highlighted in bold.
Coefficient estimate z value P
1km
2
scale (Intercept) –5.41 ± 1.95 2.59 0.009
Elevation –0.004 ± 0.001 2.56 0.01
Slope 0.24 ± 0.07 3.38 < 0.001
Wood –0.08 ± 0.06 1.30 0.19
Wood
2
0.75 × 10
–3
± 0.61 × 10
–3
1.22 0.22
Wood-edge 0.03 ± 0.05 0.66 0.50
Urban –4.18 ± 0.04 0.01 0.99
Snake 0.63 ± 0.19 3.24 0.001
25 km
2
scale (Intercept) –4.43 ± 2.93 1.49 0.14
Elevation –0.003 ± 0.001 1.99 0.046
Wood 0.22 ± 0.11 1.96 0.05
Wood
2
–0.002 ± 0.001 1.80 0.07
Wood-edge –0.07 ± 0.07 1.11 0.27
Urban –0.17 ± 0.12 1.37 0.17
Snake 0.32 ± 0.18 1.72 0.09
Habitat selection of short-toed snake eagle 181
model showed that birds selected areas characterised by lower elevations for nesting
with a tendency towards intermediate values of wood cover (with a peak around 50%;
see Fig. 1) and snake species richness and had a smaller effect size compared to the
same effect at a nest-site scale (1 km
2
)(Table 2;Fig. 1). The model validation
performed by the AUC showed that final models predicted 89.2 and 85.6% of data,
respectively, for nest-sites and landscape scales.
DISCUSSION
The highlight of this study is the strong relationship between snake species
richness and the occurrence of breeding short-toed snake eagles at both spatial scales
(1 and 25 km
2
). The same link was found in other Mediterranean environments, such
as southeastern Spain and across Italy, but at a much larger scale (100 km
2
; Moreno-
Rueda & Pizarro 2007; Panuccio et al. 2015). The result that prey species richness can
explain nest-site selection at a very fine scale (1 km
2
) suggests that optimal breeding
sites are placed as close as possible to hunting areas, as has previously been shown for
another raptor species, the red kite (Milvus milvus, Pfeiffer & Meyburg 2015). High
numbers of prey species could favour the short-toed snake eagle, since different snake
species can be active at different times (Ernst et al. 2012; Rocha et al. 2014) and/or
select different microhabitats (Gomes & Almeida-Santos 2012), which increases the
probability of the predator contacting potential prey. The stronger relationship at a
small scale (1 km
2
) with respect to larger scales [both the 25 km
2
of the present study
and 100 km
2
reported in Moreno-Rueda and Pizarro (2007)] implies that nest-site
Fig. 1. —Relationship between the presence/absence of short-toed snake eagles and environmental
variables at a nest-site (1 km
2
, upper panels) and a landscape scale (25 km
2
, lower panels) in the Latium
Region, central Italy. Only significant and nearly significant relationships are shown (see Table 2).
Points of presence/absence are jittered on the Y-axis with random noise to enhance visualisation.
182 J.G. Cecere et al.
selection by the short-toed snake eagle is in accordance with prey occurrence. This
may suggest an ability for this raptor to assess habitat quality when establishing
breeding territories after returning from African wintering grounds; this is a useful
skill that has been observed in other bird species (Orians & Wittenberger 1991).
At the same time, considering that this bird mainly catches the most abundant
snake species with a preference for the largest individuals (Gil-Sánchez &
Pleguezuelos 2001), the short-toed snake eagle could favour snake richness through
a top-down regulation process. Thereby, eagles would regulate the population size of
the commonest species and reduce inter-specific competition among snakes as a
cascading effect (Moreno-Rueda & Pizarro 2007).
In addition to snake species richness, the presence of breeding pairs at a small
landscape scale was positively related to high slopes. Reliefs with deep sides are also
selected by the species breeding in the Alicante province in Spain (López-Iborra et al.
2011). The study area (Latium Region) is characterised by a high density of human
population, in particular in the areas surrounding Rome and in the flat areas. Therefore,
we can expect that short-toed snake eagles prefer rugged areas for breeding, where
human pressure is lower than elsewhere. Alternatively, eagles could select these rugged
sites in order to take advantage of rising thermal updrafts, which are used for soaring
and searching for food, as has also been hypothesised for juvenile Bonelli’seagles
(Hieraaetus fasciatus) selecting steeper slopes for their temporary settlements
(Balbontín 2005). Low elevations were preferred by short-toed snake eagles for breeding
(average elevation of occurrence: 424 m, SD: 237 m asl) at both small and large spatial
scales. The short-toed snake eagle arrives from wintering grounds in March, when
temperatures are still cold at higher elevations for finding reptiles. Moreover, considering
that elevation is strongly correlated with the average temperature during the breeding
season (r = –0.96; n = 40 actual nesting sites; average temperature in May–August from
WorldClim; Hijmans et al. 2005), we can conjecture that low elevations were selected due
to higher temperatures, which in turn can also favour snake activity. At a large scale,
breeders preferred areas characterised by intermediate values of wood cover and avoided
both heavily open and wooded landscapes. The species needs open areas for hunting
(Bakaloudis et al. 1998), but also woods for nesting (López-Iborra et al. 2011).
Interestingly, the amount of wood-edge, which is an indirect measure of habitat hetero-
geneity, was not a key factor explaining the presence of the species. However, other
studies have reported thatecotonal habitats are the preferred hunting areas for the short-
toed snake eagles (Sánchez-Zapata & Calvo 1999). The finer spatial scale of our analysis
in comparison with previous studies could probably account for this difference, since
landscape heterogeneity arises at larger spatial scales.
In conclusion, this study provides further support for the role of short-toed snake
eagles as sentinel species for Mediterranean habitats, and highlights the link between
the location of nesting sites and the occurrence of human-modified landscapes char-
acterised by high prey richness.
ACKNOWLEDGEMENTS
We thank Stefano Ricci, Valter Ceccarelli, Massimo Brunelli, Francesca Zintu, Fabio
Borlenghi, Alberto Sorace, Ferdinando Corbi, Luigi Corsetti, Silvano Roma, and Emiliano De
Santis for their great work during the short-toed snake eagle survey carried out in Latium in 2007
under the framework of the BirdMonitoring 2007 project, funded by Direzione Regionale
Habitat selection of short-toed snake eagle 183
Ambiente e Cooperazione tra i Popoli, Regione Lazio and managed by LIPU and the Regional
Agency of Parks (ARP Lazio). We thank Javier Balbontín and Miguel Ferrer for constructive
comments on a previous draft of the manuscript.
Pascual López-López is supported by a ‘Juan de la Cierva-incorporación’postdoctoral
grant of the Spanish Ministry of Economy and Competitiveness (Reference IJCI-2014-19190).
Simona Imperio is supported by the Project of Interest “NextData”(PNR 2011-2013). This study
also benefited from collaboration with the European Union’s Horizon 2020 project
ECOPOTENTIAL (No. 641762).
DISCLOSURE STATEMENT
No potential conflict of interest was reported by the authors.
FUNDING
This work was supported by the Direzione Regionale Ambiente e Cooperazione tra i Popoli,
Regione Lazio; LIPU; Regional Agency of Parks (ARP Lazio); a Juan de la Cierva-incorporación’
postdoctoral grant of the Spanish Ministry of Economy and Competitiveness [IJCI-2014-19190];
Project of Interest “NextData”[PNR 2011–2013]; and the European Union’s Horizon 2020 project
ECOPOTENTIAL [641762].
ORCID
Jacopo G. Cecere http://orcid.org/0000-0002-4925-2730
REFERENCES
Agostini N, Mellone U. 2008. Does migration flyway of short-toed snake-eagles breeding in
central Italy reflect the colonization history? J Raptor Res. 42:158–159.
Amores F, Franco A. 1981. Alimentation et écologie du Circaète Jean-le-Blanc dans le sud de
l’Espagne. Alauda. 49:59–64. French.
Bakaloudis DE, Vlachos CG, Holloway GJ. 1998. Habitat use by short-toed eagles Circaetus
gallicus and their reptilian prey during the breeding season in Dadia Forest (north-eastern
Greece). J Appl Ecol. 35:821–828.
Balbontín J. 2005. Identifying suitable habitat for dispersal in Bonelli’s eagle: an important issue
in halting its decline in Europe. Biol Conserv. 126:74–83.
Barton K. 2013. MuMIn: multi-model inference. R package version 1.9.13. Available from: http://
CRAN.R-project.org/package=MuMIn
BirdLife International. 2015.Circaetus gallicus. The IUCN Red List of Threatened Species 2015: e.
T22734216A80165019; [cited 2016 Dec 7]. Available from: 10.2305/IUCN.UK.2015-4.
RLTS.T22734216A80165019.en
Blasi C. 1994. Fitoclimatologia del Lazio [Phytoclimatology of Lazio]. Roma: Università La
Sapienza e Regione Lazio, Assessorato Agricultura-Foreste, Caccia e Pesca. Italian.
Burnham KP, Anderson DR. 2002. Model selection and multimodel inference: a practical
information-theoretic approach. 2nd ed. Berlin: Springer-Verlag.
Ceccarelli V, Ricci S. 2007. Monitoraggio delle coppie di Biancone Circaetus gallicus nella ZPS
“Comprensorio-Tolfetano-Cerite”. Alula. 14:3–7. Italian.
184 J.G. Cecere et al.
Cramp S, Simmons KEL. 1980. The birds of western palearctic. Vol II. London: Oxford
University Press.
Ernst CH, Orr JM, Creque TC, Laemmerzahl AF, Hartsell TD. 2012. Annual and daily activity
cycles of snakes in northern Virginia, USA. Herpetol Bull. 121:23–28.
Gil-Sánchez MJ, Pleguezuelos JM. 2001. Prey and prey-size selection by the short-toed eagle
(Circaetus gallicus) during the breeding season in Granada (south-eastern Spain). J Zool.
255:131–137.
Gomes CA, Almeida-Santos SM. 2012. Microhabitat use by species of the genera Bothrops and
Crotalus (Viperidae) in semi-extensive captivity. J Venom Anim Toxins Incl Trop Dis.
18:393–398.
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. 2005. Very high resolution inter-
polated climate surfaces for global land areas. Intern J Climat. 25:1965–1978.
López-Iborra GM, Limiñana R, Pavón D, Martínez-Pérez JE. 2011. Modelling the distribution
of short-toed eagle (Circaetus gallicus) in semi-arid Mediterranean landscape: identifying
explanatory variables and their implications for its conservation. Eur J Wildl Res. 57:83–
93.
López-López P, García-Ripollés C, Aguilar JM, García-López F, Verdejo J. 2006. Modelling
breeding habitat preferences of Bonelli´s eagle (Hieraaetus fasciatus) in relation to topo-
graphy, disturbance, climate and land use at different spatial scales. J Ornithol. 147:97–
107.
Moreno-Rueda G, Pizarro M. 2007. Snake species richness and shrubland correlate with the
short-toed eagle (Circaetus gallicus) distribution in south-eastern Spain. Ann Zool Fennici.
44:314–320.
Newton I. 1979. Population ecology of raptors. London: T. & A.D. Poyser.
Orians GH, Wittenberger JF. 1991. Spatial and temporal scales in habitat selection. Am Nat.
137:S29–S49.
PanuccioM,AgostiniN,PremudaG.2012. Ecological barriers promote risk minimization and
social learning in migrating short-toed snake eagles. Ethol Ecol Evol. 24:74–80.
Panuccio M, Lucia G, Agostini N, Ottonello D, Bogliani G. 2015. Motion capacity, geography
and ecological features explain the present distribution of a migratory top predator. Ecol
Res. 30:181–190.
Pfeiffer T, Meyburg BU. 2015. GPS tracking of Red Kites (Milvus milvus) reveals fledgling
number is negatively correlated with home range size. J Ornithol. 156:963–975.
R Core Team. 2016. R: A language and environment for statistical computing. Vienna, Austria: R
Foundation for Statistical Computing. Available from: https://www.R-project.org/
Robin X, Turck N,Hainard A, Tiberti N,Lisacek F, Sanchez JC, Müller M.2011.pROC:anopen-
source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics. 12:77.
Rocha MC, Hartmann PA, Winck GR, Cechin SZ. 2014. Seasonal, daily activity, and habitat
use by three sympatric pit vipers (Serpentes, Viperidae) from southern Brazil. An Acad
Bras Cienc. 86:695–706.
Sánchez-Zapata JA, Calvo JF. 1999. Raptor distribution in relation to landscape composition in
semi-arid Mediterranean habitats. J Appl Ecol. 36:254–262.
Scalisi M, Capizzi D, Monaco A, Pizzol I, Sinibaldi I. 2011. Rete Ecologica Regionale
(REcoRd_Lazio): approccio metodologico e primi risultati. In: Bruschi M, Scalisi M,
editors. Verso un piano per il sistema delle aree protette del Lazio. Roma: Edizioni ARP;
p. 55–102. Italian.
Sergio F, Caro T, Brown D, Clucas B, Hunter J, Ketchum J, McHugh K, Hiraldo F. 2008. Top
predators as conservation tools: ecological rationale, assumptions, and efficacy. Annu Rev
Ecol Evol Syst. 39:1–19.
Sergio F, Newton I. 2003. Occupancy as a measure of territory quality. J Anim Ecol. 72:857–865.
Sergio F, Newton I, Marchesi L, Pedrini P. 2006. Ecologically justified charisma: preservation
of top predators delivers biodiversity conservation. J Appl Ecol. 43:1049–1055.
Habitat selection of short-toed snake eagle 185
Sergio F, Schmitz OJ, Krebs CJ, Holt RD, Heithaus MR, Wirsing AJ, Ripple WJ, Ritchie E,
Ainley D, Oro D, et al. 2014. Towards a cohesive, holistic view of top predation: a
definition, synthesis and perspective. Oikos. 123:1234–1243.
Speranza L, Cecere JG. 2008.L’importanza degli ambienti rurali per gli uccelli e il progetto
BirdMonitoring. Roma: LIPU –BirdLife International partner. Italian.
Zuur AF, Ieno EN, Smith GM. 2007. Analysing ecological data. Berlin: Springer Science &
Business Media.
186 J.G. Cecere et al.