Assessing reintroduction outcome: comparison of
the juvenile post-edging dependence period
between wild and reintroduced Bonelli’s eagles in
two Mediterranean islands
University of Valencia
PASCUAL LÓPEZ-LÓPEZ ( Pascual.Lopez@uv.es )
University of Valencia
Grupo de Rehabilitación de la Fauna Autóctona y su Hábitat (GREFA)
Ecologia Applicata Italia, Termini Imerese (PA)
Grupo de Rehabilitación de la Fauna Autóctona y su Hábitat (GREFA)
JUAN JOSÉ IGLESIAS-LEBRIJA
Grupo de Rehabilitación de la Fauna Autóctona y su Hábitat (GREFA)
MARIO LO VALVO
University of Palermo
Grupo de Rehabilitación de la Fauna Autóctona y su Hábitat (GREFA)
Ecologia Applicata Italia, Termini Imerese (PA)
LIFE Bonelli. Govern de les Illes Balears. Santa Eugènia. Mallorca. Balearic Islands
MASSIMILIANO DI VITTORIO
Ecologia Applicata Italia, Termini Imerese (PA)
Keywords: island environment, movement ecology, post-edging dependence period, raptors,
Posted Date: April 11th, 2022
Islands are key areas for biodiversity; however, they are extremely sensitive to anthropic actions. This has
led to local species extinctions, especially large predators such as raptors. Consequently, reintroduction
and conservation projects aimed at reversing population decline of endangered species have recently
gained popularity. Nevertheless, their relatively elevated cost and chance of failure make them
controversial, hence assessing their effectiveness is essential. Within the early stages of raptors, the post-
edging dependency period (PFDP) is the one in which individuals must face dangers without having
completely developed their skills. Thereby, comparing PFDP patterns concerning reintroduced and wild
individuals is of major interest as it would help to plan and improve future conservation actions. We
analyzed the behavior of 38 juvenile Bonelli’s eagles (15 reintroduced and 23 wild) tracked through GPS
telemetry, tagged as nestlings in two insular environments. The study period encompassed a total of
nine-year movement data from reintroduced chicks in Mallorca (Spain) and wild chicks from Sicily (Italy).
Movement parameters (i.e., age of rst ight, age of dispersal, length of the PFDP, revisits to the natal or
release area, and residence time in them) were analyzed together with their behavior during the PFDP for
reintroduced and wild individuals. Similar movement patterns were obtained for both origins, although
wild individuals revisited the natal site more often and dispersed earlier. Behavior was also similar, it
varied throughout the PFDP, observing a more abrupt progress in wild individuals and an earlier
development of travelling and hunting behaviors. Observed differences are probably related to food
availability, which can improve body condition and thus delay onset of dispersal, together with parental
presence, which can prompt an earlier ending of the PFDP by encouraging juvenile independence.
Islands are key areas for biodiversity and for this reason many of them were included in the 25 Global
Biodiversity Hotspots (Myers et al. 2000). The study of island environments has traditionally led to great
advances within the eld of ecology and conservation biology (MacArthur and Wilson, 1963; Diamond
1975) and this has led to consider islands as natural laboratories (Vitousek et al 2002). However, islands
are not exempt from human activities, indeed they have been experiencing them since prehistory (Duncan
et al. 2013) and more than 60% of recorded extinctions were insular endemic species (Gates and Donald
2000; Tershy et al. 2015). Due to their top position in food webs, large predators have experienced rapid
population declines worldwide (Peltonen and Hanski 1991; Rosenzweig and Clark 1994). Consequently,
different conservation plans including large predators like raptors have been developed to date (O’Rourke
2014; Dzialak et al 2007). These plans range from population monitoring and environmental
management in order to improve local conditions for particular species or ecosystems, to reintroduction
projects of endangered species, among others. In fact, overall extinction risk is greater for species that
inhabit insular environments than continents (Hoffmann et al. 2010), and hence several conservation
projects are being addressed to halt the decrease in population numbers of birds that have entered on the
brink of extinction. One of the most popular actions of these projects is to reintroduce individuals in those
areas where species went extinct in the past.
Reintroduction and relocation are specic methods widely used in conservation programs, and the study
of their biological signicance has been increasing dramatically since the 1990s (Seddon et al. 2007;
Efrat et al. 2020). Among raptors, successful reintroductions in islands include species such as the
Peregrine falcon (
) (Tordoff and Redig 2001), the Mauritius kestrel (
(Nicoll et al. 2004) and Seychelles’ Kestrel (
) (Watson 1981; Ferrer et al. 2019). However,
reintroduction projects are often controversial, mainly due to their potential high economic cost (Ferrer et
al. 2017) and their possible lack of success (Fischer and Lindermayer 2000), which is often perceived in
detriment of conservation budget for wildlife in some areas (Seddon et al. 2007; Armstrong and Seddon
2008) instead of considering the sum of this and other conservation techniques a way of adding
synergetic measures up.
One of the most crucial stages for birds of prey when colonizing new territories or settling in existing ones
is the juvenile dispersal period or natal dispersal (Bowler and Benton 2005), which can be dened as the
movements performed by juveniles of long-lived species, from their natal site to their rst breeding site
(Howard 1960, Greenwood 1980, Geeenwod and Harvey 1982). In many cases, the departure from the
natal site is preceded by exploratory movements during the phase known as the post-edging juvenile
dependence period (hereafter PFDP), which spans from the rst ight from the nest to the onset of
dispersal (Ferrer 1993; Balbontín and Ferrer 1993; Ramos et al. 2019). During this time, individuals
usually perform exploratory ights of different lengths and eventually return to their natal area until the
actual dispersal starts (Kenward et al. 1993; Walls and Kenward 1995; Cadahía et al. 2008; Weston et al.
2013). The PFDP is thus a critical stage of the early life of juveniles, where they must face natural and
anthropic mortality risks in parallel to the development of ying, hunting and social skills (Weathers and
Sullivan 1989; Ferrer 1992).
Global population decline of raptors (McClure et al. 2018; McClure and Rolek 2020) has led to the
implementation of conservation programs worldwide. In Europe, European Union-funded projects
addressed specically to the conservation of endangered raptors throughout different European
countries, including some of their islands, in Portugal, Spain, Italy, Cyprus, and Greece. Although the
juvenile PFDP and onset of dispersal have been studied in raptors by means of radio-tracking and
satellite telemetry (e.g., Cadahía et al. 2007b; Monti et al. 2012; López-López et al. 2014; Rahman et al.
2015) there is little knowledge on this stage of the life cycle of reintroduced individuals in comparison to
wild individuals, especially in insular environments. This information is crucial for reintroduction projects
in which much of the captive breeding efforts can be spoilt in the early stages of this process (Robert et
al. 2015). To evaluate the effectiveness of these conservation measures, monitoring of different
populations in which conservation actions are being performed is necessary (IUCN – SSC Species
Conservation Planning Sub-Committee 2017). Nowadays, with the development of remote tracking
technologies, the use of dataloggers allows not only accurate wildlife tracking for assessing conservation
plans but also broadens the understanding of animal ecology in terms of movement ecology and animal
behavior (Kays et al. 2015). Cutting-edge technologies, such as telemetry, provide researchers with
massive amounts of precise locations, enabling the study of movement patterns and the interactions
with other individuals or the environment. This technology can report crucial information of an
individual’s location, accelerometry and physiological parameters (López-López 2016) throughout
different stages of its life cycle. Dataloggers have been widely used to study raptors’ dispersal (Cadahía
et al. 2007b, Cadahía et al. 2010; Soutullo et al 2006) and, to a lesser extent, PFDP (López-López et al.
2014; Krüger and Amar 2017).
Here we analyzed the juvenile PFDP in a threatened raptor, the Bonelli’s eagle (
), in two
different islands of the Mediterranean Basin in relation to the origin of the individuals: reintroduced birds
in Mallorca (Balearic Islands, Spain) in comparison to wild birds in Sicily (Italy). Our study is thus aimed
at evaluating whether the individual’s origin or sex has an effect on the natural behavior during the PFDP
and onset of dispersal of the species, all of it from an island environment perspective by means of
GPS/GSM telemetry. We examined this period by establishing different parameters at which individuals
performed essential movements in addition to the different behaviors and their variation throughout the
whole PFDP. If pre-release and post-release actions were successful, we would expect wild and
reintroduced individuals to accomplish the different parameters at a similar age, except for the onset of
dispersal, which presumably might take place later in captive-reared individuals as they are fed ad-libitum
and, ultimately, do not account for parental presence to modify their behavior.
Materials And Methods
The Bonelli’s eagle is a medium-sized long-lived raptor, unevenly distributed in the Palearctic, Afrotropical
and Indomalayan regions. This eagle has experienced a considerable decline during the last decades,
especially in Europe (BirdLife International 2015) where it shows a circum-Mediterranean distribution.
These circumstances led to the Bonelli’s eagle being considered by the IUCN as “nearly threatened” in
Europe (BirdLife International 2015). Bonelli’s eagles usually breed between January and June nesting in
cliffs and trees. On average (± standard deviation), juveniles perform their rst ight at the age of 63 ± 6
days and begin dispersal at an age of 163 ± 17 days, with a PFDP spanning 98 ± 18 days (Real et al.
This study took place in two Mediterranean islands: Sicily (Italy) and Mallorca (Spain), where European-
funded Bonelli’s eagle conservation actions are being performed. Through the second half of the last
century, Bonelli’s eagle populations in these islands decreased principally because of habitat degradation,
direct persecution by poachers, poisoning, and electrocution on power lines(Real and Mañosa 1997; Real
et al. 2001; BirdLife International 2015; Hernández-Matías et al. 2015; Iglesias et al. 2018). It even became
locally extinct in Mallorca during the late 60s (Viada et al. 2015) and the Italian population diminished in
such a way that they currently only breed in Sicily, becoming extinct from Sardinia (López-López et al.
2012; Di Vittorio et al. 2012). To counteract this situation, the European Union funded two projects: the
“LIFE Bonelli” (LIFE12 NAT/ES/000701) for the conservation of Bonelli’s eagle in Mallorca and the Iberian
Peninsula, and the “LIFE ConRaSi” (LIFE14 NAT/IT/001017) for the conservation of Bonelli’s eagle, along
with Egyptian Vulture (
) and Lanner Falcon (
) inSicily. Both
islands are relatively large and close to mainland. Mallorca has an area of 3640 km2 and it is 220 km
apart from the Iberian Peninsula, whereas Sicily has a total extent of 25711 km2 and is isolated from the
Italic peninsula by the Strait of Messina which is 3 km wide at its narrowest point. These two islands
have areas with an abrupt orography that provides cliffs and forests, which constitute a suitable
environment for the establishment of new breeding territories (Cramp and Simmons 1980). They also
contain scrubland areas and non-irrigated agricultural areas, which favor the presence of potential prey,
including rabbits, pigeons and lizards (Di Vittorio et al. 2012; Viada et al 2015).
Overall, 44 individuals were tagged from 2011 to 2019; 27 wild chicks were tagged in Sicily from 2017 to
2019, whereas 17 reintroduced juveniles were tagged in Mallorca from 2011 to 2016. Wild chicks were
marked at the nest at the age of 45–50 days approximately following standard procedures (e.g., Cadahía
et al. 2007a). Reintroduced individuals were either bred in captivity in different centers in Spain and
France (GREFA, Vendée and Ardèche dependents of UFCS/LPO) (n = 11) or extracted from natural nests
from different localities in Andalusia (Spain) (n = 9). When aged 45–55 days, reintroduced individuals
were tagged and relocated to Mallorca. They were released through hacking cage and fed
avoiding contact with humans once the juveniles were inside the hacking. The hacking cage consisted of
a big cage with an articial nest in the inside together with perches. Juveniles spend the rst days
enclosed inside this articial nest to protect them from predators. After 7–10 days the nest is opened, and
juveniles can move around the cage. Inside this hacking cage juveniles were fed
prey, when juveniles were around 75 days live prey was provided. Once all the juveniles were able to hunt
and were in an adequate physical condition, hacking cage was opened for their release (Viada and
Iglesias 2017). Hacking cages were located along the Serra de Tramuntana, in the northwest of Mallorca
Island. Around the hacking area, juveniles were provided with platforms for external feeding, which was
supplied until none of the released birds further visited the hacking area. Wild individuals were raised by
their parents and were all tagged with solar-powered E-OBS dataloggers. Diet studies conducted in Sicily
during the breeding season by collecting pellets from the nest showed that Bonelli’s eagles fed mainly on
pigeon and wild rabbit (Di Vittorio et al. 2018).
Juveniles were tagged with GPS/GSM transmitters. Reintroduced individuals were tagged with solar-
powered dataloggers manufactured by Ecotone 30 g (n = 3), E-OBS 48 g (n = 6), NorthStar 45 g Geotrack
s/n GC106 (n = 1) and Microwave PTT-100 45 g GPS/Argos solar (n = 7). Wild individuals were raised by
their parents and were all tagged with solar-powered E-OBS dataloggers. Transmitters were mounted in a
backpack conguration designed to ensure that the harness would fall off at the end of the tag’s life
(Garcelon 1985), which is approximately 3 years, as it varies depending on the sampling rate and battery
size (Rempel and Rogers 1997). Although it has been proven that satellite backpack tagging of raptors
has no effect on their ecology and behaviour (Sergio et al. 2015), Gracelon’s harnessing method was
selected because of its demonstrated lack of effect on soaring raptors (Garcia et al. 2021). The mass of
the equipment, including the harness, metal ring and tag, was below 3% of the bird’s body mass, which is
within recommended limits (Kenward 2000). Juveniles’ sex was determined by morphometrics and
genetic determination following Palma et al. (2001).
Given that transmitters were initially programmed to record locations at different time intervals depending
on battery levels, we subsampled locations at ve minutes intervals to homogenize the overall dataset.
Data were retrieved, stored and downloaded from the online data repository Movebank
("http://www.movebank.org"). Only birds that completed the entire PFDP were considered for this study.
PFDP parameters and spatial analysis
To characterize the major events of the PFDP we calculated the following parameters: i) age of rst ight;
ii) age of dispersal; iii) length of the PFDP; iv) residence time; and v) number of revisits to the nest or
hacking site. The age of rst ight was calculated by means of the “recurse” R package (Bracis et al.
2018). This package allows establishing a central point and a circular area around it for which it counts
the number of visits and time spent inside and outside the area. Thereby, we set the central point as the
coordinates of the nest or hacking location. The radius was set to 50 m, which accounts for twice the
nominal error of the GPS. As a result, we obtained the number of revisits and residence time around their
nest or release site (i.e., nest or hacking). First ight day was set as the rst time that juveniles registered
a revisit, being recorded more than 50 m away for more than 30 minutes apart from the nest or hacking
site, as long as it was followed by regular revisits through the following days, together with visual
inspection of telemetry information.
To assess the onset of dispersal we calculated the Net Squared Displacement (NSD), which measures the
squared distance between the origin and further locations to gauge movements patterns (Bunneeld et al.
2011; Cagnacci et al. 2016). Here, NSD was measured from the nest/hacking site location to all
subsequent locations to obtain the day at which juveniles started dispersal. The onset of dispersal was
distinguished by an abrupt departure in the NSD which was not followed by any return to the natal area in
the following month (Supplementary Material S1: Fig. S1 – S2). The total length of the PFDP was thus
calculated as the difference between the onset of dispersal and the date of the rst ight. PFDP
parameters were compared by means of Wilcoxon rank sum tests either regarding juveniles’ origin or sex.
Provided that we performed multiple comparisons, we adjusted the p-values using the Benjamini –
Hochberg method implemented in the “p.adjust” function of the “Stats” R package (Benjamini and
We analyzed temporal variations in behavior by classifying individual’s behavior through an unsupervised
approach to multivariate data clustering, the Expectation Maximization binary Clustering (EMbC) (Garriga
et al. 2016). This algorithm allows annotation of behaviours through correlation as it considers the
velocity and the turning angle of the individual between consecutive GPS locations recorded at regular
intervals. Local estimates of velocity and turning angle establish segmentation limits, allowing
categorization of behavior into four groups: high velocity-high turning angle (HH), high velocity-low
turning angle (HL), low velocity-high turning angle (LH) and low velocity-low turning angle (LL). These
four classes were classied into behaviors: HH as active hunting, HL as travelling/relocating, LH as
searching/foraging and LL as resting (for further details on this method see Garriga et al. 2016; review in
Bennison et al. 2018).
We performed Linear Mixed Models (LMM) (Zuur et al. 2009) to assess differences in movement
parameters in relation to sex, origin, and age. These parameters were the age of rst ight, age of
dispersal, length of the PFDP, residence time and number of revisits. In each LMM we settled as
dependent variable each of the parameters, independent variables were “origin” and “sex”, and “year” was
established as a random factor. EMbC results were also analyzed with LMM. In this case, the dependent
variable was each of the movements obtained (i.e., hunting, travelling, foraging, and resting), “origin”,
“sex” and “week” (i.e., week after edging) were considered as independent variables, and “year” as
random factor. We compared and ranked the different models using the Akaike’s Information Criterion
(AIC) (Akaike 1973). The model with the lowest AIC value was selected as the best one. When two or more
models were considered as valid (i.e., AIC values differed in less than two AICc units), modelling average
was conducted to evaluate the contribution of each independent factor by means of the MuMin R
package (Barton 2020). All analyses were performed in R Statistic software version 4.0.2 (R Core Team
2020). Statistical signicance was set al α < 0.05.
From the 44 birds initially tagged, a total of 38 individuals were considered in this study: 15 from
Mallorca and 23 from Sicily. Individuals which did not complete the dependence period or whose track
was lost during this time span were not included. A total of 846 087 GPS locations were recorded and
analyzed in this study.
Overall, wild individuals tended to develop sooner than reintroduced ones (Fig. 1). The rst ight (Fig. 1a)
in wild individuals took place at an age of 64 ± 6 days (range 56 – 75 days), while reintroduced
individuals ew at 68 ± 8 days (range 57 – 90 days), although no signicant differences were found (U =
= 0.146). The onset of dispersal (Fig. 1b) occurred at a signicantly earlier age in wild birds
(158 ± 18 days; range 125 – 211 days) than in reintroduced individuals (168 ± 22 days; range 137 – 223
days) (U = 252.50,
= 0.018). Thus, the whole dependence period (Fig. 1c), from rst ight to dispersal,
was shorter for wild birds (90 ± 20 days; range 54 – 151 days) than for reintroduced ones (100 ± 25 days;
range 62 – 158 days) although no signicant differences were found (U = 22,
With regards to residence time (Fig. 1d), wild individuals spent less time in the nest area (7.43 ± 8.18
days; range 0.003 – 28.39 days) than reintroduced ones did in the hacking site (13.47 ± 12.04 days;
range 0.03 – 34.34 days) although no signicant differences were found (U = 222,
= 0.144). In contrast,
the number of visits to the hacking site (Fig. 1e) for reintroduced individuals (17 ± 27 revisits; range 1 –
98 revisits) were signicantly fewer than those of wild individuals to the nest (60 ± 50 revisits; range 1 –
187 revisits) (U = 70,
Pooling wild and reintroduced birds, no differences between sexes were found for any of the analyzed
> 0.05). Age of rst ight was 66 ± 9 days (range 56 – 90 days) and 67 ± 5 days (range
57 – 75 days) for males and females, respectively (U = 211,
= 0.346). Males began dispersal at a
slightly earlier age (157 ± 20 days; range 125 – 211 days) than females (163 ± 21 days; range 126-223
days) (U = 220,
= 0.228), resulting in a total dependence period of 91 ± 23 days for males (range 62 –
151) and 97 ± 21 days for females (range 54 – 158) (U = 218,
= 0.25). The residence time was 9.81 ±
10.94 days (range 0.003 – 34.34 days) and 9.82 ± 9.49 days (range 0.06 – 29.48 days) for males and
females, respectively, with no signicant differences observed (U = 194,
= 0.663). Similarly, no
differences in the number of visits to the nest or hacking site were observed for males (49 ± 50 revisits;
range 1 – 187 revisits) and females (43 ± 45 revisits; range 1 – 151 revisits) (U = 192,
= 0.702) (Fig. 1).
The best models differed for each parameter (Table 1). The age of rst ight and the length of the PFDP
were not affected neither by sex, origin, or their combination. The number of revisits were accounted for
the origin, the additive effect of sex and origin, and their interaction. After model averaging of the best
LMMs, origin inuenced age of dispersal (Z = 2.158;
= 0.003) and number of revisits (Z = 2.641;
0.008). Finally, all factors were included as predictors of residence time in all models.
When analyzing the movements from the different individuals, we observed that searching/foraging
behavior decreased over time as travelling and active hunting increased (Fig. 2). LMM analyses showed
that the differences in the share of time spent on each behavior varied as the PFDP progressed, mainly
for hunting, traveling, and foraging (Table 2), although it did not have an effect on the resting behavior.
Nevertheless, these transitions in the percentage of time invested in each behavior tent to be earlier and
more abrupt in wild individuals than in reintroduced ones, especially for the travelling behavior (Fig. 2). In
fact, at the end of the PFDP, at week 14, half of the individuals had already begun dispersal.
Regarding origin (Fig. 3), we observed that throughout the whole juvenile PFDP, most of the time was
invested in searching/foraging and resting activities; searching/foraging accounted for 47.93% of the
time for wild and a 49.62 % of the time for reintroduced birds. Wild individuals spent 11.57% of the time
in active hunting activities while reintroduced ones invested 6.14%. Concerning movements of travelling
and resting, they represented only 9.51% and 30.99% of the time, respectively, for wild individuals. In
comparison, reintroduced individuals invested more time in resting, accounting for 37.90% of the time,
and slightly less in travelling, which represented the 6.34% (Fig. 3).
New advances in wildlife tracking have allowed obtaining a large number of records with great precision
that would have never been recorded with traditional methods such as ringing or radio-tracking. However,
one of the key challenges of studying wildlife with GPS/GSM telemetry is the lack of observational data
that endows recorded locations with biological sense. To ll this gap, new tools that allow the
categorization of movements in an automatized and objective way have arisen, enabling researchers to
analyze not only ecological parameters but also behaviors. Thanks to the development of devices that
continuously track individuals and analytical tools that allow the study of movement data, here we have
thoroughly analyzed the PFDP and given biological signicance to individuals’ movements.
Our results showed that the PFDP resulted in a similar time span for wild and reintroduced individuals,
with both performing the rst ight at a similar age. This rst ight can only take place once juveniles are
fully feathered, which occurs at an age of approximately 60 days (Gil-Sánchez 2000), and are in an
appropriate physical condition. Our results agree with the ones obtained from the study conducted on
Montagu’s harrier (
), in which reintroduced and wild juveniles showed no differences
during the PFDP and similar behavior (Amar et al. 2000).
Wild individuals depend on their parents during the juvenile PFDP for food provisioning, however, towards
the end of the PDFP, raptor parents tend to decrease investment in their offspring, reducing the amount of
prey delivered to the juveniles (Ceballos and Donázar 1990; Arroyo et al. 2001). This reduction
encourages juveniles to develop hunting and ying skills and promotes their departure from their natal
area to avoid intraspecic competition (Trivers 1974). In fact, parental aggression towards the juveniles
has been reported for the Spanish imperial eagle (
) at the end of the PFDP (Alonso et al.
1986) and harpy eagle (
) (Urios et al. 2017). Reintroduced individuals do not account for
this parental presence, on the contrary, they are fed in order to keep them in the release site as long as
possible to favor their return and settlement there. In this situation, juvenile Bonelli’s eagles depart from
the natal site presumably when they acquire an optimal body condition as seen in other raptors (Ferrer
1992; Walls et al. 1999; Delgado et al. 2010). However, raptor juveniles tend to expand their PFDP and
remain in the parental territory to benet from their hunting areas and protection from other conspecics
(Weston et al. 2018). The fact that reintroduced individuals disperse later supports the idea that the onset
of dispersal is mostly driven by the individual, although parental presence can prompt it. Also, it is worth
noting that food availability does not seem to prevent dispersal departure from their natal territory, as
juveniles which were fed
eventually abandon their hacking site, although it could delay the
onset of dispersal as seen in Vergara et al. (2010), where food supplemented kestrels tended to acquire
independence later. Finally, Bonelli’s eagle, as other raptors, is bounded to climate traits, which affect not
only its distribution but also its breeding success (Ontivero and Pleguezuelos 2003; Carrascal and Seoane
2009; Di Vittorio and López-López 2014). Therefore, the warmer climate of Sicily could have favored an
earlier development of wild eagles.
The age at which individuals performed their rst ight, began dispersal and the total length of the PFDP
were similar to those obtained by Real et al. (1998) who found that rst ight occurred at an average age
of 63 ± 6 days after hatching (range 52–66 days), dispersal began at an average age of 163 ± 17 days
after hatching (range 143–177 days) and thus the whole PFDP lasted for 98 ± 18 days (range = 77–113
days, n = 5 in all cases). Although age of rst ight showed similar results between both populations, in
the case of reintroduced juveniles, they were not able to perform the rst ight until the hacking cages
were opened for them by the team, so probably there were some individuals that could have own for the
rst time earlier. However, rst ight is related to feather development, and it has been shown that
feeding of captive juveniles favors feather growth (Lacombe et al. 1994). Also, food
supplementation has been shown to allow juveniles to acquire a better physical condition and
homogenizes their development, minimizing differences among them (Muriel et al. 2015), so probably
body condition of hacked juveniles was similar when allowing them to leave the facility. In this way,
reintroduced juveniles’ age of rst ight was similar not only to the one obtained for the wild population
but also to that obtained in previous studies. The duration of the PFDP recorded here was longer than
that reported in Balbontín and Ferrer (2005) which lasted for 77 ± 19 days (range = 50–114 days, n = 28).
These differences were probably because they radio-tracked the chicks once a week and considered
dispersal started when chicks were 3.5 km apart from the nest during two consecutive observations. This
distance and timespan could correspond to a one-time excursion performed by the juvenile and not
necessary the emancipation itself as discussed by Cadahía et al. (2007a). In fact, Nygård et al. (2016)
dened an excursion in Golden Eagle (
), as movements performed further than 10 km
apart from the nest with a return to the surroundings of the nest, less than 5 km apart. Different tracking
techniques, radio-tracking versus GPS/GSM telemetry, as well as the use of unbiased methods such as
recursive analyses, can explain differences in the results obtained, with much more precision and
accuracy in the later (Thomas et al. 2011).
During the whole dependence period, all individuals remained in the vicinity of the nest or hacking site a
similar amount of time, although the difference in the number of times individuals visited the nest area
suggest that parental prey deliveries in wild individuals take place around this area (Bustamante 1995).
Reintroduced individuals made little use of the hacking cage, thus using the platforms around the area
When observing the movements performed by the eagles throughout the whole dependence period, we
found that there were not substantial differences between wild and reintroduced birds. Overall, changes in
the share of time spent on each behavior throughout the PFDP were more gradual and occurred later in
reintroduced individuals than in wild ones. This progressiveness on the percentage of each movement
observed in all individuals and its delay is consistent with the development of ight skills, in which ights
that involve apping (i.e., searching/foraging) are rstly acquired and perfectioned with practice, and
more complex movements (i.e., active hunting and travelling/relocating) are later developed (Ruaux et al.
2020). In the Golden eagle, at the beginning of the dependence period, apping ights were found, and
complex movements developed later (Walker 2008), indeed, Weston et al. (2018) showed that the PFDP
could be divided into two phases: ontogenic phase and maintained phase. During the ontogenic phase,
eagles rst increase their mobility, during the rst 10 weeks (68 days), and later, this mobility was
maintained until week 14 (99 days). However, the ability to perform complex movements seems to be
acquired sooner if there is parental presence (Ruaux et al. 2020). Juveniles learning from parents has
also been observed in the Osprey (
), where hand-raised chicks acquired hunting skills
without the presence of other conspecics (Schaadt and Rymon 1982).
According to Balbontín and Ferrer (2005), two different hypotheses could explain the onset of dispersal:
Resource Competition Hypothesis (RCH) (Howard 1960; Murray 1967) and the Ontogenic Switch
Hypothesis (OSH) (Holekamp 1986). The rst one explains emancipation as a response to competition
with their conspecics for food resources, so under abundance of food they will not leave the parental
territory. The OSH states that when resources are abundant, juveniles achieve earlier the optimal body
condition and therefore emancipate earlier. Here, as Balbontín and Ferrer (2005), we observed that
behavior in Bonelli’s eagle juveniles might support the RCH, as reintroduced juveniles, which were fed
, delayed the age of dispersal. Similar results were obtained in juvenile golden eagles, which
despite acquiring enough physical capacity, remained in their parental territory (Weston et al. 2018).
Studies aimed at assessing the outcome of reintroductions such as this one are essential to provide new
insights in conservation biology, offering crucial information on the ecacy of different techniques with
applications for future reintroduction projects. Our research highlights the importance of reintroductions
together with an exhaustive monitoring of reintroduced individuals. Overall, wild and reintroduced eagles
behaved in a similar way and neither sex nor year had an effect on the PFDP. In this way, the techniques
used during the whole process did not affect the development and behaviour of individuals, as they do
not differ substantially from the wild ones. Our results also illustrate that although parents are not
responsible for the onset of dispersal, they can prompt an earlier ending of the dependence period and in
conditions of high food availability, juveniles delay their onset of dispersal.
Data Availability Statement
All data used in this study are publicly available upon request to data managers in the online data
repository Movebank (www.movebank.org), project “Bonelli's eagle Mallorca. Life Bonelli / AQUILA a-
LIFE” (project ID = 1019304600) and project “LIFE ConRaSi - Bonelli's eagle” (project ID = 268207241).
OE-C and PL-L contributed equally to this paper. OE-C and PL-L conceived the ideas and designed
methodology; GC, MDV, MG, JJI, MLV, JM, SM, and CV collected the data; OE-C and PL-L analyzed the
data; OE-C and PL-L led the writing of the manuscript. All authors contributed critically to the drafts and
gave nal approval for publication. This research did not receive any specic grant from funding
agencies in the public, commercial, or not-for-prot sectors.
Bonelli’s eagles’ raw data were provided by two projects co-funded by European Union: “LIFE Bonelli”
(LIFE12 NAT/ES/000701) in Mallorca and “LIFE ConRaSi” (LIFE14 NAT/IT/001017) in Sicily. The authors
declare that no grants or other support were received during the preparation of this manuscript.
Competing interests’ statement
The authors have no relevant nancial or non-nancial interests to disclose.
Alonso JC, González LM, Heredia, B, GonzálezJL (1987) Parental care and the transition to independence
of Spanish Imperial Eagles Aquila heliaca in Doñana National Park, southwest Spain. Ibis 129: 212– 224.
Akaike H (1974) A New Look at the Statistical Model Identication. IEEE Trans Automat Contr19:716-723.
Amar A, Arroyo B E, Bretagnolle V (2000) Post-edging dependence and dispersal in hacked and wild
Montagu's Harriers Circus pygargus. Anim Behav 63:235-244.
Armstrong D P, Seddon, P J (2008) Directions in reintroduction biology.Trends Ecol Evol
Arroyo B E, Cornulier T De, Bretagnolle, V (2002) Parental investment and parent–offspring conicts
during the postedging period in Montagu’s harriers. Anim Behav 63:235–244.
Balbontín, J, Ferrer, M (2005) Factors affecting the length of the post-edging period in the Bonelli’s Eagle
Hieraaetus fasciatus. Ardea 93:189–198.
Balbontín, J, Ferrer, M (2009) Movements of juvenile Bonelli's Eagles Aquila fasciata during dispersal.
Bird Study 56:86-95.
Barton K (2020) MuMIn: Multi-Model Inference. R package version 1.43.17. https://CRAN.R-
Benjamini Y, Hochberg Y (1995) Controlling the False Discovery Rate: A Practical and Powerful Approach
to Multiple Testing. J R Stat Soc Ser B57:289-300.
Bennison A, Bearhop S, Bodey T W, Votier S C, Grecian W J, Wakeeld E D, Hamer K C, Jessopp M (2017)
Search and foraging behaviors from movement data: A comparison of methods. Ecol Evol 8:13–24.
BirdLife International (2015) Aquila fasciata. The IUCN Red List of Threatened Species 2015:
e.T22696076A60132541. Downloaded on 16 September 2020.
Bowler D E, Benton T G (2005) Causes and consequences of animal dispersal strategies: relating
individual behaviour to spatial dynamics. Biol Rev 80:205-225.
Bunnefeld N, Börger L, van Moorter B, Rolandsen C M, Dettki H, Solberg E J, Ericsson G (2011) A model‐
driven approach to quantify migration patterns: individual, regional and yearly differences. J Anim Ecol
Bustamante, J (1995) The duration of the post-edging dependence period of Ospreys Pandion haliaetus
at Loch Garten, Scotland. Bird study 42:31–36.
Bracis C Bildstein K Mueller T (2018) Revisitation analysis uncovers spatio-temporal patterns in animal
movement data.Ecography. 10.1111/ecog.03618, http://dx.doi.org/10.1111/ecog.03618.
Cadahía L, López-López P, Urios V, Negro J J (2007a)Estimating the onset of dispersal in endangered
Bonelli’s Eagles Hieraaetus fasciatus tracked by satellite telemetry: A comparison between methods.Ibis.
Cadahía L, Urios V, Negro J J (2007b)Bonelli's Eagle Hieraaetus fasciatus juvenile dispersal: hourly and
daily movements tracked by GPS. Bird Study. https://doi.org/10.1080/00063650709461484
Cadahía L, López-López P, Urios V, Negro J J (2010)Satellite telemetry reveals individual variation in
juvenile Bonelli’s eagle dispersal areas. Eur J Wildl Res. https://doi.org/10.1007/s10344-010-0391-z
Cagnacci F, Focardi S, Ghisla A, Van Moorter B, Merrill E H, Gurarie E, Heurich M, Mysterud A, Linnell J,
Panzacchi M, May R, Nygård C R, Rolandsen C Hebblewhite M (2016) How many routes lead to
migration? Comparison of methods to assess and characterize migratory movements. J Anim Ecol 85:54-
Carrascal L M, Seoane J (2009)Factors affecting large-scale distribution of the Bonelli’s eagle Aquila
fasciata in Spain.Ecol Res https://doi.org/10.1007/s11284-008-0527-8
Ceballos O, Donazar J A (1990)Parent‐offspring Conict during the Post‐edging Period in the Egyptian
Vulture Neophron percnopterus (Aves, Accipitridae) Ethology. https://doi.org/10.1111/j.1439-
Cramp S, Simmons K E L (1994) The birds of the western Palearctic. Oxford: Oxford University Press.
Di Vittorio M, López-López P (2014) Spatial distribution and breeding performance of Golden Eagles
Aquila chrysaetos in Sicily: implications for conservation. Acta Ornithol 49:33-45.
Di Vittorio M, Lo Valvo M, Di Trapani E, Sanguinetti A, Ciaccio A, Grenci S, Zafarana M, Giacalone G, Patti
N, Cacopardi S, Rannisi P, Scuderi A, Luiselli L, La Grua G, Cortone G, Merlino S, Falci A, Spinella G, López-
López P (2019) Long-term changes in the breeding period diet of Bonelli’s eagle (Aquila fasciata) in Sicily,
Italy. Wild. Res. doi.org/10.1071/WR18081
Di Vittorio M, Sará M, López-López P (2012) Habitat preferences of Bonelli’s Eagles Aquila fasciata in
Sicily. Bird Study. https://doi.org/10.1080/00063657.2012.656577
Diamond J M (1975) Assembly of species communities. In: J.M. Diamond M. L. Cody (eds.) Ecology and
Evolution of Communities, Cambridge: Harvard University Press, pp 342-444
Duncan R P, Boyer A G, Blackburn T M (2013) Magnitude and variation of prehistoric bird extinctions in
the Pacic. PNAS. https://doi.org/10.1073/pnas.1216511110
Dzialak M R, Lacki M J, Vorisek S (2007) Survival, mortality, and morbidity among peregrine falcons
reintroduced in Kentucky. J Raptor Res https://doi.org/10.3356/0892-
Efrat R, Hatzofe O, Miller Y, Berger‐Tal O (2020) Determinants of survival in captive‐bred Griffon Vultures
Gyps fulvus after their release to the wild . Conserv Sci Pract 2:e308.
Ferrer M (1992) Regulation of the period of postedging dependence in the Spanish Imperial Eagle Aquila
adalberti. Ibis 134, 128-133.Ferrer, M. (1993) Juvenile dispersal behaviour and natal philopatry of a long‐
lived raptor, the Spanish Imperial Eagle Aquila adalberti. Ibis 135: 132-138.
Ferrer C, Ferrer J, Ferrer M (2019) Apparent Natural Recolonization of an Island by the Seychelles Kestrel
(Falco araea) J Raptor Res 53: 355-357.
Ferrer M, Morandini V, Baguena G, Newton I (2018) Reintroducing endangered raptors: A case study of
supplementary feeding and removal of nestlings from wild populations. J Appl Ecol
Fischer J, Lindenmayer D B (2000) An assessment of the published results of animal relocations. Biol
Garcelon D K (1985) Mounting backpack telemetry packages on bald eagles.Arcata: Institute for Wildlife
Studies.García, V., Iglesias‐Lebrija, J. J., Moreno‐Opo, R. (2021)Null effects of the Garcelon harnessing
method and transmitter type on soaring raptors.Ibis 163: 899-912.
Garriga J, Palmer J R B, Oltra A, Bartumeus F (2016)Expectation-maximization binary clustering for
behavioural annotation. PLoS One. R package version 2.0. 3.
Gates S, Donald P F (2000) Local extinction of British farmland birds and the prediction of further loss.J
Gil-Sánchez J M (2000) Efecto de la altitud y de la disponibilidad de presas en la fecha de puesta del
águila-azor perdicera (Hieraaetus fasciatus) en la provincia de Granada (SE de España) Ardeola, 47, 1-
8.Greenwood, P.J. (1980) Mating systems, philopatry and dispersal in birds and mammals.Anim Behav
Greenwood P J, Harvey P H (1982) The natal and breeding dispersal of birds. Ann Rev Ecolog Syst 13: 1-
Hernández-Matías A, Real J, Parés F, Pradel R (2015)Electrocution threatens the viability of populations
of the endangered Bonelli's eagle (Aquila fasciata) in Southern Europe. Biol Conserv 191:110-116.
Butchart SH, Carpenter K E, Chanson J,
Collen B, Cox N A, Darwall W R T, Dulvy N K, Harrison L R, Katariya V, Pollock C M, Quader S, Richman N I,
Rodrigues S L, Tognelli M F, Vié J,
et al. (
The impact of conservation on the status of the world's
Holekamp K E (1986) Proximal Causes of Natal Dispersal in Belding’s Ground Squirrels (Spermophilus
Beldingi) Ecol Monogr 56:365–391.
Howard W E (1960) Innate and Environmental Dispersal of Individual Vertebrates. Am. Midl Nat 63:152–
IUCN – SSC Species Conservation Planning Sub-Committee. (2017) Guidelines for Species Conservation
Planning. Version 1.0. Gland, Switzerland: IUCN. xiv + 114 pp.
Kays R, Crofoot M C, Jetz W, Wikelski M (2015) Terrestrial animal tracking as an eye on life and planet.
Kenward R E (2000) A manual for wildlife radio tagging. London: Academic press.
Kenward R E, Marcström V, Karlbom M (1993) Post-nestling behaviour in goshawks, Accipiter gentilis: I.
The causes of dispersal. Anim Behav 46:365-370.
Krüger S, Amar A (2017) Insights into post-edging dispersal of Bearded Vultures Gypaetus barbatus in
southern Africa from GPS satellite telemetry. Bird Study 64:125-131.
Lacombe D, Bird D M, Hibbard K A (1994) Inuence of reduced food availability on growth of captive
American kestrels. Can J Zool 72: 2084-2089.
López-López P (2016) Individual-based tracking systems in ornithology: welcome to the era of big data.
López-López P, Sarà M, Di Vittorio M (2012) Living on the edge: assessing the extinction risk of critically
endangered Bonelli’s Eagle in Italy.PLoS One 7:e37862.
López-López P, Gil J A, Alcántara M (2014)Post-edging dependence period and onset of natal dispersal
in bearded vultures (gypaetus barbatus): New insights from gps satellite telemetry. J Raptor Res.
MacArthur R H, Wilson E O (1967) The Theory of Island Biogeography. Princeton: Princeton University
Press.McClure, C. J., Rolek, B. W. (2020) Relative conservation status of bird orders with special attention
to raptors. Front Ecol Evol 8:593941.
McClure C J, Westrip J R, Johnson J A, Schulwitz S E, Virani M Z, Davies R, Symes A, Wheatley H,
Thorstorm R, Amar A, Buij R, Jones VR, Williams NP, Buechley ER, Butchart, S. H. (2018) State of the
world's raptors: Distributions, threats, and conservation recommendations. Biol
Conserv.https://doi.org/10.1016/j.biocon.2018.08.012.Monti F, Sforzi A, Marie Dominici J (2012) Post-
edging dependence period of ospreys pandion haliaetus released in central Italy: Home ranges, space
use and aggregation.Ardeola.https://doi.org/10.13157/arla.59.1.2012.17
Muriel R, Ferrer M, Balbontín J, Cabrera L, Calabuig C P (2015)Disentangling the effect of parental care,
food supply, and offspring decisions on the duration of the postedging period. Behav Ecol 26: 1587-
Murray Jr B G (1967) Dispersal in vertebrates. Ecology, 48:975-978.
Myers N, Mittermeier R A, Mittermeier C G, Da Fonseca G A B, Kent J (2000) Biodiversity hotspots for
conservation priorities. Nature 403:853–858.
Nicoll M A C, Jones C G, Norris K (2004) Comparison of survival rates of captive-reared and wild-bred
Mauritius kestrels (Falco punctatus) in a re-introduced population. Biol Conserv 118:539-548.
Nygård T, Jacobsen K-O, Johnsen T V, Systad G H (2016) Dispersal and Survival of Juvenile Golden
Eagles (Aquila chrysaetos) from Finnmark, Northern Norway. J Raptor Res 50:144–160.
Ontiveros D, Pleguezuelos J M (2003) Inuence of climate on Bonelli’s eagle’s (Hieraaetus fasciatus V.
1822) breeding success through the Western Mediterranean. J Biogeogr.https://doi.org/10.1046/j.1365-
O’Rourke E (2014) The reintroduction of the white-tailed sea eagle to Ireland: People and wildlife. Land
use policy. https://doi.org/10.1016/j.landusepol.2013.10.020
Palma L, Mira S, Cardia P, Beja P, Guillemaud T, Ferrand N, Cancela M L Da Fonseca LC (2001) Sexing
Bonelli’s eagle nestlings: morphometrics versus molecular techniques. J Raptor Res 35:187-193.
Peltonen A, Hanski I (1991) Patterns of island occupancy explained by colonization and extinction rates
in shrews. Ecology 72:1698-1708.
Rahman M L, Batbayar N, Purev-Ochir G, Etheridge M, Dixon A (2015) Inuence of nesting location on
movements and survival of juvenile saker falcons falco cherrug during the post-edging dependence
Ramos R F, Silva J P, Carrapato C, Rocha P, Marques P A, Palmeirim J M (2019) Spatial behaviour of
Spanish Imperial Eagle Aquila adalberti juveniles during the dependence period revealed by high-
resolution GPS tracking data. J Ornithol 160: 463-472.
Real J, Grande J M, Mañosa S, Sánchez-Zapata J A (2001)Causes of death in different areas for
Bonelli's Eagle Hieraaetus fasciatus in Spain. Bird study, 48:221-228.
Real J, Mañosa S, Codina J (1998) Post-nestling dependence period in the Bonelli's Eagle. Ornis Fennica,
Rempel R S Rodgers A R (1997) Effects of differential correction on accuracy of a GPS animal location
system.J Wildl Manage61:525–530.
Robert A, Colas B, Guigon I, Kerbiriou C, Mihoub J B, Saint‐Jalme M, Sarrazin F (2015) Dening
reintroduction success using IUCN criteria for threatened species: a demographic assessment.Anim
Conserv, 18: 397-406.
Rosenzweig M L, Clark C W (1994) Island extinction rates from regular censuses. Conserv Biol 8:491-494.
Ruaux G, Lumineau S, De Margerie E (2020) The development of ight behaviours in birds: Bird ight
development. Proc R Soc B Biol Sci.https://doi.org/10.1098/rspb.2020.0668rspb20200668
Schaadt CP Rymon LM (1982) Innate shing behavior by ospreys. J Rapt Res, 16:61–62
Seddon PJ, Armstrong D P, Maloney R F (2007) Developing the science of reintroduction biology. Conserv
Sergio F, Tavecchia G, Tanferna A, López Jiménez L, Blas J, De Stephanis R, Marchant T A, Kumar N
Hiraldo F (2015)No effect of satellite tagging on survival, recruitment, longevity, productivity and social
dominance of a raptor, and the provisioning and condition of its offspring. J Appl Ecol, 52: 1665-1675.
Soutullo A, Urios V, Ferrer M (2006)Post-edging behaviour in Golden Eagles Aquila chrysaetos : onset of
juvenile dispersal and progressive. Ibis 148:307–312.
Tershy B R, Shen K W, Newton K M, Holmes N D, Croll D A (2015) The importance of islands for the
protection of biological and linguistic diversity. BioScience. https://doi.org/10.1093/biosci/biv031
Tordoff H B, Redig P T (2001) Role of genetic background in the success of reintroduced peregrine
falcons. Conserv Biol 15:528-532.
Thomas B, Holland J D, Minot E O (2011) Wildlife tracking technology options and cost considerations.
Trivers R L (1974) Parent-offspring conict.Integr Comp Biol 14:249-264.
Urios V, Muñiz-López R, Vidal-Mateo J (2017) Juvenile dispersal of harpy eagles (Harpia harpyja) in
Ecuador. J Raptor Res, 51: 439-44.
Viada C, Parpal L, Morro B, Mayol J (2015) El águila de Bonelli (Aquila fasciata) en Mallorca: Su
extinción y su reintroducción. In: Servei de Protecció d’Espècies (Eds.) Llibre Verd de Protecció d’Espècies
a les Balears, Balearic Islands, pp 283-294.
Viada C, Iglesias JJ (2017) La jaula hacking: nuevo sistema para el LIFE Bonelli. In:Life Bonelli (Eds.)
Recuperación integral de las poblaciones de águila perdicera en España.Navarra, pp 111-121.
Vergara P, Fargallo J A, Martínez-Padilla J (2010)Reaching independence: food supply, parent quality,
and offspring phenotypic characters in kestrels. Behav Ecol, 21: 507-512.
Vitousek P M (2002) Oceanic islands as model systems for ecological studies.J Biogeogr.
Walker D G (1987) Observations on the post‐edging period of the Golden Eagle Aquila chrysaetos in
England. Ibis. https://doi.org/10.1111/j.1474-919X.1987.tb03163.x
Walls S S, Kenward R E (1995) Movements of radio-tagged Common Buzzards Buteo buteo in their rst
year. Ibis 137:177-182.
Watson A J (1981) Population ecology, food and conservation of the Seychelles Kestrel (Falco araea) on
Mahè (Doctoral dissertation, University of Aberdeen)
Weathers W W, Sullivan, K. A. (1989) Juvenile foraging prociency, parental effort, and avian reproductive
success. Ecol Monogr 59:223-246.
Weston E D, Whiteld D P, Travis J M, Lambin X (2013) When do young birds disperse? Tests from
studies of golden eagles in Scotland. BMC ecology 13:42.
Weston E D, Whiteld D P, Travis J M, Lambin X (2018) The contribution of ight capability to the post‐
edging dependence period of golden eagles. J. Avian Biol., 49: e01265.
Zuur A, Ieno E N, Walker N, Saveliev A A, Smith, G. M. (2009) Mixed effects models and extensions in
ecology with R. New York: Springer Science Business Media.
Table 1.- Model selection ranking for each parameter (i.e., age of dispersal, age of first
flight, PFDP length, number of revisits and residence time) in relation to origin, age, sex,
their additive effects and their interaction. Best models are highlighted in bold. (df =
degrees of freedom, AIC = Akaike Information Criterion, AICw = model weight).
Parameter Model df AIC ∆AIC AICw
First flight Intercept 3 -53.459 0.000 0.859
Origin 4 -49.302 4.157 0.107
Sex 4 -46.735 6.725 0.030
Origin + Sex 5 -42.622 10.837 0.004
Origin + Sex + Origin*Sex 6 -37.134 16.325 0.000
Age dispersal Intercept 3 -40.543 0.000 0.677
Origin 4 -38.578 1.965 0.254
Sex 4 -35.413 5.130 0.052
Origin + Sex 5 -32.961 7.581 0.015
Origin + Sex + Origin*Sex 6 -28.585 11.957 0.002
PFDP length Intercept 3 5.507 0.000 0.772
Origin 4 9.034 3.527 0.132
Sex 4 10.049 4.542 0.080
Origin + Sex 5 13.617 8.110 0.013
Origin + Sex + Origin*Sex 6 16.914 11.407 0.003
Revisits Origin 4 141.118 0.000 0.463
Origin + Sex 5 142.490 1.372 0.233
Origin + Sex + Origin*Sex 6 142.726 1.609 0.207
Intercept 3 145.049 3.931 0.065
Sex 4 146.430 5.312 0.032
Residence Time Origin + Sex + Origin*Sex 6 182.204 0.000 0.356
Origin 4 183.454 1.251 0.190
Origin + Sex 5 183.623 1.419 0.175
Intercept 3 183.999 1.795 0.145
Sex 4 184.152 1.948 0.134
Table 2.- Model selection ranking for each behavior obtained by means of the Expectation
Maximization Binary Clustering (EMbC) algorithm (i.e., resting, foraging, travelling and
hunting) in relation to origin, age, sex, their additive effects and their interaction. Best
models are highlighted in bold. (df = degrees of freedom, AIC = Akaike Information
Criterion, AICw = model weight).
Movement Model df AIC ∆AICc AICw
Resting Intercept 3 -519.261 0.000 0.889
Origin 4 -514.754 4.507 0.093
Sex 4 -511.368 7.893 0.017
Week 22 -405.127 114.135 0.000
Sex + Origin + Week 24 -393.253 126.009 0.000
Sex + Origin + Week + Origin*Sex*Week 76 -224.237 295.025 0.000
Foraging Week 22 -208.067 0.000 0.969
Sex + Origin + Week 24 -201.208 6.859 0.031
Sex 4 -162.891 45.176 0.000
Intercept 3 -162.688 45.379 0.000
Origin 4 -156.887 51.179 0.000
Sex + Origin + Week + Origin*Sex*Week 76 -35.3233 172.743 0.000
Travelling Week 22 -729.816 0.000 0.998
Sex + Origin + Week 24 -716.607 13.209 0.001
Intercept 3 -713.564 16.252 0.000
Sex 4 -708.759 21.057 0.000
Origin 4 -707.531 22.285 0.000
Sex + Origin + Week + Origin*Sex*Week 76 -465.041 264.775 0.000
Hunting Week 22 -1172.54 0.000 0.943
Sex + Origin + Week 24 -1166.94 5.604 0.057
Origin 4 -1097.35 75.186 0.000
Intercept 3 -1096.12 76.425 0.000
Sex 4 -1091.92 80.622 0.000
Sex + Origin + Week + Origin*Sex*Week 76 -861.499 311.042 0.000
Comparison of ve parameters used to characterize the post-edging dependence period of wild and
reintroduced juvenile Bonelli’s eagles in two Mediterranean islands, Sicily and Mallorca, respectively.
Boxplots represent median as well as 25% and 75% quartiles. Outliers are shown as dots.
Percentage of time spent by juvenile Bonelli’s eagles on the four categories of behavior computed using
the unsupervised Expectation Maximization binary Clustering (EMbC) method throughout the juvenile
post-edging dependence period by origin and sex. Grey areas represent standard error.
General share of time of juvenile Bonelli’s eagles spent on the four categories of behavior obtained by
means of the Expectation Maximization Binary Clustering (EMbC) algorithm by origin (i.e., wild or
reintroduced) during the PFDP.