Movement ecology of long-distance migrants: insights from the Eleonora s falcon and other raptors. Ugo Mellone

Transcription

1 Movement ecology of long-distance migrants: insights from the Eleonora s falcon and other raptors Ugo Mellone

2

3 Movement ecology of longdistance migrants: insights from the Eleonora s falcon and other raptors Ugo Mellone PhD Thesis University of Alicante, 2013

4 Cover: the tracks represent schematically the migration routes of Eleonora s falcons. Juveniles: yellow; adults: red (autumn) and blue (spring). Details in paper 3. Cover layout, graphics and photographs Ugo Mellone. Published papers are reprinted with permission from the publishers.

5 Ecología del movimiento de migradores de larga distancia: Ejemplos con el Halcón de Eleonora y otras rapaces Tesis presentada por Ugo Mellone para optar al grado de Doctor en Biología por la Universidad de Alicante. La tesis se presenta como compendio de artículos, y solicitando la mención de Doctor internacional. Director: Dr. Vicente Urios Moliner Co-Director: Dr. Pascual López López Universidad de Alicante Centro Iberoamericano de la Biodiversidad (CIBIO) 2013

6

7 CONTENTS Introduction 9 General methods..13 Results and discussion: an overview..14 Future perspectives..16 This thesis is based on the following papers Section one: long-distance migration I II López-López P., Limiñana R., Mellone U. &Urios V From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Landscape Ecology 25: Mellone U., López-López P., Limiñana R. & Urios, V Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. International Journal of Biometeorology 55: III Mellone U., López- López P., Limiñana R., Piasevoli G. & Urios V The transequatorial loop migration system of Eleonora s falcon: differences in migration patterns between age classes, regions and seasons. Journal of Avian Biology...41 IV V Mellone U., Limiñana R., López-López P. & Urios Regional and agedependent differences in the effect of wind on the migratory routes of Eleonora's Falcon. Manuscript Mellone U., Klaassen R.H.G., García-Ripollés C., Limiñana R., López-López P., Pavón D., Strandberg R., Urios V., Vardakis M. & Alerstam T Interspecific comparison of the performance of soaring migrants in relation to morphology, meteorological conditions and migration strategies. PLoS ONE 7(7): e Section two: short range movements VI Mellone U., López-López P., Limiñana R. & Urios V Wintering habitats of Eleonora's Falcons Falco eleonorae in Madagascar. Bird Study 59: VII Mellone U., López-López P., Limiñana R. & Urios V., Summer pre-breeding movements of Eleonora s Falcons Falco eleonorae revealed by satellite telemetry: implications for conservation. Bird Conservation International...91 VIII Mellone U., Urios V, Rguibi-Idrissi H., Limiñana R., Benhoussa A. & López-López P., Ranging behaviour of Eleonora s Falcons Falco eleonorae during chickrearing. Acta Ornithologica 47: Conclusions Introducción, sintesis y conclusiónes en Español Agradecimientos/Acknowledgements/Ringraziamenti

8

9 CAPTIONS OF THE PICTURES Section one: Eleonora s falcon, Columbretes islands (Spain) I: the Sahara desert during night (Morocco) II: Eleonora s falcon, Columbretes islands (Spain) III: adult Eleonora s falcon, Columbretes islands (Spain) IV: wind shaping the dunes of the Sahara desert (Morocco) V: migrating Short-toed eagle, Strait of Gibraltar (Spain) Section two: adult Eleonora s falcons, Antikythira (Greece) VI: winter locations of Eleonora s falcons in Madagascar (created with Google Earth) VII: coniferous forest, Sierra de Segura (Spain) VIII: Essauoira, the town seen from the island (Morocco) 7

10

11 INTRODUCTION Migration: living in an endless summer I borrow this title from a nice paper by Shaffer et al. (2006) on the tracking of Sooty shearwaters, because it resumes very brilliantly what being a migratory bird means. Migration is a strategy to maximise fitness in a seasonal environment (Alerstam et al. 2003), e.g. regular and seasonal movements between breeding and wintering grounds, occurring roughly during the same period each year. Migration behaviour has evolved independently across several taxa, e.g. insects, fishes, reptiles and mammals, but is among birds that it is most widespread. These movements occur in response to seasonal changes in the abundance of food resources, normally with the non-breeding period spent at lower latitudes than the breeding one (Newton 2008). In this way migratory birds can exploit enough resources allyear around, focusing on survival during winter, in order to reproduce during summer. An obvious question would be why migratory birds do not spend the whole year at low latitudes, where resources are less fluctuating? There are many reasons to explain this, and, in particular: 1) to avoid competition with resident birds, 2) to exploit the peak of resources at higher latitudes during summer and 3) to ensure a safe nest site, which is not always available at lower latitudes (Newton 2008). Evolution has rewarded the individuals that took into account these factors, shaping very fascinating life-histories. Personally, I find amazing to think that many bird species (for example, many waders) usually considered as Arctic species, indeed stay in the Arctic just a couple of months to reproduce, while they live most of the year in the tropics! These risky strategies persist because fitness benefits associated with the reproduction at higher latitudes outweigh the costs of long-distance migration. So, what have the birds to do during migration? In comparison to the shortdistance movements that take place in the breeding and in the wintering areas, migratory journeys can last thousands of kilometres, during which birds have to replenish their energy stores, find the right way and, obviously, fly. Migratory behaviour shows a strong variation among different species and also within species, according to population, season, sex, age (experience), landscape properties and meteorological conditions (Alerstam 2011). Different populations can experience different selective pressures, which promote variation in migration distances and location of wintering areas (Panuccio et al. 2013). It has been shown that longdistance migration is also associated to higher daily flight range in comparison with short-distance one (Alerstam 2003, Strandberg et al. 2008). Motivational asymmetries due to different migration goals (e.g., breeding vs. wintering) arise since birds are expected to migrate more quickly when they are en route to the breeding grounds, thus minimizing time rather than energy expenditure (Kokko 1999). Therefore, a timeselected (i.e., faster) migration would be expected in spring rather than in autumn (Alerstam 2006). Moreover, males are more time constrained than females because they have to ensure a high quality breeding site and defend it (Rubolini et al. 2004). In this scenario, immature individuals are expected to migrate slower than adults, being not motivated by breeding duties. Other age differences arise since experienced individuals have better navigation abilities (Thorup et al. 2003, Agostini 2004) and do better than juveniles when 9

12 fuelling before migration, selecting the optimal departure in relation to weather conditions (Newton 2008, Morganti et al. 2011). The migratory performance is strongly influenced also by external factors like landscape properties and meteorological agents (Klaassen et al. 2008, Borher et al. 2012). Landscape properties promote variation in migration speed since some bird species fly slower in order to take advantage of foraging opportunities when moving over benign habitats, and increase their migration speed when they have to overcome ecological barriers, such as deserts and sea surfaces (Strandberg et al. 2009). Ecological barriers can also influence routes geometry, for example shaping large detours (Mellone et al. 2011). The desert crossing can be very dangerous, triggering also deleterious carry-over effects on breeding output (Strandberg et al. 2010). It has been recently shown that yearly variations in environmental conditions occurring even at a very local scale can have strong effects for a whole migration system (Tøttrup et al. 2012). Among raptors, pure soaring migrants with a low aspect ratio wing morphology, such as eagles, vultures and buzzards are predicted to carry with them enough fat in order to complete migration without foraging, while flapping migrants have to perform stop-overs or a fly-andforage strategy (Smith et al. 1986; Alerstam and Hedenström 1998; Hake et al. 2003; Panuccio et al. 2006). Concerning meteorological conditions, birds take advantage of winds to save both time and energy during migrations (Alerstam 2003), although strong crosswinds can be very dangerous, especially when flying over ecological barriers (Klaassen et al. 2011). Also thermal uplift can boost migration speed (Borher et al. 2012), especially interacting with topography, since it has been shown that the same species can show diverse responses according to different orographic features (Mandel et al. 2011). The aim of the first section of this thesis is to analyse how all these factors interact to shape the migration strategies of Eleonora s falcons (Falco eleonorae) tracked by Argos satellite telemetry. Moreover, I also present an interspecific comparison among four raptors species tracked by GPS satellite telemetry that share the same flyway of Eleonora s falcons during the crossing of the Sahara desert. The main questions that I tried to answer are (with the relative chapter given in brackets): - Do Eleonora s falcons adjust their migration speed and routes in relation to ecological barriers? (1, 3) - Is such variation dependent from flight speed, time budget and/or straightness? (1, 3) - How individual flexibility, landscape characteristics and age interact with wind conditions to shape the migration routes? (2, 4) - Which are the cues that Eleonora s falcons use to orientate? (1, 2, 3, 4) - Which factors promote behavioural differences between age classes in respect to speed and stop-over behaviour? (3) - Are there seasonal differences in geometry of migration routes promoting a loop migration, and why? (3) - Which internal and external factors affect the migratory performances of soaring birds crossing the Sahara and are there interspecific differences? (5) Conservation of migratory birds Since migratory birds spend different parts of their life cycle in different places located thousands of kilometres 10

13 apart each other, they are exposed to a high variety of risks, and conservation efforts aimed at protecting them, for example, during the breeding season, can be jeopardized by threats occurring during the wintering one, and viceversa (Martin et al. 2007). It has been clearly shown that long-distance migrants are declining at a faster rate than other species (Sanderson et al. 2006). Moreover, carry-over effects, e.g. events occurring in a given place during a given season, but having an impact also on processes occurring elsewhere in another season, can act even at a very fine scale, for example affecting the process of sexual selection (Reudink et al. 2009). Therefore, conditions experienced in non-breeding grounds may greatly affect the population dynamic of a species (see Zwarts et al 2009). As I will show in this thesis, the scenario is even more complex, since some species also use discrete staging areas apart those belonging to the breeding and the wintering areas (prebreeding areas, see figure 1). All these issues highlight that is very important to identify the main areas, habitat and threats related to birds distribution during each part of their life cycle, in order to design the most effective conservation actions. The aim of the second section of this thesis is to describe the short range movements of the Eleonora s falcons taking place during the wintering, prebreeding and breeding period, identifying the main habitat used by the species during these stages of the life cycle and discuss their relative importance under a conservation perspective. Conservation of migratory birds from a flyway point of view is a subject of growing attention also from conservation agencies (Brouwer 2009). Given the very large range used throughout the year and the variety of Figure 1. The annual cycle of Eleonora s falcon. Percentages refers to the proportion of time spent in each phase. Numbers in brackets refer to the papers of this thesis. Data from paper 3. environments where this species forages, an integrate conservation strategy could benefit, on a very wide scale, also other species, both migrants and residents, since raptors has been proposed as among the most effective umbrella/flagship species (Sergio et al. 2008). The main questions I kept in mind for the second section of this thesis are the following (chapter in brackets): - Which are the main habitat used by the Eleonora s falcon during the nonbreeding periods and which are the main threats there? (6 and 7) - How, during the breeding season, Eleonora s falcons do integrate the use of their colony with that of nearby areas outside it, used either for foraging or for other purposes? (8). 11

14 The study species The main species with which this Ph.D. thesis deals is the Eleonora s falcon, although a chapter has also been dedicated to the migration strategies of other four species of raptors. The Eleonora s falcon is a medium sized raptor that breeds generally on small islands, almost exclusively in the Mediterranean basin, although some colonies are present also in Canary islands and on the Atlantic coast of Morocco (Walter 1979). The species breeds mainly on inaccessible cliffs, although, when terrestrial predators are absent, it can also breed on the ground, as on the island of Mogador (chapter 8). I must admit that the affirmation by H. Walter in his monograph on the species, I have yet to see a breeding habitat of Eleonora s falcon that would not be worthy of a picture is very appropriate, as I had the chance to verify it in the colonies of Spain and Morocco, where I carried out the field work for this thesis, and also in other colonies I visited, such as those in Sardinia, Sicily, Croatia and Greece. This species adjusts its breeding season in order to coincide with the postbreeding migration of trans-saharian passerines (late August-September), being the only European bird species breeding in this period (Gangoso et al. 2013). Outside the breeding season, Eleonora s falcons feed mainly on insects (Walter 1979). The global population has been recently estimated between 13,000 and 14,000 breeding pairs, occurring mainly in Greece (90%; Dimalexis et al. 2008). Although the species is classified by the IUCN as least concern, it is included in Annex I of the Directive 2009/147/EC on the conservation of wild birds and belongs to the category SPEC 2 of BirdLife International, which means that is a species with an unfavourable European conservation status, and with more than half of the global breeding or wintering population concentrated in Europe. Indeed, the fact that the species is highly concentrated during both breeding and wintering suggests that it might be susceptible of a global decline even in the case of local perturbations. Unlike bigger raptors with relatively shorter and broader wings (low aspect ratio, sensu Kerlinger 1989), such as eagles and vultures, that migrate by exploiting thermal convection using soaring flight, the Eleonora s falcon is characterised by higher aspect ratio and therefore is more adapted to flapping flight. This morphological characteristic allows Eleonora s falcon to migrate on a broad front, without concentrating where sea-crossing are shorter. Almost the whole population of Eleonora s Falcon spend the winter in Madagascar (Walter 1979), that is located ca km (straight line) from the Spanish breeding colonies of the Western Mediterranean. Therefore, twice each year Eleonora s falcons perform a transequatorial journey, facing a huge variety of landscape and climatic conditions, and being therefore an excellent model to study the above mentioned topics. Similarly as I did for the Eleonora s falcon along the thesis, in chapter 4 I also analysed the migratory performance in relation to morphology, meteorological conditions and migration strategies of four other species of raptors: the Osprey (Pandion haliaetus), the Marsh harrier (Circus aeruginosus), the Egyptian vulture (Neophron percnopterus) and the Shorttoed eagle (Circaetus gallicus). My research group and the group where I did my first external research stay (University of Lund) were working on these species, using the same GPS transmitters, and therefore we felt 12

15 opportune to do this analysis within my PhD thesis. Unfortunatley, it has been impossible to include the Eleonora s falcon in this analysis, due to the differences in tracking devices, and consequently in data quality. In short, Eleonora s falcons are too light to carry GPS transmitters at present (> 22g), and therefore they have been tracked only with Argos ones (9.5g, see the next paragraph for details). The four above mentioned species were tagged during the breeding season in very different European countries (Sweden, Spain and Italy), but share all the same wintering grounds (Sub-Saharan Africa), and therefore the last part of their migratory flyway over the Sahara Desert. Finally, unlike the Eleonora s falcon, they fly mainly by soaring flight, exploiting thermal currents, and are thus diurnal migrants. For this reason, they are easier to observe during active migration than the Eleonora s falcon, and for me it has been very interesting to study them also through satellite telemetry after many field seasons of visual observations in islands and bottle-neck across the Mediterranean basin (Agostini et al. 2004, Panuccio et al. 2005, Lucia et al. 2011). GENERAL METHODS The findings presented in this thesis are based on satellite telemetry, a technology that allows to follow individual birds all over the world (Argos 2011). Basically, satellites collect the information (time and coordinates) sent by small transmitters mounted on each bird and powered by solar panels. Then, these data are collected and processed in the Argos centre of Toulouse (France) that makes them available on the internet in a very short time (according to the duty cycle, even a few minutes after their collection). Within the researches presented here, the weight of the transmitter has been always lower than the 5% of the birds body mass and, keeping in mind their average working duration, they have been mounted with a teflon harness locked in order to break after a few years. The transmitters used weigh from 9.5 grams (Eleonora s falcon) to 22-45g (the heavier species of paper 5). While the lighter transmitters allow to obtain only Argos location, that vary their accuracy according to the location class (from m to 100+ km in the worst cases), bigger ones send GPS quality data, with a maximum nominal error of 18 meters. Therefore, Argos locations have been filtered, using the procedures detailed in each paper. The majority of Eleonora s falcons tracked in this study were captured in the Western Mediterranean colonies of Spain, in particular in the Columbretes Islands in autumn and in the Balearic Islands in autumn 2007 and Adult birds were trapped using dho-gaza nets and a stuffed Eagle Owl (Bubo bubo) as a decoy, while juveniles were taken directly from their nests a few days before fledging. All birds were weighed, measured, ringed and sexed using molecular methods. Finally, all birds were equipped, using a teflon harness, with Microwave Telemetry s 9.5 g solar-powered satellite transmitters (Figure 2). Locations were collected using the Argos system, and the details on location classes and duty cycles are given in each chapter. As a general method, usually I downloaded the data of each bird using the Satellite Tracking and Analysis Tool (Coyne and Godley 2005), and then I carried out the basic calculations (time, distance, speed) and the consequent data filtering using Microsoft Excel, always visually inspecting, simultaneously, the same data through Google Earth. This software is very 13

16 useful to compare at a glance bird movements with habitat and topography, stimulating questions to which, afterwards, I tried to give an answer through more detailed analyses carried out using databases of meteorological data (chapters 2, 4 and 5) or GIS software (chapters 6 and 7). Figure 2. Eleonora s falcon with satellite transmitter. RESULTS AND DISCUSSION: AN OVERVIEW Section one: long-distance migration Paper 1: ecological barriers I analysed several aspects of the migration behaviour of Eleonora s falcons crossing the African continent during autumn migration. Eleonora s falcons adjust their speed according to the landscape characteristics of the crossed regions. In particular, they show higher daily distances during the crossing of ecological barriers such as the Sahara Desert and the Mozambique Channel, thanks to higher flight speeds and, especially, to a circadian migration pattern in which stationary behaviour is reduced to the minimum, with very few stops during the day and a significant amount of nocturnal migration during the desert crossing, as well as long nostop flight over the sea. The results also highlight that Eleonora s falcons rely on an orientation system that works during both day and night, which suggests that it is not related to landmarks visibility. Finally, no evidence has been found that the equatorial rainforest is an ecological barrier for this species during migration, as it has been shown for a similar species, the Hobby (Falco subbuteo). Paper 2: route flexibility I analysed the spring migration routes of three Eleonora s falcons during the spring crossing of the Indian Ocean, between Madagascar and Africa, a journey lasting km. Two individuals were followed in different years, allowing getting new insights on individual plasticity with regards the selection of migration routes. The differences among routes were much stronger among different years within the same individual than among different individuals within the same year. By plotting the routes over meteorological maps (i.e. vorticity patterns) I show that falcons are able to change their routes to minimize mortality risk in response to low pressure areas. The paper proposes the definition of meteorological barriers to describe areas where weather conditions are less favourable actively avoided by migrant birds. Moreover, the results highlighted again that Eleonora s falcons were able to successfully navigate in a featureless environment such as the open ocean. 14

17 Paper 3: effects of age, region and season on migration patterns I investigated whether differences in migratory behaviour and routes according to age classes, regions and seasons do occur and which factors promote them among migrating Eleonora s Falcons tagged with satellite transmitter in Spain and Croatia. I found that during autumn migration no age differences occur when crossing the Sahara desert, but in the remaining African regions, juveniles are more prone than adults to fly at a slower and more tortuous rate, as well as exhibiting longer stop-overs, particularly in the Sahel region. Such differences might be promoted by a lower foraging and premigratory fattening efficiency in juveniles. During spring, routes were significantly more eastern than during autumn, with Ethiopia being an important area for stop-over. This loop migration could be accounted for seasonal variation in the distribution of trophic resources and involves an eastward shift apparently fixed in Eleonora s falcons innate circannual program. These results are an illustration of how a migratory species make the most of seasonal resources on a continental scale throughout its annual cycle, changing the movement patterns in response to both internal (age) and external (habitat) factors. Paper 4: wind influence on migratory routes Within the autumn migration system of the Eleonora s falcon, I investigated at a daily scale the effect of tailwinds on forward movement rates, and of crosswinds on perpendicular ones. Data from different regions and age classes were analysed separately. The results showed that the effect of wind on movement rates was not uniform along the migratory journey, being stronger in the farthest region from the migration goal, the Sahara desert, which is also the most hazardous one. In the Sahel, the results are more conflicting, perhaps because daily movements are more shaped by the distribution of food resources. In Equatorial Africa, daily movement rates were mainly affected by crosswinds. It remains unclear which orientation mechanism allow Eleonora s falcons to reach such a narrow wintering area compensating at the same time for wind displacement. Paper 5: interspecific comparison of migratory performance I compared the migratory performance of four raptor species during their crossing of the Sahara desert at both an hourly (flight speed and altitude) and a daily scale (daily distance). Interspecific differences were stronger in spring, with long-distance migrants (>5000 km: Osprey and Marsh harrier) being faster than species that migrate shorter distances (Egyptian vulture and Shorttoed eagle). While time of day (that can be considered as a proxy of thermal strength) is the most important factor affecting variation in hourly speed, tailwind support is the most important one affecting daily distance, at the expense of day length and thermal strength. Controlling for wind assistance, that was less favourable during spring, the three species involving adult individuals (i.e. excluding Short-toed eagles) are, as expected, faster during spring. In the light of the predictions concerning future changes of worldwide wind patterns these results raise interesting questions. Section two: short range movements Paper 6: wintering habitats I analysed habitat preferences of three Eleonora s falcons wintering in 15

18 Madagascar, two of them tracked during two consecutive years. The results show that the birds showed site fidelity across years and selected degraded humid forests and cultivated areas close to pristine humid forest, probably to take advantage from a spillover edge effect of their insect preys into open areas where hunting is easier than inside the pristine humid forest. The current loss of pristine humid forests in Madagascar (that probably act as a source of prey) is a cause of concern with respect to the conservation of the species, given that the great majority of the global population spends the winter in Northern Madagascar. This study was the first detailed analysis on the habitat selection of a Palearctic migrant in Africa, and it could be useful to focus conservation efforts and further field studies. Paper 7: pre-breeding movements I analysed four summering events belonging to three Eleonora s falcons. Before settling in the colony, the falcons made inland movements in areas up to c.400 km distant from them, visiting several habitats, from forests to arable lands, from mountains to coastal areas, probably taking advantage of local concentrations of insects in a period of the year in which the breeding areas still do not offer foraging opportunities. This is the first study unravelling the exact connections between breeding colonies and these inland areas. Moreover, a literature review of anecdotal field observations confirms the importance of these areas, suggesting that our results could be applied to a wider proportion of individuals. Poisoning events occurred in Greece highlight that perturbations occurring in these pre-breeding foraging areas could seriously threaten Eleonora s Falcons with dramatic consequences at the population level. Therefore, conservation measures should be implemented not only at breeding and wintering grounds, but also in these temporary staging areas. Paper 8: ranging behaviour during breeding Transmission problems affecting the performance of the Argos system in the Mediterranean basin prevented its use to study ranging behaviour of Eleonora s Falcons during the breeding season in Europe. For the first time, it was possible to perform such a study by tagging birds on the Atlantic coast of Morocco. Two adult females breeding in the colony of Essaouira spent most of the time at sea during mornings, stayed mainly inland during afternoons, and rested in the colony during nights. Although most locations were less distant than 50 km away from the colony, extreme movements took also place to areas located more than 100 km away. Locating and protecting these inland areas used for resting and foraging may be of interest for the conservation of the species. FUTURE PERSPECTIVES This thesis provide answers to some of the questions that I proposed in the introduction, but some issues are still unsolved and other questions are arising as the knowledge improve. Future studies on the movement ecology of the Eleonora s falcon will hopefully take advantage of higher spatial and temporal resolution of new tracking devices, as well of reduced costs, to investigate with higher precision and more robust sample size how these birds accomplish their journeys mixing pure migratory behaviour with the fly-and-forage and/or stop-overs strategies, and how 16

19 the events occurring during migration, wintering and the pre-breeding movements affect individual conditions and breeding output. In fact, one of the most challenging tasks for migration research is to unravel carry-over effects between different life-history phases. The same fine scale data may also be used to study habitat selection during stop-over, which is also a very poorly studied topic. Although the prebreeding movements may act as a buffer against perturbations experienced during wintering (in comparison with species passing directly from the wintering grounds to the breeding ones), the effect of habitat destruction in Madagascar remains an important issue that deserves particular attention. My results concerning habitat selection during winter have been confirmed also by the researches involving Sardinian and Greek birds (Gschweng et al. 2012, Kassara et al. 2012), and therefore is now possible to identify with great precision the wintering priority areas for the great majority of the species population, which is not a common fact among other long-distance migrants, converting the Eleonora s falcon in one of the best-studied migratory raptors during the non-breeding season. The high interannual individual variability in migratory routes (see chapters 2 and 3) is striking when compared to the high site fidelity to wintering and pre-breeding areas (chapters 6 and 7). Still, it remains unclear which orientation system do Eleonora s falcon use, given the astonishing fine-tuned navigation abilities exhibited by both adults and juveniles, migrating alone (Thorup et al. 2010). The ability of sensing infrasound radiated from steep-sided topographic features, as suggested by Gschweng et al. (2009) is a both promising and fascinating hypothesis that remains to be tested. In this respect, also the seasonal differences in migration routes deserve great attention: perhaps the spring routes resemble more than the autumn ones the routes used by the ancestors of the Eleonora s falcon to colonize the Mediterranean basin tens of thousands of years ago (see Wink and Ristow 2000, López-López et al. 2009). These statements are valid, obviously, for many other species of long-distance migrants on which hopefully we will experience a dramatic increase in the knowledge of their habits outside the breeding season thanks to a higher diffusion and miniaturization of tracking devices. In particular, it would be very interesting to compare the findings of this thesis with other phylogenetically similar species, such as, for example, the Sooty falcon (Falco concolor), the Red-footed falcon (Falco vespertinus) and the Amur falcon (Falco amurensis). Understanding which internal and external factors affect the movement patterns of migratory birds and how these processes affect their breeding output is a fundamental issue under the current and the future projected scenarios of climate change. Building the bridge between movement ecology, population dynamics and species conservation is the next challenge. REFERENCES Agostini N Additional observations of age-dependent migration behaviour in western honey buzzards Pernis apivorus. Journal of Avian Biology 35: Agostini N, Premuda G, Mellone U, Panuccio M, Logozzo D, Bassi E, Cocchi L Crossing the sea en route to Africa: autumn migration of some Accipitriformes over two central Mediterranean islands. Ring 26: Alerstam T Bird migration speed. In: Berthold P, Gwinner E, Sonnenschein E, editors. pp Springer-Verlag, Berlin. 17

20 Alerstam T Strategies for the transition to breeding in time-selected bird migration. Ardea 94: Alerstam T Optimal bird migration revisited. Journal of Ornithology 152: S5-S23 Alerstam T, Hedenström A The development of bird migration theory. Journal of Avian Biology 29: Alerstam T, Hedenström A, Åkesson S Long-distance migration: evolution and determinants. Oikos 103: Argos Argos User s Manual. Worldwide tracking and environmental monitoring by satellite. CLS, Toulouse. Bohrer G. Brandes D, Mandel JT, Bildstein KL, Miller TA, Lanzone M, Katzner T, Maisonneuve C, Tremblay JA Estimating updraft velocity components over large spatial scales: contrasting migration strategies of golden eagles and turkey vultures. Ecology Letters 15: Brouwer J The Flyway Approach to conserving migratory birds. Its necessity and value. Report to UNEP/Convention on Migratory Species, Bonn, March Coyne MS, Godley BJ Satellite tracking and analysis tool (STAT): an integrated system for archiving, analyzing and mapping animal tracking data. Marine Ecology Progress Series 301:1 7 Gangoso L, López-López P, Grande JM, Mellone U, Limiñana R, Urios V, Ferrer M, Ecological specialization to fluctuating resources prevents long-distance migratory raptors from becoming sedentary on islands. PLoS ONE 8(4): e Gschweng M, Kalko EKV, Querner U, Fiedler W, Berthold P All across Africa: highly individual migration routes of Eleonora s falcon. Proceedings of the Royal Society of London B 275: Gschweng M, Kalko EKV, Fiedler W, Berthold P, Fahr J Multi-temporal distribution modelling with satellite tracking data: predicting responses of a long-distance migrant to changing environmental conditions. Journal of Applied Ecology 49: Hake M, Kjellen N, Alerstam T Agedependent migration strategy in Honey buzzards Pernis apivorus tracked by satellite. Oikos 103: Kassara C, Fric J, Gschweng M, Sfenthourakis S Complementing the puzzle of Eleonora s Falcon (Falco eleonorae) migration: new evidence from an eastern colony in the Aegean Sea. Journal of Ornithology 153: Kerlinger P Flight Strategies of Migrating Hawks. University of Chicago Press. Klaassen RHG, Strandberg R, Hake M, Alerstam T Flexibility in daily travel routines causes regional variation in bird migration speed. Behavioral Ecology and Sociobiology 62: Klaassen RHG, Hake M, Strandberg R, Alerstam T Geographical and temporal flexibility in the response to crosswinds by migrating raptors. Proceedings of the Royal Society of London B 278: López-López P, Limiñana R, Urios V Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zoological Studies 48: Lucia G, Agostini N, Panuccio M, Mellone U, Chiatante G, Tarini D, Evangelidis A 2011.Raptor migration at Antikythira, in southern Greece. British Birds 104: Mandel JT, Bohrer G, Winkler DW, Barber DR, Houston CS, Bildstein KL Migration path annotation: cross-continental study of migration-flight response to environmental conditions. Ecological Applications 21: Mellone U, Limiñana R, Mallìa E, Urios V, Extremely detoured migration in an inexperienced bird: interplay of transport costs and social interactions. Journal of Avian Biology 42: Morganti M, Mellone U, Bogliani G, Saino N, Ferri A, Spina F, Rubolini D Flexible tuning of departure decisions in response to weather in black redstarts migrating across the Mediterranean Sea. Journal of Avian Biology 42: Newton, I The Migration Ecology of Birds. Academic Press, London. Panuccio M, Agostini N, Mellone U Autumn migration strategies of honey buzzards, black kites, marsh and montagu's harriers over land and over water in the central Mediterranean. Avocetta 29: Panuccio M, Agostini N, Wilson S, Lucia G, Ashton-Booth J, Chiatante G, Mellone U, Todisco S Does the Honey-buzzard feed during migration? British Birds 99: Panuccio M, Mellone U, Muner L, Differential wintering area selection in Eurasian Marsh Harrier (Circus aeruginosus): a ringing recoveries analysis. Bird Study. DOI: / Rubolini D, Spina, Saino N Protandry and sexual dimorphism in trans-saharan migratory birds. Behavioral Ecology 15:

21 Sergio F, Caro T, Brown D, Clucas B, Hunter J, Ketchum J, McHugh K, Hiraldo F Top predators as conservation tools: ecological rationale, assumptions, and efficacy. Annual review of ecology, evolution, and systematics 39: 1 19 Smith NG, Goldstein DL, Bartholomew GA Is long-distance migration possible for soaring hawks using only stored fat? Auk 103: Strandberg R, Alerstam T, Hake M, Kjellén N Short-distance migration of the Common Buzzard Buteo buteo recorded by satellite tracking. Ibis 151: Strandberg R, Klaassen RHG, Olofsson P, Alerstam T Daily travel schedules of adult Eurasian Hobbies Falco subbuteo variability in flight hours and migration speed along the route. Ardea 97: Strandberg R, Klaassen RHG, Hake M, Alerstam T How hazardous is the Sahara Desert crossing for migratory birds? Indications from satellite tracking of raptors. Biology Letters 6: Thorup K, Alerstam T, Hake M, Kjellén N Bird orientation: compensation for wind drift in migrating raptors is age dependent. Proceedings of the Royal Society of London B 270: S8-S11. Thorup K, Holland RA, Tøttrup AP, Wikelski M Understanding the migratory orientation program of birds: extending laboratory studies to study free-flying migrants in a natural setting. Integrative and Comparative Biology 50: Tøttrup A., Klassen RHG, Kristensen MW, Strandberg R, Vardanis Y, Lindström Å, Rahbek C, Alerstam T, Thorup K Drought in Africa Caused Delayed Arrival of European Songbirds. Science 338: Reudink MW, Marra PP, Kyser TK, Boag PT, Langin KM, Ratcliffe LM Nonbreeding season events influence sexual selection in a long-distance migratory bird. Proceedings of the Royal Society of London B 276: Sanderson FJ, Donald PF, Pain DJ, Burfield IJ, van Bommel FPJ Long-term population declines in Afro-Palearctic migrant birds. Biological Conservation 131: Shaffer SA, Tremblay Y, Weimerskirch H, Scott D, Thompson DR, Sagar PM, Moller H, Taylor GA, Foley DG, Block BA, Costa DP Migratory shearwaters integrate oceanic resources across the Pacific Ocean in an endless summer. Proceedings of the National Academy of Sciences USA 103: Walter H Eleonora s falcon. Adaptations to prey and habitat in a social raptor. University of Chicago Press. Wink M, Ristow D 2000 Biology and molecular genetics of Eleonora s falcon Falco eleonorae, a colonial raptor of Mediterranean islands. In: Raptors at risk (ed. WWGB/Hancock House), pp Zwarts L, Bijlsma RG, van der Kamp J, Wymenga E Living on the edge: wetlands and birds in a changing Sahel. KNNV Publishing. 19

22 Section one: long-distance migration 20

23 I 21

24

25 Landscape Ecol (2010) 25: DOI /s RESEARCH ARTICLE From the Mediterranean Sea to Madagascar: Are there ecological barriers for the long-distance migrant Eleonora s falcon? Pascual López-López Rubén Limiñana Ugo Mellone Vicente Urios Received: 5 October 2009 / Accepted: 5 February 2010 / Published online: 21 February 2010 Ó Springer Science+Business Media B.V Abstract We examined the connection between landscape characteristics and behaviour of a longdistance migratory raptor. Our main goal was to test whether long-distance migratory birds adjust their migration programme according to the different characteristics of the habitats crossed during the journey with special emphasis in the so-called ecological barriers, inhospitable environments such as deserts, ice fields, seas and mountain ranges, where the opportunities to fulfil energy requirements are low or absent and environmental factors could be extremely severe. To this end, 11 Eleonora s falcons were tracked by satellite telemetry in their ca km autumn migration route from colonies located in Western Mediterranean to their wintering grounds in Madagascar during 2007 and Our results show that Eleonora s falcons migrated during day and night-time, adjusting migration speed and daily distance in relation to the crossed region. Unlike other migrant species, Eleonora s falcons did not P. López-López (&) Cavanilles Institute of Biodiversity and Evolutionary Biology, Terrestrial Vertebrates Group, University of Valencia, Polígono de la Coma s/n, Paterna, Valencia, Spain Pascual.Lopez@uv.es; lopez.pascual@gmail.com P. López-López R. Limiñana U. Mellone V. Urios Grupo de Investigación Zoología de Vertebrados, CIBIO, Universidad de Alicante, Apdo. 99, Alicante, Spain avoid ecological barriers by making unnecessary detours around them or converging on narrow corridors. Nocturnal migration and higher daily distances were observed when flying across the Sahara Desert and the Mozambique Channel. The circadian pattern of activity budget shows that Eleonora s falcon relies on an internal navigation mechanism that works during both day and night. Finally, our results suggest that the Sahara is an ecological barrier not only for passerines but also for raptors migrating within the Palaearctic-African flyway. Keywords Falco eleonorae Long-distance migration Navigation Orientation Route convergence Satellite tracking Introduction During their migratory movements between breeding and wintering ranges, birds face a variety of landscapes that can greatly affect their migration paths and schedules (Klaassen et al. 2008). Detailed understanding of the connection between landscape characteristics and behaviour of migrating birds is important in the light of current global changes. Moreover, this is particularly important in the case of long-distance migratory species for which changes in environmental conditions could affect timing of 123

26 804 Landscape Ecol (2010) 25: reproduction and migratory behaviour. In particular, ecological barriers have constrained the evolution of migration pathways. These are inhospitable environments such as deserts, ice fields, seas and mountain ranges, where the opportunities to fulfil energy requirements are low or absent and environmental factors could be extremely severe (e.g. extreme temperatures or adverse wind conditions; Newton 2008; Strandberg et al., in press). For example, there is a general agreement that the Sahara Desert is the major ecological barrier of the Palaearctic-African migration system, especially for small birds such as passerines (Schmaljohann et al. 2007), as Himalaya and Karakorum mountain ranges are for migratory birds in Asia (Combreau et al. 2009). Recent studies have suggested that the equatorial rainforest could be an obstacle for migrating falcons (Strandberg et al. 2009a). On the other hand, vast open ocean could be a migration corridor for landbirds, since this environment provides a windassisted passage relatively free of pathogens and predators (Gill et al. 2009), challenging previous hypotheses and even the physiological limits of migratory birds. In order to overcome an ecological barrier, migrating birds can choose among different strategies. Birds can either (1) make a detour to avoid crossing the barrier; (2) concentrate along routes that involve a shorter crossing; (3) try to maximise the migration speed reducing the travelling time when flying over the barrier; and even (4) the combination of the second and third scenarios. In the first scenario, a clear significant change in migration direction when approaching the barrier would be expected, whereas, in the second scenario, converging routes rather than scattered ones would be expected. In the third case, the expected behaviour would be the achievement of higher travelling speed during the barrier crossing (eventually increasing flight speed and reducing the number of stops, e.g. flying during night; Alerstam 2009), counterbalanced by more stops when flying across more suitable regions. To test these hypotheses, we used satellite-based telemetry to investigate the case of an extreme longdistance migratory raptor, the Eleonora s falcon (Falco eleonorae), in the 9000 km migration between breeding colonies in western Mediterranean Sea and wintering areas in Madagascar. Eleonora s falcon is one of the smallest bird species that is possible to track with satellite transmitters without exceeding 3% of bird s body mass (Kenward 2001), and for this reason, while there are several studies on migration of large raptors, those dealing with medium and small-sized species have been absent until recently. In addition, Eleonora s falcon crosses a huge variety of different environments such as large water bodies, deserts and dense forests that presumably could be acting as ecological barriers, and thus this species may provide interesting insights into the behavioural response of migratory birds to landscape characteristics. Materials and methods Study species The Eleonora s falcon is a cliff-nesting raptor that usually breeds on isolated small islands and feeds mainly on small birds and insects (Ferguson-Lees and Christie 2001). A unique characteristic of the species is that it adjusts its breeding season to coincide with the post-breeding autumn migration of its small passerine prey, usually in late August and early September, making it one of the latest raptor breeding seasons in the Northern Hemisphere (Walter 1979). The global population has been recently estimated between 13,000 and 14,000 breeding pairs (Dimalexis et al. 2008) after a strong decrease in population numbers in past decades, mainly due to poisoning in foraging areas, decrease of food abundance and human disturbance at colonies (Walter 1979; Dimalexis et al. 2008). Unlike bigger raptors, such as eagles and vultures that migrate by exploring thermal convection using soaring flight, the Eleonora s falcon is characterised by higher wing aspect ratio and therefore is more adapted to flapping flight (Spaar 1997). This allows Eleonora s falcons to migrate irrespective of large water bodies (Kerlinger 1989; Meyer et al. 2000) and therefore it is a good model to study landscape effects on avian migration patterns. Animal tagging, PTT programming and study area The 11 Eleonora s falcons tracked in this study were captured in the Balearic Islands in autumn 2007 and 2008, and in the Columbretes Islands in autumn 2008 (Table 1), both in Spain. Birds were trapped using 123

27 Landscape Ecol (2010) 25: Table 1 Signal transmission data and histories of 11 Eleonora s falcons fitted with satellite transmitters (PTTs) in their autumn migration from Western Mediterranean colonies to Madagascar during 2007 and 2008 Country End of transmission Distance (km) Distance/ day (km/d) Migration end Migration duration Tagging site Migration onset ID# Sex Age Weight Tagging date Male Adult /09/2007 Balearic Is. 26/10/ /12/ /01/2008 Madagascar Male Juvenile /09/2007 Balearic Is. 27/10/ /12/ /04/2008 Madagascar Male Juvenile /09/2007 Balearic Is. 24/10/2007 Spain Female Subadult /08/2008 Columbretes Is. 30/09/ /11/ /10/2009 Madagascar Female Adult /08/2008 Columbretes Is. 21/10/ /11/2008 Algeria Female Subadult /08/2008 Columbretes Is. 01/09/2008 Spain Female Adult /09/2008 Balearic Is. 20/10/ /11/ Male Adult /09/2008 Balearic Is. 21/10/ /11/ /04/2009 Algeria Female Juvenile /09/2008 Columbretes Is. 22/10/ /11/2008 Lybia Female Adult /09/2008 Columbretes Is. 20/10/ /11/ Male Juvenile /09/2008 Columbretes Is. 23/10/ /01/2009 Niger dho-gaza nets and a stuffed Eagle Owl (Bubo bubo) as a decoy (Bub 1991). All birds were weighed, measured, ringed and sexed using molecular methods. Birds were equipped with Microwave Telemetry s 9.5 g solar-powered PTT-100 platform transmitter terminals (PTTs) affixed to their backs using a Teflon harness (Kenward 2001; Limiñana et al. 2007, 2008). PTTs were programmed with a duty cycle of 8 h on/15 h off for the first 3 months of operation, and for subsequent months, the duty cycle consisted of 12 h on/58 h off as described in López- López et al. (2009). Locations were collected using the Argos system, and only locations assigned to location classes (LCs) 3, 2, 1 and 0 were used for the analyses. These LCs are a measure of reliability provided by Argos and they have nominal accuracies of \150, , and [1000 m, respectively (Argos 1996). Lower accuracy LCs (A, B) were used only when in agreement with normal travel rates (speed and direction; Strandberg et al. 2009b). Locations belonging to class Z were not used. Also, to avoid biases associated with temporal auto-correlation, positions obtained less than 1 h after the previous one were excluded from the analyses (Limiñana et al. 2007, 2008). When more than one location was available within a given hour, we used the one of highest quality. All data were retrieved and managed using the Satellite Tracking and Analysis Tool (Coyne and Godley 2005) and are publicly available at MoveBank ( Timing of migration The onset of migration was estimated to be the middle day between the last location of Eleonora s falcons in Spain and the first location in Africa, given that the birds do not perform any pre-migratory movement and depart directly to Africa from the breeding areas (López-López et al. 2009). Similarly, the date of the end of the migration was estimated as the first location of Eleonora s falcons in Madagascar. Random tracks simulations In order to test whether the observed migration routes converged towards predetermined targets, e.g. areas where a strong route convergence would be unlike to arise by chance, and thus significantly different from a direct migration route, we generated random tracks 123

28 806 Landscape Ecol (2010) 25: using observed migration parameters. This analysis followed the Strandberg et al. (2009a) rationale. To this end, we calculated the longitudinal intersections of the observed paths for each 5 latitude interval between 35 N and 15 S and calculated the mean longitude as well as the scatter in longitude (standard deviation and range) at the different latitudes. Then, the changes in longitude for each 5 longitude segments were calculated and used to generate random tracks by reshuffling these segments between the initial and the end locations at 35 N and 15 S, respectively. After each round of reshuffling the longitude at each of the latitudes (11 different latitudes in our case) for each individual was obtained. Then we calculated the longitude mean and scatter in the same way that we did for observed tracks and this procedure was repeated 1,000 times per individual. To test for non-randomness in mean longitude, the mean overall longitude and scatter for the simulated journeys was calculated for each simulation round and compared with the corresponding overall mean longitude and scatter for the observed tracks, following Strandberg et al. (2009a). Probability values for random effects were estimated from the proportion of simulations giving the same or more extreme values than those observed (see Strandberg et al. 2009a for further details on the method). To avoid possible bias in longitudinal scatter caused by nonrandomness in mean longitude and in order to detect convergences in only a single part of the migration route, simulations were also generated by dividing the journey in a single interval between 35 N and 15 S (the case explained above), two intervals, between 35 N 10 N and 10 N 15 S, and three intervals, between 35 N 15 N, 15 N 0 and 0 15 S. Migration speed and barriers To analyze if there were differences in flight speed, daily distance and time budget when crossing different habitat types, the migration routes were divided into five different regions namely Sahara, Sahel, equatorial region, SE Africa plains and Mozambique Channel (Fig. 1). Regions were defined as bands of latitudinal range as follows: (1) Sahara (from 35 N to 15 N), (2) Sahel (from 15 N to 5 N), (3) equatorial region (from 5 N to 3 S), (4) SE Africa plains (from 3 S until the start of sea crossing) and (5) the Mozambique Channel. We used successive telemetry locations to divide the real migration paths into segments and only segments between 1 and 4 h long were used for the analysis. We calculated flight speeds for each segment and following Strandberg et al. (2009b) we considered only segments with speed [5 km/h belonging to active migration (travelling), while for the others we assumed that the bird was not migrating (stationary segments). Raw data (i.e. flight speeds calculated from the real telemetry locations) showed high variation both between and within individuals in relation to regions, and therefore did not fit the canonical assumptions of standard statistical tests such as Generalized Linear Models (e.g. variance homogeneity, normal distribution of residuals). To overcome this, we performed Monte Carlo simulation analysis on individual birds (Rubinstein and Kroese 2007). To test for differences in migration speed among regions, the average speed per region was calculated with the observed data (average observed speed per region). Then, migration speeds were reshuffled within each individual and 999 simulation rounds were performed for each individual. After that, average migration speed per region was calculated for each simulation round (average simulated speed per region) and was compared with the observed average migration speed. Finally P-values were estimated as the proportion of simulations giving the same or more extreme values as observed. The null hypothesis assumed that there were no differences between observed and simulated speeds. Because the alternative hypothesis was non-directional a two tailed probability distribution was used. The threshold of significance was set at a = Daily distance across the different regions was calculated for each bird, dividing the distance between the first and last locations in each region by the elapsed hours, and then multiplying by 24 h. The number of daily travelling hours were also estimated by dividing the mean speed of migration on travelling days (km/24 h; based on locations between different days) by mean short-interval speeds (km/h; including only segments [5 km/h; Strandberg et al. 2009b). Circadian patterns of migration To analyze migratory behaviour in relation to the time of day we used only intervals shorter than 4 h (Limiñana et al. 2007; Strandberg et al. 2009b), with 123

29 Landscape Ecol (2010) 25: Fig. 1 Autumn migration routes of six Eleonora s falcons (Falco eleonorae) tracked by satellite telemetry from their breeding colonies in the Western Mediterranean to Madagascar during 2007 and Route of adult birds shown with solid line and the juvenile with dashed line. Desert and equatorial rainforest regions are highlighted in two shades of grey (adapted from Olson et al. 2001). The main geographic regions used for analyses are shown (see text for details) 123

30 808 Landscape Ecol (2010) 25: each interval assigned to an hour according to local time at the end of the segment. Local time was calculated by correcting GMT times according to each time zone. Nocturnal segments were those for which at least half the time length occurred after sunset or before sunrise. The exact time of sunrise and sunset for every location was obtained from the website Segments were considered either travelling or stationary according to the same criteria mentioned before (stationary if migration speed was \5 km/h) (Strandberg et al. 2009b). Differences in travel rates among regions were tested by means of contingency tables. Results After trapping and marking 11 Eleonora s falcons, we obtained six complete autumn migration routes from the Western Mediterranean breeding colonies to wintering areas in Madagascar, corresponding to four adults, one subadult (second calendar year) and one juvenile (Fig. 1). Detailed migration parameters, bird histories and signal transmission data are shown in Table 1. The comparison of the observed routes with randomly simulated ones did not show significant differences that would suggest convergence towards goal areas or travelling along narrow corridors (Fig. 2). Similar results were obtained when excluding the juvenile from analyses and when dividing the journey in two or three intervals (all tests nonsignificant, Table 2). Daily distances differed among regions (Kruskal Wallis test: H 4,30 = 16.18, P = 0.003), with the highest value observed in the Mozambique Channel and the lowest in the Sahel (Fig. 3). The comparison of observed speeds with expected speeds among different regions showed a high variation within individuals, with results being significant in 11 cases (Table 3). Observed speeds were lower than expected in the Sahel (four cases) and in the equatorial region (two cases), while higher speeds were observed in the Sahara and in the Mozambique Channel, respectively. The average number of daily travelling hours was 11.3 ± 4.6 h (range: 5 24), with higher values in the Mozambique Channel (24 h) and the Sahara (12.7 ± 6.6 h), and the lowest in the Sahel region (9.2 ± 3.2 h). Birds migrated during both day and night and within all regions (Fig. 4), although during night-time there Fig. 2 Comparison between a mean longitude and b scatter of longitudes (measured as standard deviation), for the six observed routes of Eleonora s falcons (connected by a solid line with triangles) and 1,000 simulated random tracks (dashed line with circles). Values were calculated at 5 latitude intervals were differences in the number of travelling segments among regions (Fig. 5; v 2 = 29.67, d.f. = 3, P \ 0.001). The nocturnal travel rate was higher in Sahara than in the Sahel (v 2 = 24.1, d.f. = 1, P \ 0.001). However, no differences occurred among the equatorial region and SE Africa plains (v 2 = 2.01, d.f. = 1, n.s.). During day-time, no differences occurred among the four regions (v 2 = 3.4, d.f. = 3, n.s.) but interestingly, of the six Saharan segments, none was a stationary one. Discussion Our results show that Eleonora s falcons tracked from Western Mediterranean islands migrate during day and night-time, travelling through inland Africa until reaching the wintering areas in Madagascar. Similar inland routes have been reported for Eleonora s falcons tracked from other Mediterranean colonies 123

31 Landscape Ecol (2010) 25: Table 2 Comparison between observed migration parameters (mean longitude and scatter in longitude measured as standard deviation) and a set of 1,000 simulated random tracks per animal Latitudinal range Individuals Mean longitude observed Mean longitude simulated Mean s.d. longitude observed Mean s.d. longitude simulated Number of random simulations P-value longitude P-value s.d. longitude 1 interval 35 N 15 S All N 15 S Adults intervals 35 N 10 N All N 15 S All N 10 N Adults N 15 S Adults intervals 35 N 15 N All N 0 All S All N 15 N Adults N 0 Adults S Adults To analyze for convergence in the migratory route, comparisons were made between the migration route as one single interval, and by dividing into two and three latitudinal intervals. Comparisons were also made including all individuals together (juvenile, subadults and adults) or only adults Fig. 3 Daily distance covered across the five main regions during autumn migration of six Eleonora s falcons tracked by satellite telemetry. Median, 25 and 75% percentiles and maximum and minimum data are shown either in Sardinia (Gschweng et al. 2008) or Greece ( satellite_map.php; unpubl. data). The Eleonora s falcons tracked in this study migrated through the Sahara Desert, the Sahelian region and the equatorial rainforest without making any detour to avoid those regions, until finally converging in SE Africa (Tanzania and Mozambique) just before crossing the Mozambique Channel. Despite individual variation, our analysis did not discover the existence of narrow migration corridors through ecological barriers (e.g. the Sahara Desert or the equatorial rainforest; Berthold 2001; Strandberg et al. 2009a). In fact, the only apparent convergence occurred in the final part of the route, which leads to the shortest route between mainland Africa and Madagascar. Our results are different than those obtained by Strandberg et al. (2009a) for the Eurasian hobby (Falco subbuteo) using the same random tracks simulation analysis, despite the close taxonomic relationship between the two species and their similar food habits (Ferguson-Lees and Christie 2001). In their migration from Northern Europe to Southern Africa, Eurasian hobbies converged in a narrow corridor after crossing the equatorial rainforest, suggesting that the rainforest acts as an ecological barrier for migratory birds, perhaps related to reduced feeding opportunities at this habitat (Strandberg et al. 2009a). Commenting on the results of satellite tracking of Eleonora s falcons breeding in Sardinia, the same authors hypothesized that this could also happen for Eleonora s falcons because the adults partially avoided the rainforest, migrating along the eastern border, and juveniles changed their direction abruptly after beginning the crossing of the rainforest (Gschweng et al. 2008). However, in our study, four of six adults crossed the equatorial rainforest, and 123

32 810 Landscape Ecol (2010) 25: Table 3 Comparison between observed migration speeds (in km/h) with randomly simulated expected speeds after 999 reshufflings for the five main regions crossed by satellite tracked Eleonora s falcons Significant results (P \ 0.05) are highlighted in bold Tag ID Region Average speed observed Average speed simulated P-value Sahara Sahel Equatorial rainforest SE Africa plains Mozambique Channel Sahara Sahel Equatorial rainforest SE Africa plains Mozambique Channel Sahara Sahel Equatorial rainforest SE Africa plains Mozambique Channel x x x Sahara x x x Sahel Equatorial rainforest SE Africa plains Mozambique Channel Sahara Sahel Equatorial rainforest SE Africa plains Mozambique Channel x x x Sahara Sahel Equatorial rainforest SE Africa plains Mozambique Channel Fig. 4 Time budget of migratory Eleonora s falcons in relation to hour of the day. Segments were divided as travelling (if migration speed exceeded 5 km/h) or stationary (migration speed \5 km/h). Data on all regions are pooled. The number of segments is indicated on the bars 123

33 Landscape Ecol (2010) 25: Fig. 5 Daily time budget of Eleonora s falcons in relation to the five main regions crossed during autumn migration from the Western Mediterranean to Madagascar. The number of segments is indicated on the bars those that avoided the rainforest were already migrating using a more easterly route. In contrast to Eurasian hobbies that winter in open habitats (Ferguson-Lees and Christie 2001, see also the map in Strandberg et al. 2009a), Eleonora s falcons live in rainforests in Madagascar during the winter, suggesting that this environment does not represent a real barrier for the species. Daily distance travelled varied among regions, due to changes in flight speed and time budget. Although is difficult to unravel a fixed pattern in these changes, it is interesting to note that the lowest speeds occurred in the same region, the Sahel, where the nocturnal travel rate was also lower, probably to counterbalance the extreme energetic consumption previously experienced during the crossing of the Sahara Desert. In fact, during the Sahara crossing Eleonora s falcons increased their nocturnal travel rate and did not stop during daytime. Locations over favourable habitats such as oases, or with certain topographic attributes (e.g. valleys) were not observed along the route. Taking into account that feeding opportunities are quite low in desert landscapes, the absence of stationary daylight segments suggest that the birds avoid foraging in this environment and try to spend as much time as possible in active migration in order to overcome the barrier as soon as possible. This is in agreement with our third scenario, for which birds would try to maximise the migration speed when flying over an ecological barrier (Alerstam 2009). Similar results have been reported for other raptor species whose migration have been investigated by mean of satellite telemetry, such as the osprey (Pandion haliaetus), honey buzzard (Pernis apivorus), marsh harrier (Circus aeruginosus) and hobby (Kjellén et al. 2001; Hake et al. 2003; Klaassen et al. 2008; Strandberg et al. 2009b, in press). These species also reached faster speeds during the Sahara crossing, exceeding the speeds recorded in other regions. The absence of differences for these species in travel rates between the equatorial region, where the main environment is the rainforest, and SE Africa, where open savannahs dominate the landscape, suggests that for these birds the rainforest does not represent an ecological barrier. The circadian pattern of activity budget showed a consistent high rate of nocturnal migration throughout all regions. Nocturnal flights of migrating raptors have been reported mostly over water (see De Candido et al and references therein), presumably because thermal updrafts are absent during both day and night, resulting in no differences in energy expenditure (Alerstam 2009). Regular nocturnal migration over land has been only shown for the Levant sparrowhawk (Accipiter brevipes) (Stark and Liechti 1993; Spaar et al. 1998) and less frequently for the Eurasian hobby (Strandberg et al. 2009b). This behaviour is more typical of other groups of migrating birds, especially passerines (Berthold 2001; Alerstam 2009). These findings suggest that these species, now including Eleonora s falcon, rely on an internal navigation mechanism that works during both day and night. Finally, our results support the hypothesis that the Sahara is an ecological barrier not only for passerines (Berthold 2001; Schmaljohann et al. 2007) but also for raptors migrating within the Palaearctic-African flyway. The crossing of the Sahara has a profound influence on survival and fitness of migrants (Strandberg et al., in press; García-Ripollés et al., in press) and, for this reason, an increase in migratory speed during the Sahara crossing is important in order to minimize the associated risks, such as starvation, disorientation, unfavourable weather conditions and even sand storms. Given that Eleonora s falcon abruptly increase their energetic consumption during 123

34 812 Landscape Ecol (2010) 25: the crossing of desert landscapes, the lower travel rates observed in the Sahel region just after having crossed the Sahara desert may reflect a strategy to replenish their energy reserves in a more productive environment. Our study also shows that long-distance migratory birds adjust their migration activity according to the different landscapes crossed during the journey, but that the response differs among individuals. In the light of the rapid shift of world biomes due to global change (Williams et al. 2007), detailed understanding of the connection between landscape characteristics and behaviour of long-distance migratory birds is of utmost importance. This is especially important in the case of long-migrant species crossing such a great variety of environments as Eleonora s falcon do, for which small changes in environmental conditions could have unexpected consequences that could jeopardize timing of reproduction and even the survival of maladjusted birds (carry-over effects), given the mismatch between migration schedules and food availability (Both et al. 2006). Forecasting how global changes will shape the future behaviour of migratory birds constitutes the next challenge. Acknowledgements The Terra Natura Foundation and the Conselleria de Medi Ambient, Aigua, Urbanisme i Habitatge of the Generalitat Valenciana financed this project. Special thanks are due to J. Jiménez and J.V. Escobar of the Servicio de Biodiversidad of the regional government. We would like to thank J. De la Puente, A. Bermejo, E. Escudero (SEO-Monticola), J.L. Martínez (GOB), M. Suárez (GOB) and T. Muñoz (GOB) who helped in trapping some Eleonora s falcons in Balearic Islands and V. Ferrís, E. Sánchez, B. Sarzo, M.A. Bartolomé and C. García who helped us in Columbretes Islands. The Conselleria de Medi Ambient of the Govern Balear kindly gave permission to trap falcons in Balearic Islands, and special thanks are due to J. Mayol and J. Muntaner. J. García, sexed the falcons, and blood samples were provided by Ll. Parpal of the Centre de Recuperació de Fauna de Balears. L.M. Carrascal, T. Alerstam and R. Strandberg kindly gave us statistical advice and helped us revise a previous draft of the manuscript. We also thank two anonymous referees and E. Gustafson for valuable comments on earlier versions of this manuscript. P. López-López and U. Mellone are supported by FPU grants of the Spanish Ministry of Science and Innovation (references AP and AP ). This paper is part of the Ph.D. of U. Mellone and complies with the current laws in Spain. References Alerstam T (2009) Flight by night or day? Optimal daily timing of bird migration. J Theor Biol 258: doi: /j.jtbi Argos (1996) User s manual. CLS/Service Argos, Toulouse Berthold P (2001) Bird migration. A general survey. Oxford University Press, New York Both C, Bouwhuis S, Lessells CM, Visser ME (2006) Climate change and population declines in a long-distance migratory bird. Nature 44:81 83 Bub H (1991) Bird trapping and bird banding. Cornell University Press, Ithaca Combreau O, Judas J, Lawrence M (2009) Importance of satellite tracking in Houbara Bustard conservation program in Asia. In: Proceedings of the 2009 bird fish tracking conference. Microwave Telemetry Inc., Elicott City Coyne MS, Godley BJ (2005) Satellite tracking and analysis tool (STAT): an integrated system for archiving, analyzing and mapping animal tracking data. Mar Ecol Prog Ser 301:1 7 De Candido R, Bierregaard RO, Martell MS Jr, Bildstein KL (2006) Evidence of nocturnal migration by Osprey (Pandion haliaetus) in North America and Western Europe. J Raptor Res 40: Dimalexis A, Xirouchakis S, Portolou D, Latsoudis P, Karris G, Fric J, Georgiakakis P, Barboutis C, Bourdakis S, Ivovič M, Kominos T, Kakalis E (2008) The status of Eleonora s falcon (Falco eleonorae) in Greece. J Ornithol 149: doi: /s Ferguson-Lees J, Christie DA (2001) Raptors: birds of prey of the world. A & C Black Publishers, Ltd., London García-Ripollés C, López-López P, Urios V (in press) First description of migration and wintering of adult Egyptian vultures Neophron percnopterus tracked by GPS satellite telemetry. Bird Study Gill RE, Tibbitts TL, Douglas DC, Handel CM, Mulcahy DM, Gottschalck JC, Warnock N, McCaffery BJ, Battley PF, Piersma T (2009) Extreme endurance flights by landbirds crossing the Pacific Ocean: ecological corridor rather than barrier? Proc R Soc B 276: Gschweng M, Kalko EKV, Querner U, Fiedler W, Berthold P (2008) All across Africa: highly individual migration routes of Eleonora s falcon. Proc R Soc B 275: doi: /rspb Hake M, Kjellén N, Alerstam T (2003) Age-dependent migration strategy in honey buzzards Pernis apivorus tracked by satellite. Oikos 103: Kenward R (ed) (2001) A manual for wildlife radio tagging. Academic Press, London Kerlinger P (1989) Flight strategies of migrating hawks. University of Chicago Press, Chicago Kjellén N, Hake M, Alerstam T (2001) Timing and speed of migration in male, female and juvenile Ospreys Pandion haliaetus between Sweden and Africa as revealed by field observations, radar and satellite tracking. J Avian Biol 32:57 67 Klaassen RHG, Strandberg R, Hake M, Alerstam T (2008) Flexibility in daily travel routines causes regional variation in bird migration speed. Behav Ecol Sociobiol 62: Limiñana R, Soutullo A, Urios V (2007) Autumn migration of Montagu s harriers Circus pygargus tracked by satellite telemetry. J Ornithol 148: Limiñana R, Soutullo A, López-López P, Urios V (2008) Premigratory movements of adult Montagu s Harriers Circus pygargus. Ardea 96:

35 Landscape Ecol (2010) 25: López-López P, Limiñana R, Urios V (2009) Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool Stud 48: Meyer SK, Spaar R, Bruderer B (2000) To cross the sea or to follow the coast? Flight directions and behaviour of migrating raptors approaching the Mediterranean Sea in autumn. Behaviour 137: doi: / Newton I (2008) The migration ecology of birds. Academic Press, London, pp Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND, Powell GVN, Underwood EC, D Amico JA, Itoua I, Strand HE, Morrison JC, Loucks CJ, Allnutt TF, Ricketts TH, Kura Y, Lamoreux JF, Wettengel WW, Hedao P, Kassem KR (2001) Terrestrial ecoregions of the world: a new map of life on earth. Bioscience 51: doi: / (2001)051[0933:TEOTWA]2.0.CO;2 Rubinstein RY, Kroese DP (2007) Simulation and the Monte Carlo method, 2nd edn. Wiley, New York Schmaljohann H, Liechti F, Bruderer B (2007) Songbird migration across the Sahara: the non-stop hypothesis rejected!. Proc R Soc B 274: Spaar R (1997) Flight strategies of migrating raptors; a comparative study of interspecific variation in flight characteristics. Ibis 139: Spaar R, Stark H, Liechti F (1998) Migratory flight strategies of Levant sparrowhawks: time or energy minimization? Anim Behav 56: Stark H, Liechti F (1993) Do Levant sparrowhawks Accipiter brevipes also migrate at night? Ibis 135: Strandberg R, Klaassen RH, Hake M, Olofsson P, Alerstam T (2009a) Converging migration routes of Eurasian hobbies Falco subbuteo crossing the African equatorial rain forest. Proc R Soc B 276: doi: /rspb Strandberg R, Klaassen RHG, Olofsson P, Alerstam T (2009b) Daily travel schedules of adult Eurasian Hobbies Falco subbuteo variability in flight hours and migration speed along the route. Ardea 97: Strandberg R, Klaassen RHG, Hake M, Alerstam T (in press) How hazardous is the Sahara Desert crossing for migratory birds? Indications from satellite tracking of raptors. Biol Lett. doi: /rsbl Walter H (ed) (1979) Eleonora s falcon. Adaptations to prey and habitat in a social raptor. The University of Chicago Press, Chicago Williams JW, Jackson ST, Kutzbach JE (2007) Projected distributions of novel and disappearing climates by 2100 AD. Proc Natl Acad Sci USA 104:

36 II 34

37 Int J Biometeorol (2011) 55: DOI /s ORIGINAL PAPER Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor Ugo Mellone & Pascual López-López & Rubén Limiñana & Vicente Urios Received: 22 June 2010 / Revised: 10 September 2010 / Accepted: 13 September 2010 / Published online: 29 September 2010 # ISB 2010 Abstract Weather conditions are paramount in shaping birds migratory routes, promoting the evolution of behavioural plasticity and allowing for adaptive decisions on when to depart or stop during migration. Here, we describe and analyze the influence of weather conditions in shaping the sea-crossing stage of the pre-breeding journey made by a long-distance migratory bird, the Eleonora s falcon (Falco eleonorae), tracked by satellite telemetry from the wintering grounds in the Southern Hemisphere to the breeding sites in the Northern Hemisphere. As far as we know, the data presented here are the first report of repeated oceanic journeys of the same individuals in consecutive years. Our results show inter-annual variability in the routes followed by Eleonora s falcons when crossing the Strait of Mozambique, between Madagascar and eastern continental Africa. Interestingly, our observations illustrate that individuals show high behavioural plasticity and are able to change U. Mellone (*) : R. Limiñana : V. Urios Grupo de Investigación Zoología de Vertebrados, University of Alicante, Apdo. 99, Alicante, Spain ugomellone@libero.it R. Limiñana ruben.lm@gimail.com V. Urios vicenteurios@yahoo.es P. López-López Cavanilles Institute of Biodiversity and Evolutionary Biology, Terrestrial Vertebrates Group, University of Valencia, C/ Catedrático José Beltrán 2, Paterna( Valencia, Spain lopez.pascual@gmail.com R. Limiñana Instituto de Investigación en Recursos Cinegéticos (CSIC-UCLM-JCCM), Ronda de Toledo s/n., Ciudad Real, Spain their migration route from one year to another in response to weather conditions, thus minimising the risk of long ocean crossing by selecting winds blowing towards Africa for departure and changing the routes to avoid low pressure areas en route. Our results suggest that weather conditions can really act as obstacles during migration, and thus, besides ecological barriers, the migratory behaviour of birds could also be shaped by meteorological barriers. We briefly discuss orientation mechanisms used for navigation. Since environmental conditions during migration could cause carry-over effects, we consider that forecasting how global changes of weather patterns will shape the behaviour of migratory birds is of the utmost importance. Keywords Argos. Eleonora s falcon. Migration. Satellite tracking. Wind. Vorticity Introduction Weather conditions are paramount in shaping birds migratory routes (Richardson 1978; Elkins 1988; Liechti 2006; Hedenström 2010). The severe fitness cost of longdistance migration has promoted the evolution of behavioural plasticity allowing for adaptive decisions on when to depart or stop in relation to local weather conditions (Alerstam 1981; and see, e.g. Saino et al. 2010). This generates a strong selective pressure for the selection of optimal wind conditions for departure and travelling, which are particularly important when birds have to face large ecological barriers (Delingat et al. 2008; Navedo et al. 2010; Saino et al. 2010), inhospitable environments such as deserts, seas and mountain ranges, where the opportunities to fulfil energy requirements are low or absent and environmental factors could be extremely severe, increasing mortality risks (Kerlinger 1989; Newton 2008; Strandberg et al. 2009; López-López et al. 2010). For example, it has

38 464 Int J Biometeorol (2011) 55: been shown that Western honey buzzards (Pernis apivorus) change their pathways when crossing the Mediterranean sea according to prevailing winds (Agostini et al. 2005; 2007), and that the departure of Bar-tailed godwits (Limosa lapponica) from Alaska when crossing the Pacific ocean in a 11,000 km non-stop flight was associated with tailwinds (Gill et al. 2009). Nevertheless, data on the impact of harsh weather conditions on migration routes are scarce. It is well known that water crossings lead to increased mortality risk (Kerlinger 1989) and could even result in episodes of massive mortality [e.g. more than 1,300 raptors were found dead along a beach of the Mediterranean coast of Israel during April 1980, as reported by Zu-Aretz and Leshem (1983)]. However, this type of risk has been somewhat under-estimated in the migration literature (Liechti 2006). Here, we describe the sea-crossing stage of the journey made by a long-distance migratory bird, the Eleonora s falcon (Falco eleonorae), in its pre-breeding migration from the wintering grounds in the Southern Hemisphere to the breeding sites in the Northern Hemisphere. In particular, we analyse the influence of weather conditions in shaping the oceanic route between Madagascar and continental Africa. We focus only on this part of the journey since raptors cannot stop when migrating over water, and hence weather conditions are expected to act more selectively here than when migrating over land. To our knowledge, these are the first data ever reported concerning repeated oceanic journeys by the same individuals in consecutive years. Hence, they provide an excellent benchmark to study behavioural plasticity on an individual basis in response to changing weather conditions across an ecological barrier. Materials and methods Study species The Eleonora s falcon is a cliff-nesting raptor that breeds mainly on small Mediterranean islands in late August and early September, adjusting its breeding season to coincide with the post-breeding migration of small passerines, its main prey during reproduction (Walter 1979). This longdistance migratory species winters in Madagascar and travels more than 18,000 km every year for breeding and wintering (Ferguson-Lees and Christie 2001). On the basis of satellite tracking data, migration routes have been recently discovered (Gschweng et al. 2008; López-López et al. 2009). The Eleonora s falcon has narrow wings and therefore is characterised by high aspect ratio (calculated as the quotient between wingspan and wing area; Kerlinger 1989), being adapted to flapping flight instead of soaring. Hence, it can migrate irrespective of large water bodies, and is more flexible than soaring species in changing route, providing a good model to study the influence of weather conditions on avian migration. Animal tagging We captured and tagged 11 Eleonora s falcons in the breeding colonies located in Balearic and Columbretes Islands (Spain) in autumn 2007 and Birds were trapped using dho-gaza nets and a stuffed Eagle Owl (Bubo bubo) as a decoy. All birds were equipped with Microwave Telemetry s 9.5 g solar-powered PTT-100 platform transmitter terminals, accounting for less than 3% of their body mass and affixed to their backs using a Teflon harness (Kenward 2001). Locations were collected using the Argos system (Argos 2008). Detailed data on animal histories, tagging methods and data retrieval are available in López- López et al. (2010). Hours are given according to local time. Wind data were downloaded from the NCEP/ NCAR Reanalysis project ( gridded/data.ncep.reanalysis.pressure.html). The reanalysis data were created using various assimilated observations/ measurements and atmospheric models, and this dataset includes a temporal coverage of observation four times per day with a spatial resolution of 2.5 x 2.5 from 1948 until the present (Kalnay et al. 1996). Data analyses On the basis of satellite telemetry data, we calculated the position of each bird at any given hour when wind data were available (every 6 hours) by interpolating locations and using as a reference the speed of each bird when migrating over the ocean. This speed was calculated by dividing the distance between the first and the last location recorded over the sea by the elapsed time. Speed data during the sea crossing was obtained for all birds but one (bird #80400); thus, for this bird, we used the lowest speed value recorded for the other birds (60 km/h) for the estimation of interpolated locations along the routes. At these interpolated locations, we collected U (West-East component) and V (South-North component) wind vectors at 850 mb, corresponding approximately to 1,300 m altitude above sea level, from which we inferred wind direction and wind speed. It has been shown by means of radar tracking that the chosen value of 850 mb roughly corresponds to the average value of a falcon s altitude when migrating over sea (Meyer et al. 2000). The data of the NCEP/ NCAR Reanalysis project have been used widely to analyse wind influence on Argossatellite-tracked birds (Shamoun-Baranes et al. 2003; Thorup et al. 2003; Klaassenetal.2010). Wind directions were analysed using circular statistics (Batschelet 1981).

39 Int J Biometeorol (2011) 55: We compared 2009 vs 2010 average wind direction measurements during the ocean crossing by means of the Watson Williams test for two samples (Zar 1984). To this end, we pooled the data of all individuals within each year (three in 2009 and two in 2010). In addition, we inspected maps of relative vorticity (downloaded from the CIMSS Tropical Cyclone website at: at the same pressure level (850 mb) of wind data, which are available for every 3 h. Vorticity is a measure of the spin of an air mass such as a low- or high-pressure weather system (Batchelor 2000). In the Southern Hemisphere, cyclonic rotation is associated with negative vorticity, thus coinciding with troughs of low pressure areas associated with cyclones or storms, which are unfavourable conditions for migrating birds (Richardson 1978). Results Overall, we obtained five spring ocean tracks belonging to three adult individuals (Fig. 1). In 2009, the three individuals migrated directly from Madagascar to Somalia, flying non-stop for about 1,500 km. The maximum longitudinal distance among tracks was 195 km. Birds started the sea crossing in a 14-h time window as follows: the first female (transmitter #80399) started migration at 1251 hours on 11 April, followed by the second female (transmitter #80400), which started migration at 1935 hours on 11 April; and finally, the male (transmitter #80402), which started migration at 0250 hours on 12 April. Flight speeds during sea crossing were 60 km/h (during 10.4 h; bird #80399) and 63.1 km/h (during 2.2 h; bird #80402). In 2010, two of the three individuals tracked the previous year (since bird #80400 stopped transmission on 30 April 2009 when it was in Algeria) migrated towards Tanzania for about 1,200 km, keeping a much more westward course than that recorded in Eleonora s falcons departed from Madagascar at 0230 hours on 13 April (transmitter #80399), and at 0435 hours on 14 April (transmitter #80402). The maximum longitudinal distance recorded between tracks during sea crossing was 78 km, and two locations belonging to different individuals were recorded almost 26 h apart each other in a narrow funnel of just 22 km. Flight speeds during sea crossing were 60.5 km/ h (during 6.7 h; #80399) and 72.6 km/h (during 7.1 h; #80402). Bird #80402 is still transmitting on September 2010 while the transmitter #80399 stopped working in Egypt on 27 April Wind directions during the whole ocean crossing differed between consecutive years, being more westward and concentrated during 2010 than 2009 (wind direction 2009 =36.3, SD=74.3 ; wind direction 2010 = 94.5, SD=1.9 ; Watson Williams test for two samples; W 15,7 =0.1568, P<0.05). Wind direction recorded at the onset of migration were strongly westward-directed in both years (mean 94.1, SD=2.2, N=5). Wind speed during oceanic crossing averaged 35 km/h for the two years (SD=48.9, N=22). The vorticity patterns at the onset of migration of the first bird beginning the oceanic crossing each year are shown in Fig. 1. Interestingly, two troughs of low pressure corresponding to negative values of relative vorticity were located in the northern Mozambique Channel just between Madagascar and Africa in 2009, forcing birds to fly northwards and acting as a barrier for a more direct route towards Tanzania (Fig. 1a). In contrast, in 2010 the onset of migration coincided just when the Mozambique Channel was free of troughs of low pressure, allowing birds to reach the African coast in a shorter route from northern Madagascar to the coast of Tanzania (Fig. 1b). Discussion Satellite telemetry allows us to gain new insights into avian migration ecology, especially regarding relevant topics such as birds response to environmental conditions as well as orientation during migration (Thorup et al. 2003; Gill et al. 2009; Hedenström 2010). Bird migration shows a high degree of plasticity at both individual and evolutionary levels (Alerstam et al. 2006; Pulido and Berthold 2010). As far as we know, the data presented here are the first report of repeated oceanic journeys of the same individuals in consecutive years. Our results show inter-annual variability in the routes followed by Eleonora s falcons when crossing the Strait of Mozambique between Madagascar and eastern continental Africa. Interestingly, individuals showed high route flexibility between years, migrating independently from each other following narrow funnels, and changing route according to weather conditions. In both years, birds started sea crossing in a short time window of 1 2 days, and data gathered during this stage suggest that they did not migrate in the same group. Recorded flight speeds suggest tailwind assistance, since the values of flight speed recorded for this species by radar tracking were only 44 km/h (Rosén et al. 1999). In both years, wind conditions at migration onset were always favourable to cross en route for Africa, thus suggesting a fine-tuned selection for the best departure time in order to minimize energy expenditure during open ocean crossing. However, wind assistance en route differed between years, being lower in 2009 than In 2009 birds used a longer flyway and were forced to fly directly towards Somalia instead of landing on Africa in advance through a shorter pathway, in contrast to what they did in 2010, probably in order to avoid the low

40 466 Int J Biometeorol (2011) 55: Fig. 1 Vorticity patterns at the onset of migration of the first Eleonora s falcon (Falco eleonorae) tracked by satellite telemetry in a 2009 and b Red solid lines Individual tracks, red dots satellite telemetry locations. Negative vorticity areas (yellow) coincide with troughs of low pressure, whereas positive vorticity areas (blue) correspond to areas of stable meteorological conditions (see Materials and methods for details of data sources) pressure area between Madagascar and Africa (Fig. 1a). Alternatively, they would have crossed the area with the highest vorticity, where weather changes can be sudden and less predictable, with stronger winds and hazardous weather such as rain and storms, which are especially dangerous during the crossing of ecological barriers (Strandberg et al. 2009). Consequently, our observations illustrate that Eleonora s falcons show high behavioural plasticity and are able to change migration route from one year to another in response to weather conditions, thus minimising the risk of long ocean crossing by selecting winds blowing towards Africa for departure and avoiding low pressure areas en route. A literature review showed that there is only one paper analysing the sea crossing behaviour of a satellitetracked bird facing extreme weather. This paper showed how a Peregrine falcon (Falco peregrinus), after having selected tailwind conditions to cross the Gulf of Mexico, met a hurricane during the crossing, retreated from it, being forced to go back to the mainland and, after having waited the hurricane to pass, completed the sea crossing successfully in a second attempt (McGrady et al. 2006). In a review of worldwide patterns of raptor migration, Bildstein (2006) stated that truly oceanic migration remains an exceptional form of raptor migration. Indeed, the 1,500 km journey observed in 2009 is, to our knowledge, one of the longest successful non-stop open water journeys ever reported for a raptor. Some species, such as the Merlin (Falco colombarius), have shown to be able to perform successfully an overwater flight of about 1,000 km from Iceland to Great Britain (Bildstein 2006), and a juvenile Oriental honey buzzard (Pernis ptilorhyncus) was tracked from Japan to China (Higuchi et al. 2005). Other raptors, such as the Amur falcon (Falco amurensis), are thought to take an even longer overwater route, from India to East Africa (Bildstein 2006). Remote tracking systems such as satellite telemetry or geolocators allow researchers to unravel the migratory routes of birds (Felicísimo et al. 2008; Gschweng et al.

41 Int J Biometeorol (2011) 55: ; López-López et al. 2009; Stutchbury et al. 2009). In addition, they provide great insights into a bird s capabilities of selecting optimal weather conditions, and serve to suggest what orientation systems might be used to accomplish their journeys. In our case, the fact that all falcons migrated during both night- and day-time (see also López-López et al. 2010), navigating across oceanic landscapes without any conspicuous target, suggest that neither their navigation skills nor their ability to compensate for wind displacement can rely on any system that simply involves landmark cues. Thus, the true orientation mechanism that Eleonora s falcons use remains unknown. Direction cues obtained from celestial objects and their related skylight polarization pattern, combined with some time-keeping mechanisms, the Earth s magnetic field or even the use of olfactory cues or sea waves as a reference are the most likely candidates (Alerstam and Petterson 1976; Akesson and Hedenstrom 2007). We have shown that weather conditions can really act as obstacles during migration, and thus, besides ecological barriers (Newton 2008; e.g. Strandberg et al. 2009, López- López et al. 2010), the migratory behaviour of birds could also be shaped by meteorological barriers. Since environmental conditions during migration could cause carryover effects (Strandberg et al. 2009), forecasting how global changes of weather patterns will shape the behaviour of migratory birds is of the utmost importance. In the near future, high-resolution detailed tracks will hopefully allow us to unravel finer-scale route deviations and to study more thoroughly orientation mechanisms, as well as bird s behavioural response to changes in local conditions in a changing world (Robinson et al. 2009). Acknowledgements The Terra Natura Foundation and the Servei de Biodiversitat of the ConselleriadeMediAmbient (Generalitat Valenciana) financed this project. Special thanks are due to J. Jiménez, J.V. Escobar, J. Mayol and J. Muntaner. J. De la Puente, A. Bermejo, E. Escudero (SEO-Monticola), J.L. Martínez, M. Suárez, T. Muñoz (GOB), V. Ferrís, E. Sánchez, B. Sarzo and M. A. Bartolomé helped in trapping activity. Two anonymous referees made valuable suggestions that improved the paper. P.L.-L. and U. M. are supported by FPU grants of the Spanish Ministry of Education (references AP and AP ). This paper is part of the PhD dissertation of U.M. and it complies the current laws in Spain. References Agostini N, Premuda G, Mellone U, Panuccio M, Logozzo D, Bassi E, Cocchi L (2005) Influence of wind and geography on orientation behavior of adult Honey Buzzards Pernis apivorus during migration over water. Acta Ornithol 40:71 74 Agostini N, Cardelli C, Gustin M (2007) Factors shaping pathways of European Honey-buzzards (Pernis apivorus) during spring migration in the Central Mediterranean basin. J Raptor Res 41:57 61 Akesson S, Hedenström A (2007) How migrants get there: migratory performance and orientation. BioScience 57: Alerstam T (1981) The course and timing of bird migration. In: Aidley DJ (ed) Animal migration. Cambridge University Press, Cambridge, pp 9 54 Alerstam T, Petterson SG (1976) Do birds use waves for orientation when migrating across the sea? Nature 259: Alerstam T, Hake M, Kjellén N (2006) Temporal and spatial patterns of repeated migratory journeys by ospreys. Anim Behav 71: Argos (2008) User s manual. CLS/Service Argos, Toulouse Batchelor GK (2000) An introduction to fluid dynamics. Cambridge University Press, Cambridge Batschelet E (1981) Circular statistics in biology. Academic, London Bildstein KL (2006) Migrating raptors of the world. Cornell University Press, Ithaca Delingat J, Bairlein F, Hedenström A (2008) Obligatory barrier crossing and adaptive fuel management in migratory birds: the case of the Atlantic crossing in Northern Wheatears (Oenanthe oenanthe). Behav Ecol Sociobiol 62: Elkins N (1988) Weather and bird behaviour. Poyser, Calton Felicísimo ÁM, Muñoz J, González-Solis J (2008) Ocean surface winds drive dynamics of transoceanic aerial movements. PLoS ONE 3(8):e2928. doi: /journal.pone Ferguson-Lees J, Christie DA (2001) Raptors: birds of prey of the world. Black, London Gill RE, Tibbitts TL, Douglas DC, Handel CM, Mulcahy DM, Gottschalck JC, Warnock N, McCaffery BJ, Battley PF, Piersma T (2009) Extreme endurance flights by landbirds crossing the Pacific Ocean: ecological corridor rather than barrier? Proc R Soc Lond B 276: Gschweng M, Kalko EKV, Querner U, Fiedler W, Berthold P (2008) All across Africa: highly individual migration routes of Eleonora s falcon. Proc R Soc Lond B 275: Hedenström A (2010) Extreme endurance migration: what is the limit to non-stop flight? PLoS Biology 8:e doi: / journal.pbio Higuchi H, Shiu HJ, Nakamura H, Uematsu A, Kuno K, Saeki M, Hotta M, Tokita KI, Moriya E, Morishima E, Tamura M (2005) Migration of honey buzzards Pernis apivorus based on satellite tracking. Ornithol Sci 4: Kalnay E et al (1996) The NCEP/NCAR 40-year reanalysis project. Bull Am Meteorol Soc 77: Kenward R (ed) (2001) A manual for wildlife radio tagging. Academic, London Kerlinger P (1989) Flight strategies of migrating hawks. University of Chicago Press, Chicago Klaassen RHG, Strandberg R, Hake M, Olofsson P, Tøttrup AP, Alerstam T (2010) Loop migration in adult marsh harriers Circus aeruginosus, as revealed by satellite telemetry. J Avian Biol 41: Liechti F (2006) Birds: blowin by the wind? J Orn 147: López-López P, Limiñana R, Urios V (2009) Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool Stud 48: López-López P, Limiñana R, Mellone U, Urios V (2010) From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Land Ecol 25: McGrady MJ, Young GS, Seegar WS (2006) Migration of a Peregrine Falcon Falco peregrinus over water in the vicinity of a hurricane. Ring Mig 23:80 84 Meyer SK, Spaar R, Bruderer B (2000) To cross the sea or to follow the coast? Flight directions and behaviour of migrating raptors approaching the Mediterranean Sea in autumn. Behaviour 137: Navedo JG, Masero JA, Overdijk O, Orizaola G, Sánchez-Guzmán JM (2010) Assessing the role of multiple environmental factors

42 468 Int J Biometeorol (2011) 55: on Eurasian Spoonbill departure decisions from stopover sites. Ardea 98:3 12 Newton I (2008) The migration ecology of birds. Academic, London Pulido F, Berthold P (2010) Current selection for lower migratory activity will drive the evolution of residency in a migratory bird population. Proc Natl Acad Sci USA 107: Richardson WJ (1978) Timing and amount of bird migration in relation to weather: a review. Oikos 30: Robinson RA, Crick HQP, Learmonth JA, Maclean IMD et al (2009) Travelling through a warming world: climate change and migratory species. Endanger Species Res 7:87 99 Rosén M, Hedenstrom A, Badami A, Spina F, Akesson S (1999) Hunting flight behaviour of the Eleonora s falcon Falco eleonorae. J Avian Biol 30: Saino N, Rubolini D, von Hardenberg J, Ambrosini R, Provenzale A, Romano M, Spina F (2010) Spring migration decisions in relation to weather are predicted by wing morphology among trans-mediterranean migratory birds. Func Ecol 24: doi: /j x Shamoun-Baranes J, Baharad A, Alpert P, Berthold P, Yom-Tov Y, Dvir Y, Leshem Y (2003) The effect of wind, season and latitude on the migration speed of white storks Ciconia ciconia, along the eastern migration route. J Avian Biol 34: Stutchbury BJM, Tarof SA, Done T, Gow E, Kramer PM, Tautin J, Fox JW, Afanasyev V (2009) Tracking long-distance songbird migration by using geolocators. Science 323:896 Strandberg R, Klaassen RHG, Hake M, Alerstam T (2009) How hazardous is the Sahara Desert crossing for migratory birds? Indications from satellite tracking of raptors. Biol Lett 6: doi: /rsbl Thorup K, Alerstam T, Hake M, Kjellén N (2003) Bird orientation: compensation for wind drift in migrating raptors is age dependent. Proc R Soc Lond B 270(Suppl):S8 S11 Walter H (1979) Eleonora s falcon. Adaptations to prey and habitat in a social raptor. University of Chicago Press, Chicago Zar JH (1984) Biostatistical analysis. Prentice-Hall, New York Zu-Aretz S, Leshem Y (1983) The sea a trap for gliding birds. Torgos 5:16 17

43 III 41

44

45 Journal of Avian Biology 000: , 2013 doi: /j X x 2013 The Authors. Journal of Avian Biology 2013 Nordic Society Oikos Subjcet Editor: Thomas Alerstam. Accepted 20 March 2013 The trans-equatorial loop migration system of Eleonora s falcon: differences in migration patterns between age classes, regions and seasons Ugo Mellone, Pascual López-López, Rubén Limiñana, Gvido Piasevoli and Vicente Urios U. Mellone (ugomellone@libero.it), P. López-López, R. Limiñana and V. Urios, Vertebrates Zoology Research Group, CIBIO Research Inst., Univ. of Alicante, ES San Vicente del Raspeig, Alicante, Spain. G. Piasevoli, Public Inst. for the Protected Natural Values Management in the County of Split and Dalmatia, Prilaz braće Kaliterna 10, HR Split, Croatia. Internal factors such as experience (e.g. age) and motivation for breeding, and external ones such as environmental conditions (e.g. meteorology and landscape characteristics) can promote differences in migratory behaviour and routes among seasons, regions and populations. Using satellite telemetry we investigated whether such differences occur and which factors promote them among migrating Eleonora s falcons breeding in the Mediterranean area (Spain and Croatia) and wintering in Madagascar. We found that during autumn migration no age differences occur when crossing the Sahara desert, but in the remaining African regions, juveniles were more prone than adults to fly at a slower and more tortuous rate, as well as exhibiting longer stop-overs, particularly in the Sahel region. Such differences might be promoted by a lower foraging and pre-migratory fattening efficiency in juveniles. During spring, routes were significantly more eastern than during autumn, resulting in a loop migration occurring in all studied populations. This could be accounted by seasonal variation in the distribution of trophic resources. Our results show that Eleonora s falcons integrate spatially seasonal changing resources on a continental scale throughout their annual cycle, changing their movement patterns in response to internal (age) and external (habitat) factors. This loop migration pattern may prove to be widespread among other Palearctic trans-continental migratory bird species. Bird migration routes and behaviour during migratory journeys can vary enormously within the same species and even within populations, according to factors such as experience (age), environmental conditions (meteorology and landscape characteristics), motivation (breeding condition) and period of the year (seasons) (Kokko 1999, Alerstam 2006, 2011, Newton 2008, Mellone et al. 2012a). Landscape can vary strikingly within and among the journeys, since during the same journey birds need to cross both unfavourable regions, e.g. ecological barriers when fuelling is not possible, and regions where conditions are more favourable, making stop-overs for refuelling possible (Klaassen et al. 2008, López-López et al. 2010). In addition, conditions in the same region can vary among seasons due to rainfall seasonality and the consequent availability of trophic resources (Klaassen et al. 2010). Due to their morphology, pure soaring migrants are expected to store enough fat in order to accomplish their migratory journeys without foraging (Smith et al. 1986, Hake et al. 2003, Panuccio et al. 2006). By contrast, raptors using powered flight need to replenish their energy stores during the journey, since it seems to be more advantageous to fly lighter and forage en route than increase drag by carrying fuel reserves (Alerstam and Hedenström 1998). Therefore, the flight speed, the geometry of migration routes and the amount of time allocated to fuel deposition are the result of the interactions among internal (experience, motivation) and external (landscape, weather conditions) factors. Overall, these factors, taken together, will ultimately shape migratory loops with seasonal differences in migration routes within the same age class (Newton 2008, Alerstam 2011). The Eleonora s falcon Falco eleonorae is a long-distance migratory raptor, which breeds colonially and almost exclusively on islands in the Mediterranean Sea, with almost the entire population wintering in Madagascar (Walter 1979). To travel between the breeding colonies and wintering areas, twice each year Eleonora s falcons perform a transequatorial journey of ca km, encountering a huge variety of environments and climatic conditions, with juveniles migrating independently of adults (Gschweng et al. 2008, López-López et al. 2009, 2010, Kassara et al. 2012). Therefore, its migration system provides an excellent model to study the effects of both internal and external determinants on migration behaviour. Our aim was to investigate whether there are differences in migratory routes among age classes, regions and periods of the year and to identify which are the most important factors promoting them. Moreover, for the first time, we report data for birds

46 belonging to the Croatian population, located in the northernmost part of the breeding range for the species (Walter 1979) and roughly half-way between the well-studied populations of Sardinia and Greece (Gschweng et al. 2008, Kassara et al. 2012). Methods Fieldwork Between 2007 and 2010, we equipped 16 Eleonora s falcons (eight adults and eight juveniles) on the Balearic and Columbretes islands (Spain) with 9.5 g Argos satellite transmitters together with two adults on Svetac island (Croatia) in In the Spanish colonies, adults were captured using dho-gaza nets and a stuffed eagle owl Bubo bubo as a decoy, while nestlings were tagged at nests a few days before fledging. Croatian adults were trapped close to their nests using mist-nets. The transmitters were attached to birds using a teflon harness backpack. Data selection In the analyses we included only those birds completing at least the crossing of the Sahara desert, totalling 26 migration tracks (Table 1). In some cases, we did not obtain the entire migration as far as Madagascar either because of transmitter s failure or the death of the bird. Moreover, one juvenile (no ) did not reach Madagascar, spending the winter between Kenya and Tanzania. Transmitters were programmed according to three different duty cycles (d.c.) as follows: 1) d.c. 1 (12 h on/18 h off), used for Spanish birds during autumn migration; 2) d.c. 2 (10 h on/48 h off) for Croatian birds during both migrations; and 3) d.c. 3 (12 h on/58 h off) for Spanish birds during spring migration, with one exception (Table 1). Each Argos data is accorded to a location class (l.c.) according to its spatial accuracy (Argos 2011). In order to plot the routes and identify the onset and ending of migrations, we considered all data collected excluding those with l.c. Z, retaining only a maximum of one location per hour (the highest quality one), and using data with l.c. 0, A and B only when in agreement with normal rates of speed and direction (Strandberg et al. 2009). During autumn migration, we considered as adult all individuals that were at least one year old and thus with prior migration experience. In five cases we obtained repeated journeys from the same individuals (i.e. six autumn routes of three individuals and four spring ones of two individuals; Table 1). Migration tracks belonging to the same individual in the same season but in different years were treated as independent cases, since migrating raptors show high inter-year route flexibility regardless of the individual (Alerstam et al. 2006, García-Ripollés et al. 2010, Klaassen et al. 2011, Mellone et al. 2011, 2012a, Vardanis et al. 2011, Limiñana et al. 2012a, b). The migration tracks were plotted on maps and overlapped with the NDVI (normalized difference vegetation index), a measure of the vegetation s greenness (Fig. 1, 2, 5, 7) that has been suggested to be an adequate surrogate of trophic resources abundance (Pettorelli et al. 2005), including grasshoppers (Trierweiler et al. 2013), that are likely to be an easy prey for Eleonora s falcons on migration. Age and regional differences during autumn migration We selected a sub-set of locations only for birds programmed with d.c. 1 (Spanish birds during autumn), retaining only one location per day as close as possible to midnight, in this way building segments lasting ca 24 h (daily segments). Then, we used these points to calculate straightness, daily distance and regional migration speed. Three regions (i.e. Sahara, Sahel, Equatorial Africa; Fig. 1) were identified following the same latitudinal criteria used in López-López et al. (2010), although segments south of the Sahel were pooled in a single category (Equatorial Africa). A preliminary visual inspection of the migratory tracks, including those already published by other authors (Gschweng et al. 2008, Kassara et al. 2012), revealed that Eleonora s falcons, unlike the lesser kestrel Falco naumanni (Limiñana et al. 2012a) do not usually perform true stop-overs, using discrete staging areas for many days. In contrast, Eleonora s falcons slow down migration speed showing less directed movements, probably in response to the distribution of their insect prey. The above-mentioned Table 1. Attributes of the birds used in the analyses. Tracking years Duty cycle Tag ID Colony Age Sex autumn spring autumn spring Balearic adult male Balearic juvenile male Columbretes subadult female Balearic adult female 2008, , Balearic adult male Columbretes adult female 2008, , Columbretes juvenile female Columbretes juvenile female Columbretes subadult male Columbretes juvenile male Columbretes juvenile male Columbretes juvenile male Svetac adult female 2009, Svetac adult female

47 Figure 1. Eleven autumn migration tracks of eight adults Eleonora s falcons. Routes belonging to the same individual in different years mantain the same symbol. The background represents the normalized difference vegetation index (NDVI) for November 2009 (legend in Fig. 4). authors reported that Eleonora s falcons performed long stop-overs but did not specify how they identified these. Therefore, in order to quantify the occurrence of this behaviour, half-way between a stop-over and a fly-andforage strategy (Strandberg and Alerstam 2007), we adopted a multiple approach as follows. We calculated the straightness of the migration tracks by dividing the distance between the tagging site and the final point of migration by the distance obtained summing up all daily segments, only for birds that accomplished the whole journey reaching the wintering grounds. We also calculated partial straightness for each region that was completely crossed. Straightness can vary between zero and one. Values close to zero indicate tortuous paths, while straight paths have values close to one (Benhamou 2004). Regional migration speed was obtained by dividing the straight distance (between the first and the last point of a region traversed) by the time elapsed (measured in number of days), e.g. for the Sahara the distance between the breeding colony and the first location in the Sahel, and so on. In order to identify short-distance days, indicative of stop-over/fly-and-forage behaviour (Klaassen et al. 2010, Limiñana et al. 2012a), we selected all days with daily distance (the distance between two points) lower than 50 km (regardless the flight direction), and also days 100 km, but only if the direction showed a northern

48 component (reversed with respect to the general direction), i.e. in the semi-circle sector from 270 to 90. The above mentioned thresholds were chosen rather conservatively, since an Eleonora s falcon can cover these distances in less than a couple of hours (Rosén et al. 1999), and then use the rest of the day for foraging. Nevertheless, this method allows the comparison of the spatial distribution of these short-distance days between age classes. Segments in Madagascar were excluded because they were all belonging to one juvenile individual that wandered before settling in the wintering area. The effect of internal (age) and external (region) effects upon the three above mentioned variables (straightness, regional speed, short-distance days), were tested by means of three general linear mixed models (GLMM) in which age, region and their interaction were fixed factors and track ID represented a random factor. Non-significant terms were removed stepwise from the models according to their p-value within the model, starting from the interactions, until we obtained models that retained only significant variables. The straightness model included only two regions since there was not enough data of juveniles in Equatorial Africa. Finally, age differences concerning these three variables were also evaluated by means of Mann Whitney two-tailed tests within each region. Regional variations in short-distance days within the same age class were analysed using the Kruskal Wallis test. All analyses were carried out using SPSS Stop-over behaviour during spring migration Data from birds migrating during spring was available at different time intervals, according to the duty cycle of the transmitters (Table 1). In order to calculate daily distance while avoiding any loss of available information, one location per duty cycle was selected and then the distance resulting from these segments was divided by the number of days in between. Segments resulting in a speed lower than 50 km day -1 were considered as indicators of stop-over or fly-and-forage behaviour. Seasonal differences in migration routes For each track we determined at which longitudes the latitudes 30 N, 20 N, 10 N, 0 and 10 S were crossed. Then, we evaluated whether there were differences in longitude between seasons at each latitude by means of Mann Whitney two-tailed tests. We calculated average routes for each season as the sum of segments defined by the average longitudes where the different latitudes were crossed. We used these routes to calculate the overall detour ratio for each season, i.e. how much routes were longer than the straight loxodrome line connecting breeding sites with wintering grounds (as in Klaassen et al. 2010). Juvenile birds during autumn migration were excluded from these analyses in order to avoid any bias, since it has already been shown that they migrate significantly further west than adults (Gschweng et al. 2008). We also excluded Croatian birds since the low sample size did not allow for the carrying out of a separate analysis. We therefore used eight tracks for autumn migration (belonging to six adults) and seven tracks for spring (belonging to three adults and two immatures). Annual cycle For birds for which year-round data was available (four cases belonging to three adult birds) we calculated the percentage of each stage within the annual cycle (autumn and spring migration, wintering, pre-breeding movements and breeding). Results Autumn migration: age differences During autumn migration, adult Spanish Eleonora s falcons spent on average 31.3 days reaching the wintering grounds (SD 9.3, n 7; Fig. 1) while the three juveniles (Fig. 2) employed 51 days on average (SD 11.8, n 3). Croatian Eleonora s falcons reached the wintering grounds in 28 days (SD 10.4, n 3; Fig. 1). The Spanish adults average route straightness was 0.83 (SD 0.05, n 7), while, for the two juveniles wintering in Madagascar, the straightness values were 0.7 and 0.64, with 0.6 for the bird wintering in mainland Africa. For Spanish adults, the actual average straight distance between breeding sites and wintering areas was 7740 km (SD 140, n 7), while the distance travelled was 9319 km (SD 631, n 7). For the two juveniles wintering in Madagascar straight distances were 7744 km and 8578 km, with, respectively, km and km being the distances travelled. For the juvenile wintering in continental Africa the distances were 6070 km and km. Autumn migration: regional differences The GLMMs revealed that straightness varied according to age (F 8.94, p 0.006), regional migration speed according to age (F 10.08, p 0.008) and region (F 16.96, p 0.001) and short-distance days according to all factors: age (F 14.62, p 0.002) region (F 9.4, p 0.001), and their interaction (F 6.43, p 0.006). Further pairwise tests showed that, as far as regional migration speed was concerned, there were no significant differences between adult and juveniles in the Sahara desert (U 11, n.s., Fig. 3b), while differences were significant during the crossing of the Sahel (U 0, p 0.003) and in Equatorial Africa (U 1, p 0.005). A similar pattern occurred with straightness (Fig. 3a, 4; Sahara: U 15, n.s.; Sahel: U 6, p 0.04; there being not enough cases for juveniles in Equatorial Africa). Short-distance days were equally distributed among adults and juveniles in the Sahara desert (Fig. 3c; U 24, n.s.; mean adults : 0.75, SD 1.5, n 8; mean juveniles : 0.22, SD 0.5, n 6), but we observed a significant imbalance towards juveniles in the Sahel (U 3, p 0. 01; mean adults : 1.5, SD 1.1, n 8; mean juveniles : 6.4, SD 3.3, n 5) and Equatorial Africa (U 4.5, p 0.02; mean adults : 0.75, SD 1.2, n 8; mean juveniles : 4.8, SD 4.4, n 5). Within adults, shortdistance days were evenly distributed among regions (K 2.9, n.s.), while among juveniles they were more

49 Figure 2. The autumn migration tracks of six juveniles Eleonora s falcons. The background represents the NDVI for November 2009 (legend in Fig. 4). frequent in the Sahel (K 11.2, p 0.004). In the Sahel, stop-over segments mainly occurred in Mali, Burkina Faso and Nigeria, around of latitude (Fig. 4). Finally, as far as repeated journeys by the same individual were concerned, we found neither route nor stop-over site fidelity during autumn (n 3 adults), with the same individual flying in the Sahara desert on a route up to 1200 km from that folowed the previous year (Fig. 1, 4). Stop-overs during spring migration During spring migration, adult Spanish Eleonora s falcons spent 27.6 days on average to reach the breeding grounds (SD 7.4, n 5) while the immature no spent 62 days (Fig. 5). Out of five tracks for Spanish adults, we found an average of 3.2 (SD 3.7, range 0 9) shortdistance days (all during April) in Ethiopia/Somalia, with one bird also stopping for six days in Cameroon. Among the two Spanish immatures, bird no stopped for 27 days in Ethiopia/Somalia and nine in Chad, while bird no stopped for 13 days in Ethiopia/Somalia, two in Chad and seven in Northern Algeria. In the two spring journeys of Croatian adults, we found seven short distance days in bird no and five in bird no , all of them in Ethiopia during April. Repeated spring journeys were available for two Spanish adults, with very different routes between years, especially when crossing the Sahara desert. For example,

50 Figure 3. Regional variation in straightness (a), migration speed (b) and number of short-distance days (c) of adults (red) and juveniles (yellow) Eleonora s falcons during autumn migration. adult no used two stop-over areas in Ethiopia only ca 200 km apart in 2009 and 2010, and then crossed the Sahara on tracks separated by up to ca 2800 km. Seasonal differences and annual cycle There were no significant differences between spring and autumn migration tracks during the crossing of the Sahara desert (30 N: U 17, n.s.; 20 N: U 21, n.s.). In contrast, in the remaining three intervals the differences were significant (10 N and 0 : U 0, p 0.001; 10 S: U 2, p 0.005), with spring routes lying further to the east than autumn ones (Fig. 1, 5, 6). These average autumn migration routes were 7% longer than an ideal straight route, while spring ones were 5% longer. The low sample size did not allow carrying out this analysis with Croatian birds, but looking at migration tracks obtained (Fig. 1, 5), the same loop migration pattern clearly was observed. Autumn and spring migration accounted for 7% and 9% of the annual cycle, while wintering, pre-breeding movements and breeding accounted for 41%, 18% and 25% respectively (n 4). Figure 4. Tracks of juveniles (a) and adults (b) Spanish Eleonora s falcons crossing the Sahel. The same colour indicates different routes in different years from the same adult individual. The background represents the NDVI for November 2009 (see legend).

51 Figure 5. Nine spring migration tracks of seven Eleonora s falcons. The background represents the NDVI for April 2010 (legend in Fig. 4). Discussion Age and regional differences during autumn migration The results of the analyses concerning straightness, regional migration speed and stop-over days all share the same age-related pattern with no differences during the Sahara crossing and significant differences in the remaining two regions, with the juveniles showing more tortuous routes, slower speeds and more stop-over days than adults, especially in the Sahel. The Sahara desert is an ecological barrier where foraging is barely possible and rapidly changing weather conditions can even result in the birds deaths (Strandberg et al. 2010). Therefore, Eleonora s falcons try to cross this region as fast as possible regardless their age class and the observed variability in travel rates is perhaps largely dependent on weather conditions. By contrast, the regions that follow and the Sahel in particular, host benign habitats where an insectivorous bird such as the Eleonora s falcon can find large amounts of food in the right season (Zwarts et al. 2009). Why then, is there such a difference in migration performance between adults and juveniles? Juveniles stopped as soon as the habitat became suitable for foraging, just south of the edge of the Sahara (Fig. 4). The same pattern was also observed for juveniles from

52 Sardinia and Greece, stopping in the same regions and landscape type (Acacia savannah) for approximately two weeks (Gschweng et al. 2008, Kassara et al. 2012). This difference between age classes might be explained by differences in fuel accumulation before starting migration due to differences in hunting skills. Similar age-dependent patterns have also been observed in other species of raptors (Restani 2000, Ueta and Higuchi 2002, Strandberg and Alerstam 2007). Irrespective of age, the Sahel belt is known to be an important stop-over area for many trans-equatorial migrants, some even stopping for several weeks (Pearson and Lack 1992, Tøttrup et al. 2011) to exploit the favourable foraging conditions occurring after the end of the summer rains. Therefore, it is possible to argue that, the Sahel in particular, is a vast, rather diffuse stop-over area, where juveniles Eleonora s falcons use a fly-and-forage strategy. Such differences ultimately trigger the overall age differences in migration duration, with juveniles needing in some cases twice the time of adults to accomplish the entire journey. Seasonal differences When comparing autumn vs. spring migration routes, two main differences arose: 1) unlike the long non-stop flights ( 1200 km) observed in spring, during autumn Eleonora s falcons try to minimize the flight distance over the Mozambique Channel, crossing where it is narrowest (ca km; Fig. 1) and then heading northeast in Madagascar, since the main wintering grounds are located in the northern part of the island (Gschweng et al. 2012, Mellone et al. 2012b); 2) around ca 5 S, during spring, Eleonora s falcons shift their course eastwards, heading towards Ethiopia and Somalia (Fig. 5, 6). Such differences provide the shape for a loop migration pattern that seems to be consistent and widespread between different years, individuals and population of Eleonora s falcons. In fact, in addition to our Spanish and Croatian birds, birds belonging to the Sardinian and Greek populations also exhibited the same pattern (Gschweng et al. 2008, Kassara et al. 2012). Loop migrations can arise either as a resut of wind, habitat or motivational differences between seasons (Klaassen et al. 2010, Alerstam 2011, Agostini et al. 2012, Limiñana et al. 2012b, 2013). Prevailing wind patterns might explain why Eleonora s falcons seek to minimize the water crossing during autumn migration but not in spring, when winds are more supportive (Kemp et al. 2010, Mellone et al. 2011). However, wind drift cannot be invoked to explain why Eleonora s falcons show such an eastward directional shift towards Ethiopia during spring instead of going to their breeding areas retracing their autumn itinerary, or simply heading northwest because winds are mainly easterlies during both seasons (Kemp et al. 2010). A similar change of course was reported for individually tracked Red-backed shrikes (Lanius collurio; Tøttrup et al. 2011), and explained in the light of more suitable habitat. Thus, if winds are not the reason, an alternative explanation might be habitat suitability: after the long non-stop flight over the ocean and before the crossing of another ecological barrier such as the Sahara desert Eleonora s falcons need to replenish their energy stores. The observed eastward shift also coincides with the region where the majority of spring short-distance Figure 6. Mean longitude plotted against latitude (at intervals of 10 ) of the migratory tracks of adult Eleonora s falcons during autumn (red) and during spring (blue). The inset shows a schematic representation of the same tracks. days occurred (i.e. Ethiopia). This region experiences rain peaks precisely during spring (the belg season, Glantz 1987; see also Pearson and Lack 1992), that might boost insect abundance (Janzen and Schoener 1968, Trierweiler et al. 2013). The high inter-year variation in rainfall patterns (Glantz 1987), as well as individual conditions, could explain the spatial and temporal variation in the occurrence of such stop-over areas (see also Tøttrup et al. 2012). According to Pearson and Lack (1992), the loop migration that many Palearctic migrants seem to perform in this area may be explained by climatic patterns, since in April Ethiopia and Somalia are green, while much of the Sahelian belt is very dry, with the opposite occurring in early autumn (compare the northern limit of the Sahelian green belt in Fig. 1 and 2 vs Fig. 5). Conclusions Are Eleonora s falcons time or energy minimizers? Despite the spring routes being slightly shorter than the autumn ones, adults performed longer stop-overs in spring than in autumn, resulting in a longer time spent during spring migration (9% vs 7% of the annual cycle), suggesting a lack of urgency in reaching the breeding areas. In fact, Eleonora s falcons do not immediately occupy their breeding colonies after spring arrival, but, instead, spend up to two months in pre-breeding areas that may be far away from the colonies (18% of the annual cycle; Mellone et al. 2013), delaying their reproduction until mid-summer, unlike other Palearctic migratory birds (Walter 1979, Mellone et al. 2012c). The fact that the two immatures also deviated eastwards to Somalia and Ethiopia during spring migration suggests that this orientation shift might also be endogenously

53 controlled, e.g. the temporal and spatial components of the loop migration strategy have been fixed in Eleonora s falcons innate circannual program (Thorup et al. 2010; Fig. 6). Eleonora s falcons spatially integrate seasonal changing resources on a continental scale throughout their annual cycle, changing their movement patterns in response to internal (age) and external factors (habitat). Some regions crossed by Eleonora s falcons during migration can be used as foraging areas and their selection as such is highly dependent upon the season, promoting a loop migration. Future research using GPS devices, with higher spatial and temporal resolution, will provide new insight into how Eleonora s falcons mix pure migrational behaviour with the fly-and-forage and stop-overs strategies, and will eventually help to identify priority areas for their conservation. Acknowledgements The Terra Natura Foundation and the Servicio de Biodiversidad (Generalitat Valenciana) funded the Spanish part of the project. Special thanks are due to all the people who helped in birds trapping. The Association for the Preservation of Biodiversity Falco donated the transmitters used in Croatia and we would like to thank Mr. Ivica Lolić for his invaluable help during field work. We acknowledge the use of the Maptool program ( ) for plotting the maps. The English of the manuscript was improved by Paul Tout. We also thank Thomas Alerstam for useful suggestions. UM is supported by an FPU grant of the Spanish Ministry of Education (reference AP ). PLL is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Economy and Competitiveness (reference JCI ). This paper is part of the PhD dissertation of UM at the Univ. of Alicante. This study complies with the current laws in Spain and Croatia. References Agostini, N., Lucia, G., Mellone, U., Panuccio, M., Von Hardenberg, J., Evangelidis, A. and Kominos, T Loop migration of adult European honey buzzards (Pernis apivorus) through the central-eastern Mediterranean. Ital. J. Zool. 79: Alerstam, T Strategies for the transition to breeding in time-selected bird migration. Ardea 94: Alerstam, T Optimal bird migration revisited. J. Ornithol. 152: S5 S23. Alerstam, T. and Hedenström, A The development of bird migration theory. J. Avian Biol. 29: Alerstam, T., Hake, M. and Kjellén, N Temporal and spatial patterns of repeated migratory journeys by ospreys. Anim. Behav. 71: Argos Argos user s manual. Worldwide tracking and environmental monitoring by satellite. CLS, Toulouse. Benhamou, S How to reliably estimate the tortuosity of an animal s path: straightness, sinuosity or fractal dimension? J. Theor. Biol. 229: García-Ripollés, C., López-López, P. and Urios, V First description of migration and wintering of adult Egyptian vultures Neophron percnopterus tracked by GPS satellite telemetry. Bird Study 57: Glantz, M Drought and hunger in Africa. Cambridge Univ. Press. Gschweng, M., Kalko, E. K. V., Querner, U., Fiedler, W. and Berthold, P All across Africa: highly individual migration routes of Eleonora s falcon. Proc. R. Soc. B. 275: Gschweng, M., Kalko, E. K. V., Fiedler, W., Berthold, P. and Fahr, J Multi-temporal distribution modelling with satellite tracking data: predicting responses of a long-distance migrant to changing environmental conditions. J. Appl. Ecol. 49: Hake, M., Kjellen, N. and Alerstam, T Age-dependent migration strategy in honey buzzards Pernis apivorus tracked by satellite. Oikos 103: Janzen, D. H. and Schoener, T. W Differences in insect abundance and diversity between wetter and drier sites during a tropical dry season. Ecology 49: Kassara, C., Fric, J., Gschweng, M. and Sfenthourakis, S Complementing the puzzle of Eleonora s falcon (Falco eleonorae) migration: new evidence from an eastern colony in the Aegean Sea. J. Ornithol. 153: Kemp, M. U., Shamoun-Baranes, J., Van Gasteren, H., Bouten, W. and Van Loon, E. E Can wind help explain seasonal differences in avian migration speed?. J. Avian Biol. 41: Klaassen, R. H. G., Strandberg, R., Hake, M. and Alerstam, T Flexibility in daily travel routines causes regional variation in bird migration speed. Behav. Ecol. Sociobiol. 62: Klaassen, R. H. G., Strandberg, R., Hake, M., Olofsson, P., Tottrup, A. P. and Alerstam, T Loop migration in adult marsh harriers Circus aeruginosus as revealed by satellite telemetry. J. Avian Biol. 41: Klaassen, R. H. G., Hake, M., Strandberg, R. and Alerstam, T Geographical and temporal flexibility in the response to crosswinds by migrating raptors. Proc. R. Soc. B. 278: Kokko, H Competition for early arrival in migratory birds. J. Anim. Ecol. 68: Limiñana, R., Romero, M., Mellone, U. and Urios, V. 2012a. Mapping the migratory routes and wintering areas of lesser kestrels Falco naumanni: new insights from satellite telemetry. Ibis 154: Limiñana, R., Soutullo, A., Urios, V. and Reig-Ferrer, A. 2012b. Migration and wintering areas of adult Montagu s harriers (Circus pygargus) breeding in Spain. J. Ornithol. 153: Limiñana, R., Romero, M., Mellone, U. and Urios, V Is there a different response to w inds during migration between soaring and flapping raptors? An example with the Montagu s harrier and the lesser kestrel. Behav. Ecol. Sociobiol. DOI: /s López-López, P., Limiñana, R. and Urios, V Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool. Stud. 48: López-López, P., Limiñana, R., Mellone, U. and Urios, V From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Land. Ecol. 25: Mellone, U., López- López, P., Limiñana, R. and Urios, V Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int. J. Biomet. 55: Mellone, U., Klaassen, R. H. G., García-Ripollés, C., Limiñana, R., López-López, P., Pavón, D., Strandberg, R., Urios, V., Vardakis, M. and Alerstam, T. 2012a. Interspecific comparison of the performance of soaring migrants in relation to morphology, meteorological conditions and migration strategies. PLoS ONE 7(7): e Mellone, U., López-López, P., Limiñana, R. and Urios, V. 2012b. Wintering habitats of Eleonora s falcons Falco eleonorae in Madagascar. Bird Study 59: Mellone, U., Urios, V., Rguibi-Idrissi, H., Limiñana, R., Benhoussa, A. and López-López, P. 2012c. Ranging behaviour of Eleonora s falcons Falco eleonorae during chick-rearing. Acta Ornithol. 47: Mellone, U., López-López, P., Limiñana, R. and Urios, V Summer pre-breeding movements of Eleonora s falcons Falco eleonorae revealed by satellite telemetry: implications for

54 conservation. Bird Conserv. Int. DOI: /S Newton, I The migration ecology of birds. Academic Press, Elsevier. Panuccio, M., Agostini, N., Wilson, S., Lucia, G., Ashton-Booth, J., Chiatante, G., Mellone, U. and Todisco, S Does the honey-buzzard feed during migration? Brit. Birds 99: Pearson, D. J. and Lack, P. C Migration patterns and habitat use by passerine and near-passerine migrant birds in eastern Africa. Ibis 134: Pettorelli, N., Vik, J. O., Mysterud, A., Gaillard, J. M., Tucker, C. J. and Stenseth, N. C Using the satellite-derived NDVI to assess ecological responses to environmental change. Trends Ecol. Evol. 20: Restani, M Age-specific stopover behavior of migrant bald eagles. Wilson Bull. 112: Rosén, M., Hedenström, A., Badami, A., Spina, F. and Åkesson, S Hunting flight behaviour of the Eleonora s falcon Falco eleonorae. J. Avian Biol. 30: Smith, N. G., Goldstein, D. L. and Bartholomew, G. A Is long-distance migration possible for soaring hawks using only stored fat? Auk 103: Strandberg, R. and Alerstam, T The strategy of fly and forage migration, illustrated for the osprey (Pandion haliaetus). Behav. Ecol. Sociobiol. 61: Strandberg, R., Klaassen, R. H. G., Olofsson, P. and Alerstam, T Daily travel schedules of adult Eurasian hobbies Falco subbuteo variability in flight hours and migration speed along the route. Ardea 97: Strandberg, R., Klaassen, R. H. G., Hake, M. and Alerstam, T How hazardous is the Sahara Desert crossing for migratory birds? Indications from satellite tracking of raptors. Biol. Lett. 6: Thorup, K., Holland, R. A., Tøttrup, A. P. and Wikelski, M Understanding the migratory orientation program in birds: extending laboratory studies to study free-flying migrants in a natural setting. Integr. Comp. Biol. 50: Tøttrup, A. P., Klassen, R. H. G., Strandberg, R., Thorup, K, Kristensen, M. W., Jorgensen, P. S., Fox, J., Afanasyev, V., Rahbek, C. and Alerstam T The annual cycle of a trans-equatorial Eurasian African passerine migrant: different spatio-temporal strategies for autumn and spring migration. Proc. R. Soc. B 279: Tøttrup, A. P., Klassen, R. H. G., Kristensen, M. W., Strandberg, R., Vardanis, Y., Lindström, Å., Rahbek, C., Alerstam, T. and Thorup, K Drought in Africa caused delayed arrival of European songbirds. Science 338: Trierweiler, C., Mullié, W. C., Drent, R. H., Exo, K.-M., Komdeur, J., Bairlein, F., Harouna, A., de Bakker, M. and Koks, B. J A Palaearctic migratory raptor species tracks shifting prey availability within its wintering range in the Sahel. J. Anim. Ecol. 82: Ueta, M. and Higuchi, H Difference in migration pattern between adult and immature birds using satellites. Auk 119: Vardanis, Y., Klaassen, R. H. G., Strandberg, R. and Alerstam, T Individuality in bird migration: routes and timing. Biol. Lett. 7: Walter, H Eleonora s falcon. Adaptations to prey and habitat in a social raptor. Univ. of Chicago Press. Zwarts, L., Bijlsma, R. G., van der Kamp, J. and Wymenga E Living on the edge: wetlands and birds in a changing Sahel. KNNV Publishing.

55 IV 53

56

57 Regional and age-dependent differences in the effect of wind on the migratory routes of Eleonora's Falcon Ugo Mellone 1, Rubén Limiñana 1, Pascual López-López 1, & Vicente Urios 1 1 Vertebrates Zoology Research Group, Instituto Universitario de Investigación CIBIO, University of Alicante, Apdo. 99, E Alicante, Spain. Summary During migration, birds can show different responses to wind in relation to distance to the goal, experience, ecological barriers and visibility of landmarks. We analysed the effect of wind (tailwinds and crosswinds) on daily movement rates (forward and perpendicular) of Eleonora s falcons using ARGOS satellite telemetry, during their trans-continental autumn migration to Madagascar, in relation to the different crossed regions and age classes of the individuals. Our results showed that the effect of wind on daily movement rates was not uniform, being stronger in the farthest region from the migration goal, the Sahara desert. In the Sahel, the results are more conflicting, perhaps because daily movements are more shaped by the distribution of food resources. In Equatorial Africa, daily movement rates were mainly affected by crosswinds. Still, it remains unclear which orientation mechanism allows Eleonora s falcons to reach such a narrow wintering area compensating also for wind displacement. INTRODUCTION During their migrations, animals have to deal with the movement of the fluid in which they actually move, such as birds and insects with winds or fishes with water currents (Chapman et al. 2012). In order to reach their goal, migratory birds must have navigational abilities and be able to determine wind direction to cope with wind conditions en route. These abilities can determine the success of the migration and thus of all the other events of the birds life cycle, since the inability to compensate for wind drift, or the selection of bad wind conditions for departure, can greatly increase mortality risk (Liechti 2006). Birds show different responses to crosswinds, which can be summarized in drift, compensation and overcompensation (Klaassen et al. 2011). Until recently, it was possible to study these topics only through tracking radars (Green & Alerstam 2002; Liechti 2006) or just by visual observations (Agostini et al. 2005), thus studying only a small portion of the whole migratory route. Nowadays, the availability of new technologies such as satellite tracking is enabling to follow migratory birds along their whole migration routes, which allows knowing their exact final destination and, consequently, enabling spatially explicit analyses on responses to wind conditions (Thorup et al. 2003, Klaassen et al. 2011). Furthermore, such analyses are also possible thanks to the easier availability of large scale data sets of meteorological data (Kemp et al. 2012). Theoretical and empirical studies on the responses of birds to wind drift allow the identification of the different mechanisms in response either

58 to the distance to the goal, the landscape characteristics or birds experience, although these factors have never been analysed simultaneously. In particular, theoretical studies predict that the more distant the goal, the highest the drifting effect, while compensation and overcompensation are expected to occur more frequently closer to the goal (Alerstam 1979b). Concerning landscape characteristics, birds would be more prone to allow drifting when crossing ecological barriers, areas where environmental conditions are more severe and foraging opportunities are scarcer. Hence, birds would have to compensate for this drift during the crossing of non-barrier areas to avoid ending in areas distant from their goal areas (Alerstam 1979a, Klaassen et al. 2011). Experience also plays an important role, since adult individuals, being already experienced and able to perform true navigation, should show higher compensation rates than juveniles, which are supposed to be not able to compensate for perpendicular displacement (Drost 1938, Thorup et al. 2003). Therefore, in order to assess which of these mechanisms are acting, here we analyse the effect of these different mechanisms within the same migratory system, the trans-equatorial autumn migration journey of the Eleonora s falcon (Falco eleonorae), investigating the effect of wind on forward and perpendicular movements, and accounting for the different passage regions and ageclasses. This is the first analysis of wind drift on a trans-equatorial land migrant (to date, only a few papers on seabirds are available in this respect, e.g. Felicisimo et al. 2008), and also the first one taking into account both age and regional differences at the same time, while previous papers took into account these aspects separately, e.g. age differences in Honey buzzards (Pernis apivorus; Thorup et al. 2003) and regional differences in adults Osprey (Pandion haliaetus) and Marsh harriers (Circus aeruginosus; Klaassen et al. 2011). Previous studies on the species reveal that meteorological conditions experienced by Eleonora s falcons along the routes, such as the presence of unfavourable meteorological conditions over the Indian Ocean, promote interyear variability in their migratory tracks (Mellone et al. 2011). In contrast, another study dealing with the whole migration journey during both seasons found no effect of wind conditions upon flight speed (Kassara et al. 2012). The Eleonora s falcon is a long-distance migratory raptor, which breeds colonially in small islands of the Mediterranean Sea and the Atlantic Ocean (Walter 1979). The wintering range of the species is very small, since almost the whole population concentrates in Madagascar (Walter 1979). To reach Madagascar, Eleonora s falcons perform a transequatorial journey of ca km, flying sometimes also during the night, crossing a huge variety of landscapes, such as deserts, savannahs and forests, with juveniles migrating independently of adults (Gschweng et al. 2008, López- López et al. 2009, 2010, Kassara et al. 2012, Mellone et al. 2013). Therefore, its migration system provides an excellent model to study the regional and age related variation of the response of migratory birds to wind conditions. METHODS Migration data Eleonora s falcons were trapped in Balearic and Columbretes Islands (Spain) in September, between 2007 and 2010 (Fig. 1; details in López- López et al. 2009, 2010, Mellone et al.

59 2013). Birds were tagged with Microwave Telemetry Inc. 9.5-gram solar-powered satellite transmitters using a Teflon ribbon harness. Transmitters were programmed to collect data on a duty cycle of 12-h ON / 15-h OFF during migration. The dataset included 12 individuals (six adults and six juveniles). Two adults were tracked during two consecutive years and hence a total of 14 autumn migration tracks were recorded. Data belonging to the same individuals were considered as independent events, since migrating raptors show high inter-year route flexibility (García-Ripollés et al. 2010; Klaassen et al. 2011, Mellone et al. 2011, 2012, Vardanis et al 2011, Limiñana et al. 2012). In order to identify migration segments, we used locations recorded during the night as start and ending point of these segments (local time: 6 pm-5 am), using only high quality locations (Argos location classes 3, 2, 1 and 0). Eleonora s falcons migrate also during night (López-López et al. 2010) and hence we also included some nocturnal movements in the segments. However, the large majority of movements within each segment corresponded to day-time migratory movements. Exploratory analyses showed that it was not possible to examine possible differences in response to winds between day and night because of small sample size and the different temporal resolution of bird s locations and wind data. The length of these segments was considered as the daily distance. Since in a few cases night locations were lacking, we also included segments lasting two days (N = 8). In this case, the daily distance was calculated by dividing the segment s length by two (Klaassen et al. 2011). To avoid including data belonging to stopovers, segments with daily distance shorter than 75 km were excluded from the analyses (see Klaassen et al. 2011). Every daily segment was assigned to a region (Sahara, Sahel, Equatorial Africa; Figure 1) according to the same criteria used in López-López et al. (2010), though segments South of the Sahel were pooled in a single category since there were no differences in flight behaviour between the two regions (López-López et al. 2010). Wind data and statistical analyses Wind data (u and v components of wind vectors) were obtained from NCEP/NCAR Reanalysis project, as provided by NOAA/OAR/ESRL PSD, Boulder, CO, USA, downloaded through the R package RNCEP (Kemp et al. 2012). Following Klaassen et al. (2011) and Limiñana et al. (2013), we used data from the pressure level of 925 hpa, as this is the altitude at which most of the migration movements of raptors occur (see also Mateos- Rodríguez & Liechti 2012). To simulate wind conditions experienced by birds in every migration segment we averaged wind data collected for the beginning of the segments at 06h, at 12h (midpoint) and at 18h (endpoint), giving twice much weight to midpoint(s) of migration day(s) in the calculations (Klaassen et al. 2011, Limiñana et al. 2013). Forward and perpendicular components of birds movement (measured as km/day), tailwinds and crosswinds (m/s), were all computed for each migration segment assuming a general migration direction of 135, which is followed by Eleonora s falcons from breeding colonies to reach Madagascar through the narrowest point of the Mozambique Channel (Gschweng et al. 2008, López-López et al. 2009, 2010, Kassara et al. 2012, Mellone et al. 2013). The effect of wind upon migratory movements was first tested using Generalized

60 Figure 1. Autumn migration segments of Eleonora s falcons used for analyses. Linear Mixed Models (GLMM), in which we fitted the track ID as random factor. Age, wind and their interaction were considered as fixed factors, carrying out different models separately for the three regions. Thereby, we carried out six different models for the three regions and for the two dependent variables (forward movement or perpendicular movement), fitting for each case the correspondent wind variable (tailwind or crosswind). Nonsignificant terms were removed stepwise from the models according to their p-value within the model, starting from the interactions, until we obtained models that retained only significant variables. Then, we tested the effect of

61 wind by means of linear regression (Thorup et al. 2003, Klaassen et al. 2011, Limiñana et al. 2013), correlating forward components of the migration segments with tailwinds and perpendicular components with crosswinds. Analyses were carried out separately for regional (three) and age (two) categories. Thereby, we computed a total of 12 regressions (i.e., three regions per two age classes per two movement components). RESULTS The results showed that wind conditions have an effect upon movement rates in all regions, albeit results for the Sahel were weaker, since there was no relationship between tailwind and forward movements in this region (Table 1). Linear regressions were performed for segments shown in Figure 1 and were significant in eight out of 12 cases, showing that wind affects migratory movements of Eleonora s falcons, although there were large variations among regions (Table 2. Figure 2). In summary, the stronger effect occurred in the Sahara desert, where both movement components were significantly related to winds in both age classes (Figure 2). In the Sahel, results were more conflictive, since adults were affected only by crosswinds and juveniles only by tailwinds. In the equatorial region, daily movement rates of both age classes were affected by crosswinds and not affected by tailwinds, although in the latter case the results obtained were almost significant (Table 2). DISCUSSION We analysed the effect of wind on a trans-equatorial overland migrant across different regions and age classes (Figure 1). Therefore, overall, the wind effect was higher during the crossing of the Sahara desert, in agreement with theoretical predictions and previous studies, since this region is an ecological barrier and is far from the final goal (Alerstam 1979a, 1979b). Apparently, in this region the tailwind effect is very strong for both adults and juveniles, and a comparison of the slope Source Denominator df F Sig. Forward component Sahara Tailwind Sahel* (Age) (.056) Equatorial Africa Tailwind Age Perpendicular component Sahara Crosswind Sahel Crosswind Age Age*Crosswind Equatorial Africa Crosswind Age Table 1. Final GLMMs with track (individual) as random factor. (* in this case no variable was retained in the final model, with Age being the last one to be removed; see methods).

62 Table 2. Linear regressions between daily rates of movement (km/day) and wind components (m/sec). Abbreviations: C.I., Confidence interval; ns, non-significant. P values of non-significant relationships are given in brackets. Forward movement versus tailwind Perpendicular movement versus crosswind Region Age N R 2 P Slope C.I. (95%) Intercept R 2 P Slope C.I. (95%) Intercept Sahara adults juveniles Sahel adults ns (0.714) juveniles ns (0.662) Equatorial Africa adults ns (0.052) juveniles ns (0.056)

63 values (Table 2) suggests that crosswinds affect more the adults than the juveniles. Considering the restricted wintering range used by the Eleonora s falcon, it may be expected that juveniles have better compensation abilities than shown by Thorup et al. (2003) for the juveniles of other species with wider wintering ranges. However, such imbalance towards adults, which are fully drifted, is striking and deserves further studies. Regardless possible age differences in the Sahara desert, southwestern heading segments (thus diverging from the general migration direction) are particularly frequent (Figure 1), especially in the lower third of the region, suggesting that the dogleg-shaped route of the species is promoted by wind conditions. The conflicting results obtained for the Sahel region could be influenced by other factors different to winds, as for example birds performing less directed movements in order to search for food resources, particularly juveniles (Gschweng et al. 2008, Kassara et al. 2012, Mellone et al. 2013). In the last region, Equatorial Africa, daily movement rates of both age classes were affected by winds in their perpendicular component, but not in their forward one, perhaps because, in an effort to reach as soon as possible their final goal ( sprint effect, see also Alerstam 2006), they try to increase the daily flight range irrespectively of wind. Concerning the effect of individual experience, GLMMs did not identify an important effect of the interaction Figure 2. Scatter plots of rates of movement (km/day) in relation to wind components (m/sec). Significant results are marked with an asterisk.

64 between age classes and wind conditions throughout the different regions. A further source of variation in the response to wind could be landmark visibility, obviously diminishing drastically during night. Therefore, the drift effect is expected to be stronger during nocturnal migration if birds use landmarks to orientate, since visual cues will not be visible (Liechti 2006, Klaassen et al. 2011). Unfortunately, it has been not possible to test this assumption in the present study, but it has been already shown that Eleonora s Falcons are able to navigate during both day and night over a featureless landscape like the Indian Ocean (Mellone et al. 2011) exhibiting low levels of route repeatability between seasons and years throughout their whole migration route (Gschweng et al. 2008, Kassara et al. 2012, Mellone et al. 2013). Hence, landmark recognition is not likely to be an important cue used by these birds to navigate, as well as suggested for other long distance migratory raptors such as the Osprey, Marsh harrier and Lesser kestrel (Falco naumanni; Alerstam et al. 2006, Vardanis et al. 2011, Limiñana et al. 2013). It seems that Eleonora s falcons migrate overall more irrespectively of wind than larger soaring raptors (Thorup et al. 2003, Klaassen et al. 2011, Mellone et al. 2012). Similar results have been also obtained in the Lesser Kestrel, a species very similar in morphology and flight mode (Limiñana et al. 2013). This seems to be confirmed also by the low explanatory power of the regressions, especially outside the Sahara desert, where they frequently stopped migration probably regardless of wind conditions. However, we cannot exclude that winds affect instantaneous movements more than our analysis is suggesting, but such effect is not visible at the daily scale due to variation in travelling hours among different days. Unfortunately, Argos transmitters do not allow computing exactly on a daily basis the number of travelling hours, and earlier analyses suggest a regional variation in this respect (López-López et al. 2010). Our findings disagree with results of Kassara et al. (2012), which did not find any effect of wind conditions on migration speed of Eleonora s falcons, probably due to the different statistical procedures used, as they classified wind direction only into four categories (thus loosing information) and bird speed data was analysed on an hourly scale (maximum 4 hours, thus without matching with spatial/temporal resolution of wind data, as acknowledged also by the authors). It still remains a mystery how these birds reach such a restricted wintering area, especially when considering juveniles migrating alone. Certainly, this ability implies that birds not only have a compass sense but they are also able to correct for the displacements by compensating significantly after it. It remains an open question if Eleonora s falcon and other birds reaching such narrow goal areas rely only on a mapbased orientation system (e.g. according to Earth s magnetic field or celestial cues) or they use any further mechanism (Thorup et al. 2010), like the capability of sensing infrasound radiated from steep-sided topographic features, as suggested by Gschweng et al. (2009; see also Hagstrum 2000, 2013). Acknowledgements The Terra Natura Foundation and the Servicio de Biodiversidad (Generalitat Valenciana) funded this project. Special thanks are due to all the people who helped in bird s trapping. We thank Michael Kemp and Agustina Di Virgilio for their advice, and Beatriz Arroyo, Dan Chamberlain and two anonymous reviewers for their valuable comments. U. Mellone is

65 supported by FPU grant of the Spanish Ministry of Education (reference AP ). P. López-López is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Economy and Competitiveness (reference JCI ). This paper is part of the Ph.D. dissertation of U. Mellone at the University of Alicante. This study complies with the current laws in Spain. REFERENCES Agostini, N., Premuda, G., Mellone, U., Panuccio, M., Logozzo, D., Bassi, E. & Cocchi, L Influence of wind and geography on orientation behavior of adult Honey Buzzards Pernis apivorus during migration over water. Acta Ornithol. 40: Alerstam, T. 1979a Optimal use of wind by migrating birds: combined drift and overcompensation. J. Theor. Biol. 79: Alerstam, T. 1979b Wind as a selective agent in bird migration. Ornis Scand. 10: Alerstam, T Strategies for the transition to breeding in time-selected bird migration. Ardea 94: Alerstam, T., Hake., M. & Kjellén, N Temporal and spatial patterns of repeated migratory journeys by ospreys. Anim. Behav. 71: Chapman, J.W., Klaassen, R.H.G., Drake, V.A., Fossette, S., Hays, G.C., Metcalfe, J.D., Reynolds, A.M., Reynolds, D.R., & Alerstam, T Animal orientation strategies for movement in flows. Curr. Biol. 21: R Drost, W.A Über den Einfluss von Verfrachtungen zur Herbstzugzeit auf den Sperber Accipiter nisus (L.). In: Proc. Int. Orn. Congress 9: García-Ripollés, C., López-López, P., & Urios, V First description of migration and wintering of adult Egyptian vultures Neophron percnopterus tracked by GPS satellite telemetry. Bird Study 57: Green, M. & Alerstam, T The problem of estimating wind drift in migrating birds. J. Theor. Biol. 218: Gschweng, M., Kalko, EKV, Querner, U., Fiedler, W. & Berthold, P All across Africa: highly individual migration routes of Eleonora s falcon. Proc. R. Soc. Lond. B. 275: Hagstrum, J.T Infrasound and the avian navigational map. J. Exp. Biol. 203: Hagstrum, J.T Atmospheric propagation modeling indicates homing pigeons use loftspecific infrasonic ʻmapʼ cues. J. Exp. Biol. 216: Kassara, C., Fric, J., Gschweng, M. & Sfenthourakis, S Complementing the puzzle of Eleonora s Falcon (Falco eleonorae) migration: new evidence from an eastern colony in the Aegean Sea. J. Orn. 153: Kemp, M.U., van Loon, E.E., Shamoun- Baranes, J. & Bouten, W RNCEP: global weather and climate data at your fingertips. Methods Ecol Evol 3: Klaassen, R.H.G., Hake, M., Strandberg, R. & Alerstam, T Geographical and temporal flexibility in the response to crosswinds by migrating raptors. Proc. R. Soc. Lond. B. 278: López-López, P., Limiñana R & Urios, V Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool. Stud. 48: López-López, P., Limiñana, R., Mellone, U. & Urios, V From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Land. Ecol. 25: Liechti, F Birds: blowin' by the wind? J. Orn. 147: Limiñana, R., Soutullo, A., Urios, V. & Reig- Ferrer, A Migration and wintering areas of adult Montagu s Harriers (Circus pygargus) breeding in Spain. J. Orn. 153: Limiñana, R., Romero, M., Mellone, U. & Urios, V Is there a different response to winds during migration between soaring and flapping raptors? An example with the Montagu s harrier and the lesser kestrel. Behav Ecol Sociobiol: DOI /s Mateos-Rodríguez, M. & Liechti, F How do diurnal long-distance migrants select flight altitude in relation to wind? Behav Ecol 23: Mellone, U., López-López, P., Limiñana, R. & Urios, V Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int. J. Biomet. 55: Mellone, U., Klaassen, R.H.G., García- Ripollés, C., Limiñana, R., López-López, P., Pavón, D., Strandberg, R., Urios, V., Vardakis, M. & Alerstam, T Interspecific comparison of the performance of soaring migrants in relation to morphology, meteorological conditions and migration strategies. PLoS ONE 7(7): e Mellone U., López- López P., Limiñana R., Piasevoli G. & Urios V The transequatorial loop migration system of

66 Eleonora s falcon: differences in migration patterns between age classes, regions and seasons. J. Avian Biol. DOI /j X x Thorup, K., Alerstam, T., Hake, M. & Kjellén, N Bird orientation: compensation for wind drift in migrating raptors is age dependent. Proc. R. Soc. Lond. B. 270 (Suppl). S8 S11. Thorup, K., Holland, R.A., Tøttrup, A.P. & Wikelski, M Understanding the Migratory Orientation Program in Birds: Extending Laboratory Studies to Study Freeflying Migrants in a Natural Setting. Integr. Comp. Biol. 50: Vardanis, Y., Klaassen, R.H.G., Strandberg, R., & Alerstam, T Individuality in bird migration: routes and timing. Biol. Lett. 7: Walter, H Eleonora s falcon. Adaptations to prey and habitat in a social raptor. London: The University of Chicago Press

67 V 65

68

69 Interspecific Comparison of the Performance of Soaring Migrants in Relation to Morphology, Meteorological Conditions and Migration Strategies Ugo Mellone 1 *, Raymond H. G. Klaassen 2, Clara García-Ripollés 1, Ruben Limiñana 1,3, Pascual López- López 1, Diego Pavón 4, Roine Strandberg 2, Vicente Urios 1, Michalis Vardakis 2, Thomas Alerstam 2 1 Estación Biológica Terra Natura, Vertebrates Zoology Research Group, CIBIO, University of Alicante, Alicante, Spain, 2 Department of Biology, Lund University, Lund, Sweden, 3 Instituto de Investigación en Recursos Cinegéticos (IREC), CSIC-UCLM-JCCM, Ciudad Real, Spain, 4 Bird Ecology Unit, Department of Biosciences, University of Helsinki, Helsinki, Finland Abstract Background: Performance of migrating birds can be affected by a number of intrinsic and extrinsic factors like morphology, meteorological conditions and migration strategies. We compared travel speeds of four raptor species during their crossing of the Sahara desert. Focusing the analyses on this region allows us to compare different species under equivalent conditions in order to disentangle which factors affect migratory performance. Methodology/Principal Finding: We tracked raptors using GPS satellite transmitters from Sweden, Spain and Italy, and evaluated their migratory performance at both an hourly and a daily scale. Hourly data (flight speed and altitude for intervals of two hours) were analyzed in relation to time of day, species and season, and daily data (distance between roosting sites) in relation to species, season, day length and tailwind support. Conclusions/Significance: Despite a clear variation in morphology, interspecific differences were generally very small, and did only arise in spring, with long-distance migrants (.5000 km: osprey and Western marsh-harrier) being faster than species that migrate shorter distances (Egyptian vulture and short-toed eagle). Our results suggest that the most important factor explaining hourly variation in flight speed is time of day, while at a daily scale, tailwind support is the most important factor explaining variation in daily distance, raising new questions about the consequences of possible future changes in worldwide wind patterns. Citation: Mellone U, Klaassen RHG, García-Ripollés C, Limiñana R, López-López P, et al. (2012) Interspecific Comparison of the Performance of Soaring Migrants in Relation to Morphology, Meteorological Conditions and Migration Strategies. PLoS ONE 7(7): e doi: /journal.pone Editor: Gil Bohrer, Ohio State University, United States of America Received February 15, 2012; Accepted May 27, 2012; Published July 2, 2012 Copyright: ß 2012 Mellone et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Satellite tracking studies were supported by the Swedish Research Council (Swedish birds, grant to TA), Terra Natura Foundation and Diputaciónde Alicante (Spanish birds), and Parco Gallipoli Cognato Piccole Dolomiti Lucane (Italian birds). UM is supported by a FPU grant of the Spanish Ministerio de Educación (reference AP ) and carried out this research during a stay at the Department of Biology and Centre for Animal Movement Research (Lund University, Sweden). RHGK was supported by a postdoctoral grant from the Swedish Research Council. RL has a postdoctoral grant (reference 10/12-C) co-funded by Consejería de Educación y Ciencia (JCCM) and the European Social Fund. PLL is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Economy and Competitiveness (reference JCI ). DP is awarded with a grant from the Research Foundation of The University of Helsinki. This paper is part of the Ph.D. dissertation of UM at the University of Alicante. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * ugomellone@libero.it Introduction The migratory performance of birds can be shaped by intrinsic factors like species morphology, experience and migration strategy [1 3] and also by extrinsic factors like landscape properties and meteorological variables [4 7]. Diurnal migrants flying over land can migrate by flapping or soaring flight. As energy consumption during flapping flight increases steeply with body mass, soaring migration is more advantageous for larger bird species [8,9]. However, soaring flight requires thermal updrafts that develop only over land and only during the day and hence, soaring migrants are more dependent on topography and circadian patterns in thermal convection than birds using flapping flight [10]. Here, we analyze migratory performance of four species of diurnal raptors using GPS satellite telemetry. We analyze a set of species tagged in different regions of Europe, differing greatly in morphology and behaviour, but sharing basically the same flyway and migration period, and thus facing the same environmental conditions during migration. We restricted our analysis to the crossing of the Sahara desert in order to avoid any variation related to food searching behaviour. Since the desert is an ecological barrier [5,11], all the species probably cross it as efficiently and safely as possible, allowing for comparison under equivalent conditions. We explore the general performance of soaring migrants crossing the Sahara desert, investigating factors like cross-country speed, flight altitude, daily migration time, daily migration PLoS ONE 1 July 2012 Volume 7 Issue 7 e39833

70 Factors Affecting Soaring Bird Migration distance, and effects of wind and thermal convection on resulting speed. It is only through the novel high-resolution satellite tracking of migrants that it is now possible to study these basic aspects of the performance of migratory birds at hourly and daily time scales [4]. In this, we compare different species making the same crossing of the Sahara desert, including osprey (Pandion haliaetus), Western marsh-harrier (Circus aeruginosus), Egyptian vulture (Neophron percnopterus) and short-toed eagle (Circaetus gallicus). These four species differ with respect to body mass and morphology as well as to migration distances: while ospreys and Western marsh-harriers travel around 5000 km between the breeding and the wintering grounds, Egyptian vultures and short-toed eagles breed roughly 2000 km closer to the wintering grounds (Table 1). On the basis of these differences, we formulated the following hypothesis and tested the related predictions (Table 2): 1) Wing Loading Hypothesis Species with lower wing loadings are able to exploit weaker and narrower thermals as they have lower minimum sink speeds and smaller turning radia, and are thus expected to initiate migration earlier in the morning and continue migration until later in the afternoon than species with higher wing loadings [12,13]. However, the gliding speed between thermals is higher for species with higher wing loadings, which more than compensates for a reduced climbing performance, resulting in higher cross-country speeds [9,12]. Thus, as soon as the thermal conditions are good enough to support the species with higher wing loadings (i.e. strong and wide thermals), we expect these larger soaring migrants to travel faster than the species with lower wing loadings. 2) Day Length Hypothesis Day length varies according to latitude and date, and might have an effect on migration speed for passerine birds through longer foraging times [14,15]. Similarly, the time for soaring is longer for longer days, and we thus may expect that day length has a positive effect on the number of travelling hours and hence, on distance covered per day for thermal soaring migrants. 3) Spring Arrival Hypothesis During spring migration the goal of adult individuals is to reach their breeding grounds and start reproduction early and therefore, competition for early arrival may promote a faster migration during spring compared to autumn [16]. 4) Time Constraint Hypothesis Species migrating overall longer distances (.5000 km, like the osprey and the Western marsh-harrier; Table 1) are expected to be more time-selected than short-distance migrants [17,18], and thus should travel faster, achieving higher daily distances. The two latter hypotheses both suggest that birds use behavioural strategies to increase migration speed and thus reduce the duration of migration, depending on both season (prediction 3) and migration distance (prediction 4) and, as a consequence, the most extreme time-selected strategies are expected for adult individuals of long-distance migrants during spring migration [19]. Moreover, our aim was also to study how meteorological factors like wind conditions and thermal energy interact with the factors discussed in the previous hypothesis, in particular at a daily scale. Wind assistance is of paramount importance, since it can allow birds to increase their flight range irrespectively of fuel expenditure [17], thus saving both time and energy. Thermal convection, the rising movement of a portion of air, is affected by other meteorological factors (temperature gradient with altitude Table 1. Sample sizes and characteristics of the four species involved in this study. Species N ind N tracks spring N tracks autumn Body mass (kg) Wing span (m) Wing area (m 2 ) Aspect ratio Wing loading (kg/m 2 ) Migration distance (km) osprey Western marsh-harrier Egyptian vulture short-toed eagle Sources: [8,27,34,36,41 44]. doi: /journal.pone t001 PLoS ONE 2 July 2012 Volume 7 Issue 7 e39833

71 Factors Affecting Soaring Bird Migration Table 2. Summary of the hypothesis tested on soaring migrants crossing the Sahara desert. Hypothesis Wing loading Day length Spring arrival Time constraint Prediction Species with lower wing fly more hours, while species with higher wing loading show higher gliding speed between thermals The day length enhances the daily distance Adult individuals migrate faster during spring, due to breeding motivation Species migrating longer distances (.5000 km) travel faster than those migrating shorter distances doi: /journal.pone t002 determines atmospheric stability) as well as by topography. Climbing rate is of greatest importance for the resulting crosscountry speed of soaring migrants [20] and is dependent upon thermal convection irrespectively of birds morphology [8]. Therefore, we investigated the response of birds to these factors (day length, wind, thermal convection) at different temporal scales (hourly and daily). Results The migratory performance was evaluated for the raptors crossing of the Sahara desert along the flight routes shown in Figure 1. (a) Hourly Scale Concerning the circadian pattern of migration, speeds were generally slower early in the morning and late in the afternoon, being much higher around noon. They showed an asymmetrical distribution in relation to local noon (Figure 2a), with some differences between species. The highest cross-country speeds were attained soon after local noon, when speeds were most commonly in the range km/h (Figure 2a). In particular, Western marsh-harriers started movements earlier in the morning than the other species, showing higher speeds during the first of the six periods of the day (Figure 2a and 3; Kruskal-Wallis test; H 3, 443 = 63.78, p,0.001; see also figure 4) but flying slower just after midday (period four; Kruskal-Wallis H 3, 516 = 14.01, p = 0.003; see average values in Figure 3). The same difference in the fourth period also occurred when taking into account only travelling segments (with movement.5 km/h; cf methods), with the Western marsh-harrier being the slower species (Kruskal-Wallis H 3, 507 = 10.92, p = 0.012). Hourly flight speeds were significantly different among species only during spring (based on travelling segments within all time periods; ANOVA F 3,617 = 7.26, p,0.001; autumn: n.s.), with a Tukey post-hoc test revealing that this was due to the difference between ospreys and Egyptian vultures (p,0.001). Analyzing the hourly speed data (travelling segments) separately for each species showed that seasonal differences occurred in the osprey, being faster during spring (speed spring = 33.1 km/h vs. speed autumn = 29.7 km/h; F 1,509 = 7.3, p = 0.007), and in the Egyptian vulture, being faster in autumn (26.5 spring vs autumn km/h; F 1,671 = 7.9, p = 0.005). Figure 1. Migration tracks of raptors crossing the Sahara desert. Each segment connects two roosting locations. Spring tracks are on the left and autumn tracks on the right. The colours of the segments indicate the species as Osprey (blue), Marsh Harrier (green), Egyptian Vulture (orange), or Short-toed Eagle (red). Black arrows represent average wind directions (see Results). Sample sizes are given in table 1. doi: /journal.pone g001 PLoS ONE 3 July 2012 Volume 7 Issue 7 e39833

72 Factors Affecting Soaring Bird Migration Figure 2. Circadian pattern of speed and altitude. Speed (a; N = 2835) is given in km/h and altitude in meters above ground (b; N = 2659). Both parameters are plotted in relation to time of day (calculated as minutes from noon, see Methods for details). doi: /journal.pone g002 PLoS ONE 4 July 2012 Volume 7 Issue 7 e39833

73 Factors Affecting Soaring Bird Migration Figure 3. Boxplots showing speed (km/h) in relation to time of day categories (see Methods). The colours indicate the species as Osprey (blue), Marsh Harrier (green), Egyptian Vulture (orange) and Short-toed Eagle (red). For each time category the average speed (SD) is given, including both stopping and travelling segments. doi: /journal.pone g003 Altitude values showed a similar pattern as speeds, being asymmetrical with respect to midday, and the highest altitudes, commonly ranging between 200 m and 1500 m above the ground, were recorded during the four afternoon hours immediately after local noon (Figure 2b). Taking into account only altitude data related to travelling segments there was a significant variation among the four species (Kruskal-Wallis H 3, 1650 = 11.34, p = 0.01), with short-toed eagles flying on average lower than the other species (Mann-Whitney U test, all the three post-hoc comparisons p,0.05) while there were no significant altitude differences among flight altitude of ospreys, Western marsh-harriers and Egyptian vultures. (b) Daily Scale Concerning travelling hours, there were interspecific differences during both seasons (Table S1; ANOVA spring: F 3,128 = 18.17, p,0.001; autumn: F 3,242 = 6.95, p,0.001) with Western marshharriers and ospreys spending more time flying than the other two species during both seasons (Tukey test: p,0.05). The model concerning daily distance and including all data showed a significant effect of species and its interaction with season (Table 3). In particular, during spring there were significant differences among the four species (ANOVA, F 3,196 = 14.25, p,0.001), with the Egyptian vulture and the short-toed eagle travelling shorter daily distances than the osprey and the western marsh-harrier (Tukey test, p,0.05; see mean values in Figure 5). In contrast, during autumn, there were no significant interspecific differences (F 3,342 = 1.9, p.0.05). Concerning the species-specific models, in all species a positive effect of tailwind arose, and in all species except the short-toed eagle also season was significant (Table 4). Thus, for these species, when controlling for other factors, daily distances were significantly higher during spring than during autumn (Table 4). In the Egyptian vulture there were also significant effects of the interactions season*day length, and season*tailwind (Table 4). When investigating the nature of these interactions it appeared that the effect of day length on daily distance was negative during spring and positive during autumn. However, it should be noted that during spring this effect can be discarded as spurious, since PLoS ONE 5 July 2012 Volume 7 Issue 7 e39833

74 Factors Affecting Soaring Bird Migration Figure 4. Time budgets and sample sizes in relation to time of day (N = 2835). Coloured parts indicate travelling segments, while grey parts represent stopping segments. Morning includes period 1 and 2, midday period 3 and 4, and afternoon 5 and 6 (see Methods). doi: /journal.pone g004 day length correlated negatively with tailwind (R = 20.34, p = 0.001, N = 89), while during autumn there was no such relationship. Therefore, during spring tailwind had an overriding effect, whereas during autumn daily distance was affected by both following winds and day length. Concerning the interaction season*tailwind, the effect of tailwind was positive during both seasons, but stronger during autumn (slopes of regressions of daily distance vs. tailwind: spring = 9.8 km/day for every meter/sec of tailwind; autumn = 18.5). (c) Spatial and Seasonal Variation of Environmental Factors Average absolute wind conditions (direction and speed) were similar in the two seasons (spring: 36u691.2, m/sec; autumn: 26u686.3, m/sec; Figure 1). Tailwinds were positively related to latitude during spring and negatively during autumn (spring: R = 0.42, p = , N = 200; autumn: R = 20.32, p = , N = 347; Figure 6), this is to say that during both seasons wind assistance was stronger for the last part of the Sahara crossing. Furthermore, winds were always more favourable during autumn than during spring, in all four species (all four comparisons p,0.006; Figure 6). Day length was longer during spring than during autumn for Western marsh-harriers and short-toed eagles, and conversely for the Egyptian vulture (all p,0.001; Figure 7), while there was no difference for the osprey. Therefore, the Egyptian vulture was the only species experiencing more favourable factors simultaneously (tailwind and day length) in the same season (autumn). Discussion In this study we compared the behaviour and migratory performance of different bird species migrating across the same regions and during the same periods. Our results showed that migratory performance is remarkably similar across species that differ greatly in morphology and breeding range. Despite these general similarities in performance, we also found that some responses to external factors are season- and species-specific. Influence of Morphology In agreement with our predictions based on morphology, the Western marsh-harrier was the species initiating its daily flight first, showing fast movements early in the morning, when the other species were still mostly at their night roosting sites. However, in agreement again with predictions based on morphology, the harriers travelled at lower speeds than the other larger raptor species after midday, which is the best time for soaring flight (Figure 2a and 3). Thus, the three heavier species seem to take advantage of the stronger thermals that occur during the central hours of the day whereas the Western marsh-harrier exploit also weaker thermal conditions or use flapping flight early in the morning to a higher degree than the other species. By starting migration earlier and thus flying more hours per day (Figure 2a), Western marsh-harriers covered the same daily flight distance as for example the osprey (Figure 5), despite a slower hourly speed at noon when thermals were strongest. Overall, both hourly speed and altitude showed the same diurnal distribution, peaking just after midday (Figure 2), presumably due to a time lag between irradiation by the sun and the development of thermals. Influence of Day Length We found only little support for our second prediction that day length would have a positive effect on daily distance. In the Egyptian vulture, during spring, winds had an overriding effect, and as wind were more favourable in the northern part of the crossing (Figure 6), where days were shorter, the strong positive effect of wind hides any effect of day length. During autumn, there was no such correlation between wind and day length, but both factors seemed to enhance daily distance simultaneously, increasing the flight range of Egyptian vultures. Unexpectedly, for the short-toed eagle, there was no relationship between daily distance and day length in any season, despite this species is a typical soaring migrant [8]. In addition, there was no seasonal difference in daily distance for this species. A explanation for the lack of day length and seasonal effects in this species would be that they were non-breeders, probably less time-constrained and hence not fully motivated to use the whole daily thermal soaring window to reach PLoS ONE 6 July 2012 Volume 7 Issue 7 e39833

75 Factors Affecting Soaring Bird Migration Figure 5. Mean daily distances (±1 SE; N = 547 segments) per season and species. Sample sizes are given in table S2. doi: /journal.pone g005 their destinations as fast as possible. Thus, in the future it would be interesting to analyze the behaviour of adult short-toed eagles, perhaps more prone to take advantage of the whole daily flight window and thus more dependent on day length. Influence of Thermal Conditions Considering the time of the day as a proxy of the strength of thermal currents, we showed that, at an hourly scale, thermal convection had a profound effect on the diurnal variation of speed and altitude (Figure 2). The large variation in flight altitude shows wide scatters also within the same periods (Figure 2b). Raptors constantly change altitude from a few hundred metres above ground, where they start to climb in a new thermal, up to high altitudes, sometimes approaching 2000 m above the ground, Table 3. GLM mixed model with daily distance as dependent variable, including all data. Source Denominator df F Sig. Intercept Species Season Species*Season doi: /journal.pone t003 where they leave the thermal to glide off in the migratory direction. At the daily scale, we excluded thermal energy from the models (see Methods) and we found a strong effect of winds on all species and seasons. Our findings are of broad interest, but it must be considered that they are related only to the Sahara desert, where thermal conditions are probably generally very good (being thus less likely than wind to be an important agent affecting daily distance) and therefore our analyses fail to demonstrate a clear effect of thermal variation on the daily scale. In this scenario, it is interesting to highlight that Mandel et al. [21], found regional differences between different migratory flyways of Turkey vultures (Cathartes aura) in North America, with individuals flying over flat plains being more affected by wind conditions than individuals following mountain chains, which were more thermal constrained. Chevallier et al. [22] reported that wind did not significantly affect the daily distance covered by migrating black storks (Ciconia nigra) during trans-saharan soaring migration, but comparison with our results is difficult due to the differences in statistical approaches. Differences in Travel Speed between Species and Seasons Controlling for wind assistance, that was less favourable during spring, at a daily scale the three species involving adult individuals (i.e. excluding short-toed eagles) were faster during spring (Table 4), in agreement with our third prediction (spring migration strategy promoting higher speeds) [16]. This effect was particularly evident for birds that had to cover longer migration distances, (i.e. PLoS ONE 7 July 2012 Volume 7 Issue 7 e39833

76 Factors Affecting Soaring Bird Migration Table 4. Final GLM models for each species, with daily distance as dependent variable. Wald Sig. Spring Autumn osprey Intercept Tailwind Season (13) 265 (9.3) Western marsh-harrier Intercept Tailwind Season (20.1) (15) Egyptian vulture Intercept Tailwind Season (18.5) (11.4) Day length Day length*season Tailwind*Season short-toed eagle Intercept Tailwind Season n.s (8.2) (8.2) For each season, the marginal means of daily distance (km/day; with SD in parentheses), corrected for effects of other significant factors besides season (tailwind, day length), are given). See Methods for details. doi: /journal.pone t004 ospreys and Western marsh-harriers), in agreement with our fourth prediction (individuals migrating longer distances are more time constrained). Perhaps this pattern is a consequence of a tradeoff between speed and total duration of migration in respect to the overall migration distance [18], with species facing a longer distance being more time-selected [17]. How did the osprey and Western marsh-harrier achieve higher speeds in spring, after accounting for the effects of wind and other environmental factors, in comparison with autumn and with the other two species? Our results suggested that, while in the case of the Western marshharrier the higher daily distances in spring were obtained mostly because of more time spent flying, for the osprey there was also an effect of higher flight speeds. One way of increasing cross-country speed among soaring migrants is to use a strategy of mixed soaring and flapping flight and using partially powered glides, as for example common cranes (Grus grus) commonly do [23]. The gain in resulting cross-country speed by this strategy comes at an increased cost of energy consumption per unit distance covered compared to pure soaring migration, as evaluated by Pennycuick et al. [23]. It seems likely that ospreys, and perhaps also Western marsh-harriers, adopt such a strategy in order to reduce time to complete their spring migration (at a cost of increased energy consumption) while Egyptian vultures and short-toed eagles behave more strictly as pure soaring migrants. More detailed studies are needed to confirm and evaluate if and to what degree flapping flight is included in the soaring migratory behaviour among different seasons and among different species during their Sahara crossing. Models revealed that tailwind was by far the best predictor of daily distance, since all species during both seasons were affected by the tailwinds that they experienced during their journeys. In agreement with Shamoun-Baranes et al. [24] tailwinds were more favourable in autumn compared to spring, but this pattern was only reflected in the daily distances covered by the Egyptian vulture, that during autumn also took advantage from longer days (Figure 7). The fact that ospreys and Western marsh-harriers were faster during spring, despite winds being less favourable, highlights the stronger time pressure during this season, where the birds seem to have adopted a strategy to achieve high travel speeds in order to reach the breeding grounds as soon as possible (prediction 3) [16]. Overall, differences among the four species were smaller than expected on the basis of their behaviour when crossing other regions, like the Mediterranean Sea, where the species showed more differences concerning their tendency to cross open water and consequently also in their flight behaviour [25 27]. All the species, and especially the three ones with higher wing loading behaved as true soaring migrants, restricting their flight activity to the daily hours when thermals developed. This behaviour is in contrast to species, which in addition to soaring flight, regularly use sustained flapping flight during their Sahara crossing such as the Eurasian hobby (Falco subbuteo), the Eleonora s falcon (Falco eleonorae) and the lesser kestrel (Falco naumanni) that extend their travelling time also during the night to gain considerably longer daily flight range (sometimes.500 km/day) [5,28 29]. Interspecific differences arose mainly in spring, when osprey and Western marsh-harriers were faster than Egyptian vultures and short-toed eagles, perhaps because of a combined effect of longer total migration distance and breeding pressure to migrate faster (prediction 3 and 4). Future studies including individuals of different species from the same breeding areas and belonging to the same age class could improve our knowledge concerning which factors determine migratory performance of long-distance migrating birds, possibly also using new devices which would allow studying movements at a finer temporal scale than the one presented here. A detailed understanding of the behaviour of migrating birds when crossing the Sahara desert region could help to disentangle the adaptations developed to cope with scarcity of food and water and the harsh weather conditions [30]. This is particularly important also in the light of the deleterious effects on fitness that these birds can experience thousands of kilometres away, at their breeding grounds, in case of adverse conditions met during the crossing of this ecological barrier [11]. In this scenario, it should be noted that, even if there is no clear indication of desert expansion [31], sand storms are becoming much more frequent [32]. In conclusion, our results suggested that, while thermal strength was the most important factor on a hourly basis, at the daily scale wind was more important, allowing birds to significantly increase their daily flight ranges over the Sahara desert. The effect of day length was much less marked if it had any effect at all. The impact of winds upon migration speed raises interesting questions about future effects of global change on migratory perfomance, at least in the light of the predictions concerning worldwide wind patterns [33]. Materials and Methods (a) Selection of Migration Data We used data from four raptor species tagged with 22 g, 30 g or 45 g ARGOS/GPS PTT-100 from Microwave Telemetry Inc. (horizontal accuracy of locations: 618 m), in Spain, Italy and Sweden between 2006 and The dataset included 22 individuals completing 73 journeys (Table 1, Figure 1). Egyptian vultures were tagged as adults in Spain [34], and short-toed eagles as nestlings in Spain and Italy [35,27]. Ospreys and Western marsh-harriers were tagged as adults in Sweden [4,36]. Hence, all PLoS ONE 8 July 2012 Volume 7 Issue 7 e39833

77 Factors Affecting Soaring Bird Migration Figure 6. Spatial and temporal variation of wind assistance (all species pooled). The maps show daily segments. Colours indicate the extent of wind support: green = tailwind (.2.5 m/s), blue = weak wind (2.5 to 2.5 m/s), red = headwind (, 2.5 m/s). The bars show the frequency of different tailwind classes in different latitudinal bands. doi: /journal.pone g006 the tagged individuals apart from the short-toed eagles were experienced breeding adults. As there were no differences in the results of the analyses for short-toed eagles during their first autumn migration (age: juvenile) and the subsequent ones (age: immature), we pooled all the autumn migration data of this species for further analyses. To select the data, we considered as Sahara desert the latitudinal band between 31u and 17.5u N, a journey of ca km. We included the data between the last roosting site before entering the desert and the first one after the desert crossing. Three sampling units were identified: (i) hourly flight speed; (ii) altitude above ground (accounting for an hourly scale); and (iii) daily distance (accounting for a daily scale) [37]. The hourly flight speed was calculated at two-hour intervals, while the daily distance was calculated as the loxodromic distance between two consecutive roosting sites. A few days with daily distance,50 km were excluded from the analyses [38]. The altitude above ground was obtained for the same two-hour intervals that were used to calculate hourly flight speed, by subtracting the ground altitude (obtained from geonames.org/) from the bird s altitude recorded by the transmitter (nominal accuracy 622 m) at the beginning of the segment. (b) Time and Day Length In order to standardize the time of two-hour segments irrespectively of longitudinal and seasonal differences we calculated the lapse from local noon considering the median time (for speeds) or the ending time (for altitudes). These values were thus negative before local noon and positive afterwards. The time of local noon for each location was calculated according to the PLoS ONE 9 July 2012 Volume 7 Issue 7 e39833

78 Factors Affecting Soaring Bird Migration Figure 7. Day length experienced by the birds during each daily segment, plotted against Julian day. The same variables are expressed also as box-plots. doi: /journal.pone g007 formulae provided by NOAA ( programs/eap/models.html). These time values were grouped into six time categories according to a step of 140 minutes, from early morning ( 420, 280), until late afternoon (281, 420), through the intermediate categories ( 420, 280; 279, 140; 139,0; 1, 140; 141, 280; 281, 420). Day length was expressed as that one experienced by each individual throughout a given day, thus as the difference (in minutes) between the sunset at the roosting site of arrival and the sunrise at the roosting site of departure. The exact times of sunrise and sunset for each roosting site (when sun is degrees below horizon) were calculated using the abovementioned formulae. (c) Meteorological Variables Data were obtained from the NCEP/NCAR Reanalysis project, as provided by the NOAA/OAR/ESRL PSD, Boulder, CO, USA ( using R package RNCEP [39]. Since only four measurements per day were available, we used meteorological variables only for the analyses at the daily scale, collecting the data at 12:00 hours for the midpoint of each daily segment. Wind data (U and V wind components) were extracted for a pressure level of 925 hpa, which corresponds to an altitude of about 750 m a.s.l, and thus was the closer measurement to the cruising altitude of our tracked birds [38]. The U and V wind components were combined into single wind vectors from which we calculated a tailwind component in relation to the overall direction of each journey as the wind speed multiplied by cosine for the angle between wind and migration direction. Conditions characterized by tailwind component.2.5 m/s were designated as following winds, tailwind component between 2.5 and 2.5 m/s as weak winds or crosswinds and tailwind component, 2.5 m/s as opposed winds (Figure 6). Concerning thermal energy, we performed exploratory analyses testing the effect of two different variables: 1) a temperature gradient (uc/100 m) following Chevallier et al [22], and 2) the velocity of thermal convection (w*), following Bohrer et al. [7]. In both cases, results were non-significant or driven by spurious patterns (e.g., negative relationship of w* with tailwind) leading to a negative effect of thermal conditions on daily speed. We do not present here those results for reasons of space constraints and we run the final models on daily distance excluding any thermal variable. (d) Statistical Analyses After an inspection of the frequency distribution of flight speeds at the two-hour intervals, we divided the segments either as stopping or travelling according to a threshold of 5 km/h [5,28 29,40]. Travelling hours were computed as the elapsed time between the onset-time of the first travelling segment of the day and the end-time of the last one, and only for days when the transmitter was working continuosly (thus the sample size was lower), in order to detect accurately the beginning and the end of the daily flight. To identify which factors shaped the daily distance, we first ran a General Linear Mixed Model including all data, with season, species and their interaction as fixed factors, and the individual as random factor. Finally, we ran four separate General Linear Models, one for each species, including, besides season, also tailwind, and day length, as well as the interaction of each one of these terms with season (a total of five terms). Nonsignificant terms were removed stepwise from the models according to their p-value within the model, starting from the interactions, until we obtained models that retained only significant variables. However, if a factor was significant only within an interaction, it was not removed as a single term. Since exploratory analyses showed that, at least in some cases (Table S2), there was an effect of latitude on daily distance, and that latitude was also correlated with wind assistance (see Results), we excluded latitude from the analyses in order to avoid spurious results. The analyses were carried out using SPSS PLoS ONE 10 July 2012 Volume 7 Issue 7 e39833

79 Factors Affecting Soaring Bird Migration Supporting Information Table S1 Daily travelling hours (sample size, average and standard deviation) for each species by season. (DOC) Table S2 Regression coefficient and significance of the relationship between daily distance and latitude for each species by season. (DOC) Acknowledgments L. Bolonio, R. Dekkers, J. De Lucas, P. García, V. García, M. Hake, R. Ibáñez, A. Izquierdo, E. Mallia, S. Morant, M. Nieto, P. Olofsson, J.M. References 1. Alerstam T, Hedenström A (1998) The development of bird migration theory. J Avian Biol 29: Thorup K, Alerstam T, Hake M, Kjellén N (2003) Bird orientation: compensation for wind drift in migrating raptors is age dependent. Proc R Soc Lond B 270 (Suppl): S8 S Klaassen RHG, Ens BJ, Shamoun-Baranes J, Exo K-M, Bairlein F (2011) Migration strategy of a flight generalist, the Lesser Black-backed Gull Larus fuscus. Behav Ecol 23: Klaassen RHG, Strandberg R, Hake M, Alerstam T (2008) Flexibility in daily travel routines causes regional variation in bird migration speed. Behav Ecol Sociobiol 62: López-López P, Limiñana R, Mellone U, Urios V (2010) From the Mediterranean Sea to Madagascar. Are there ecological barriers for the longdistance migrant Eleonora s falcon? Land Ecol 25: Mellone U, López-López P, Limiñana R, Urios V (2011) Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int J Biometeorol 55: Borher G, Brandes D, Mandel J, Bildstein K, Miller T, et al. (2012) Estimating updraft velocity components over large spatial scales: Contrasting migration strategies of golden eagles and turkey vultures. Ecol Lett 15: Spaar R (1997) Flight strategies of migrating raptors: a comparative study of interspecific variation in flight characteristics. Ibis 139: Hedenström A (1993) Migration by Soaring or Flapping Flight in Birds: The Relative Importance of Energy Cost and Speed. Phil. Trans. R. Soc. Lond. B 342: Kerlinger P (1989) Flight Strategies of Migrating Hawks. Chicago: the University of Chicago Press. 11. Strandberg R, Klaassen RHG, Hake M, Alerstam T (2009) How hazardous is the Sahara Desert crossing for migratory birds? Indications from satellite tracking of raptors. Biol Lett 6: Pennycuick CJ (2008) Modelling the Flying Bird. London: Academic Press. 13. Spaar R, Bruderer B (1997) Migration by flapping or soaring: Flight strategies of Marsh, Montagu s and Pallid Harriers in southern Israel. Condor 99: Kvist A, Lindström Å (2000) Maximum daily energy intake: it takes time to lift the metabolic ceiling. Physiol Biochem Zool 73: Bauchinger U, Klaassen M (2005) Longer days in spring than in autumn accelerate migration speed of passerine birds. J Avian Biol 36: Kokko H (1999) Competition for early arrival in migratory birds. J Anim Ecol 68: Alerstam T (2003) Bird migration speed. In: Berthold P, Gwinner E, Sonnenschein E, editors. Avian Migration. Berlin: Springer-Verlag Strandberg R, Alerstam T, Hake M, Kjellén N (2008) Short-distance migration of the Common Buzzard Buteo buteo recorded by satellite tracking. Ibis 151: Alerstam T (2006) Strategies for the transition to breeding in time-selected bird migration. Ardea 94: Pennycuick CJ (1978) Fifteen testable predictions about bird flight. Oikos 30: Mandel JT, Bohrer G, Winkler DW, Barber DR, Houston CS, et al. (2011) Migration path annotation: cross-continental study of migration-flight response to environmental conditions. Ecol Appl 21: Chevallier D, Handrich Y, Georges JY, Baillon F, Brossault P, et al. (2010) Influence of weather conditions on the flight of migrating black storks. Proc R Soc B 277: Perez-Garcıa, R. Silvaggi, K. Thorup, Á. Vela, M. Visceglia, M. Ström- Eriksson, A. Niklasson and L. Niklasson helped during field work. Author Contributions Conceived and designed the experiments: UM RHGK TA. Performed the experiments: UM RHGK TA. Analyzed the data: UM RHGK TA. Contributed reagents/materials/analysis tools: MV. Wrote the paper: UM. Performed field work: UM RHGK CGR RL PLL DP RS VU. Provided corrections and modifications to the text: RHGK CGR RL PLL DP RS VU MV TA. 23. Pennycuick CJ, Alerstam T, Larsson B (1979) Soaring migration of the Common Crane Grus grus observed by radar and from an aircraft. Ornis Scand 10: Shamoun-Baranes J, Baharad A, Alpert P, Berthold P, Yom-Tov Y, et al. (2003) The effect of wind, season and latitude on the migration speed of white storks Ciconia ciconia, along the eastern migration route. J Avian Biol 34: Bildstein KL (2006) Migrating raptors of the world. Cornell Univ Press. 26. Agostini N, Panuccio M (2010) Western Marsh harrier (Circus aeruginosus) migration through the Mediterranean sea: a review. J Raptor Res 44: Mellone U, Limiñana R, Mallìa E, Urios V (2011) Extremely detoured migration in an inexperienced bird: interplay of transport costs and social interactions. J Avian Biol 42: Strandberg R, Klaassen RHG, Olofsson P, Alerstam T (2009) Daily travel schedules of adult Eurasian Hobbies Falco subbuteo - variability in flight hours and migration speed along the route. Ardea 97: Limiñana R, Romero M, Mellone U, Urios V (2012) Mapping the migratory routes and wintering areas of Lesser Kestrels Falco naumanni: new insights from satellite telemetry. Ibis 154: Schmaljohann H, Liechti F, Bruderer B (2007) Songbird migration across the Sahara: the non-stop hypothesis rejected! Proc R Soc B 274: Tucker CJ, Dregne HE, Newcomb WW (1991) Expansion and contraction of the Sahara Desert from 1980 to Science 253: Goudie AS, Middleton NJ (2006) Desert dust in the global system. Berlin: Springer. 33. Weimerskirch H, Louzao M, de Grissac S, Delord K (2012) Changes in wind pattern alter albatross distribution and life-history traits. Science 335: García-Ripollés C, López-López P, Urios V (2010) First description of migration and wintering of adult Egyptian vultures Neophron percnopterus tracked by GPS satellite telemetry. Bird Study 57: Mellone U, Yáñez B, Limiñana R, Muñoz AR, Pavón D, et al. (2011) Summer staging areas of non-breeding Short-toed Snake Eagles. Bird Study 58: Strandberg R, Klaassen RHG, Hake M, Olofsson P, Thorup K, et al. (2008) Complex temporal pattern of Marsh Harrier Circus aeruginosus migration due to pre- and post-migratory movements. Ardea 96: Mandel JT, Bildstein KL, Bohrer G, Winklera DW (2008) Movement ecology of migration in turkey vultures. Proc Natl Acad Sci USA 105: Klaassen RHG, Hake M, Strandberg R, Alerstam T (2011) Geographical and temporal flexibility in the response to crosswinds by migrating raptors. Proc R Soc Lond B. 278: Kemp MU, van Loon EE, Shamoun-Baranes J, Bouten W (2012) RNCEP: global weather andclimate data at your fingertips. Methods Ecol Evol 3: Shamoun-Baranes J, Bouten W, Camphuysen CJ, Baaij E (2011) Riding the tide: intriguing observations of gulls resting at sea during breeding. Ibis 153: Alerstam T, Rosén M, Bäckman J, Ericson PGP, Hellgren O (2007) Flight Speeds among Bird Species: Allometric and Phylogenetic Effects. PLoS Biol 5(8): e Alerstam T, Hake M, Kjellen N (2006) Temporal and spatial patterns of repeated migratory journeys by ospreys. Anim Behav 71: Bruderer B, Boldt A (2001) Flight characteristics of birds: I. Radar measurements of speeds. Ibis 143: Pavón D, Limiñana R, Urios V, Izquierdo A, Yañez B, et al. (2010) Autumn migration of juvenile short-toed eagles Circaetus gallicus from southeastern Spain. Ardea 98: PLoS ONE 11 July 2012 Volume 7 Issue 7 e39833

80 Table S1. Daily travelling hours (sample size, average and standard deviation) for each species by season. Spring Autumn N Mean S.D. N Mean S.D. osprey Western marsh-harrier Egyptian vulture short-toed eagle Table S2. Regression coefficient and significance of the relationship between daily distance and latitude for each species by season. Spring Autumn N R P N R P osprey Western marsh-harrier Egyptian vulture short-toed eagle

81

82 Section two: short range movements 80

83 VI 81

84

85 Bird Study (2012) 59, Wintering habitats of Eleonora s Falcons Falco eleonorae in Madagascar UGO MELLONE 1, PASCUAL LÓPEZ-LÓPEZ 1,2, RUBEN LIMIÑANA 1,3 and VICENTE URIOS 1 1 Estación Biológica Terra Natura, Grupo de Investigación Zoología de Vertebrados, Instituto Universitario de Investigación CIBIO, University of Alicante, Apdo 99, E 03080, Alicante, Spain, 2 Cavanilles Institute of Biodiversity and Evolutionary Biology, Terrestrial Vertebrates Group, University of Valencia, C/Catedrático José Beltrán 2, E Paterna, Valencia, Spain and 3 Instituto de Investigación en Recursos Cinegéticos (IREC), CSIC-UCLM-JCCM, Ronda de Toledo, s/n, E Ciudad Real, Spain Capsule Eleonora s Falcons wintering in Madagascar selected degraded humid forests and cultivated areas close to pristine humid forest. Aims To identify the habitat preferences of Eleonora s Falcon Falco eleonorae on their wintering grounds in Madagascar, and to use this information to gain insights into the conservation priorities of this species. Methods A total of 11 Eleonora s Falcons were captured in Spain in 2007 and 2008 and equipped with solar-powered satellite transmitters. We obtained information on five complete wintering events for three birds, two of them tracked for two consecutive years. Data were analyzed using geographic information system-based cartography. Results The analyses showed a preference for degraded humid forests and cultivated lands within areas where pristine humid forests were the most abundant habitat type. Conclusions Eleonora s Falcons could be taking advantage from a spill-over edge effect of their insect prey into cultivated and more open areas close to humid forests. However, the importance of humid forests for Eleonora s Falcons seems to be high. The current loss of this habitat in Madagascar is a cause for concern with respect to the conservation of this long-distance migratory falcon species. A consistent body of recent literature suggests that, in contrast to resident bird species, populations of migratory species can be affected by events occurring in different parts of the world (Newton 2004). Moreover, it has recently been shown that populations of long-distance migratory birds are declining at a faster rate than short-distance migrants or resident species (Sanderson et al. 2006). Therefore, understanding the interactions between different periods of the annual cycle in migratory birds becomes of the greatest importance for their conservation (Martin et al. 2007, Bowlin et al. 2010). Eleonora s Falcons Falco eleonorae are long-distance migratory raptors, which breed mainly in islands of the Mediterranean Sea and winter mainly in Madagascar (Walter 1979). The species is listed in the Annex I of the Directive 2009/147/EC and constitutes a priority species for conservation (BirdLife International 2010). Unfortunately, long-term series of comparable Correspondence author. ugomellone@libero.it demographic data are lacking, but on the basis of recent surveys, the species seems to be stable or steadily increasing across its whole distribution range (Del Moral 2008, Dimalexis et al. 2008). Although the migratory routes have recently been mapped in detail (Gschweng et al. 2008, López-López et al. 2009, López-López et al. 2010, Mellone, López-López et al. 2011), little is known about the species habitat requirements on its wintering grounds. Such information is key for the conservation of Eleonora s Falcons, since actions for habitat preservation, especially at wintering grounds, remain largely unimplemented (BirdLife International 2010). Our knowledge of the ecology of migratory birds has been constrained by the difficulties in tracking individuals across seasons. Several methods such as traditional ringing and banding have been employed in recent decades (Berthold 2001). Nevertheless, advances in the miniaturization of satellite tracking devices are now making it possible to determine the connections between the different seasonal grounds of birds in a Q 2012 British Trust for Ornithology

86 30 U. Mellone et al. way unimaginable few years ago (Wikelski et al. 2007). Other methods, such as genetic markers, stable isotopes (Hobson 2008) and geolocators (Rodríguez et al. 2009), although complementary, do not allow fine-tuned habitat selection analyses that are possible with satellite tracking. Nevertheless, wintering-focused analyses performed with this technology have received little attention in comparison with the greater body of research dealing with migratory routes, and hence very few studies have been published so far (Gerkmann & Meyburg 2009, Jiguet et al. 2011). The island of Madagascar has been identified as one of the most important biodiversity hotspots in the world, which is strongly endangered owing to several threats mainly habitat loss, including deforestation (Myers et al. 2000, Harper et al. 2007, Irwin et al. 2010). In this context, the identification and conservation of optimal habitats for long-distance migratory species that winter in Madagascar is of the utmost importance. Besides benefits for the species itself, it could be useful also for wider conservation aims, given that raptors prove to be, in certain conditions, good biological indicators (Sergio et al. 2005, 2008). With the aim of identifying habitats where conservation actions should be focused, we analyze here the habitat preferences of Eleonora s Falcons wintering in Madagascar. According to the International Action Plan for the species, the location of the wintering areas and the description of their ecological characteristics are urgently needed and have been considered an essential priority (BirdLife International 1999, 2010). METHODS Study species Eleonora s Falcons are colonial raptors, which breed mainly on Mediterranean islands, from Spain to Greece, including several islands located in the central Mediterranean (Italy), and the northern coast of Africa (Algeria and Libya) (Walter 1979). They also breed in a few colonies located in the Atlantic Ocean, including the Canary Islands (Spain) and Mogador (Morocco). The bulk of the population (80 90%) breeds in Greece (Dimalexis et al. 2008). Winter records are almost exclusively restricted to Madagascar (Walter 1979, Gschweng et al. 2008, Kassara et al. 2011), although some observations (mainly juveniles) have been recorded on the eastern coast of Africa (Kenya, Tanzania and Mozambique) (Gschweng et al. 2008, authors unpubl. data). Since individuals belonging to different populations concentrate in such a narrow winter range, a comprehensive mapping and a complete description of habitat requirements at wintering areas is crucial in order to gain insight into conservation priorities for the species as a whole. In fact, to the best of our knowledge, only sporadic visual observations of Eleonora s Falcons in Madagascar have been reported so far (Thorstrom & René de Roland 2000, Zefania 2001). In this context, information based even on a small sample size of tracked birds could be useful, at least as a starting point for further investigations, which could be also applied to fieldwork planning. Animal tagging and transmitter terminal programming A total of 11 Eleonora s Falcons were trapped in Balearic (39830 N3800 E) and Columbretes Islands (39855 N 0840 E), located in the western Mediterranean (Spain) (Fig. 1). Birds were trapped in autumn 2007 and 2008, and were equipped with Microwave Telemetry Inc. 9.5-g solar-powered transmitter terminals (López-López et al. 2009, López-López et al. 2010). During the wintering season, transmitters were programmed to collect data (coordinates, date and time) on a duty cycle of 12 hours on/58 hours off. Since the transmitters provide data of different accuracy, only high-quality locations (location classes 1 3; maximum error, 1 km) were used for the analyses (Soutullo et al. 2007). Here, we only used wintering data from those individuals for which we recorded the complete wintering season (between arrival on the wintering grounds and the onset of the following pre-breeding migration). We used data belonging to three adult individuals, two of them tracked for two consecutive years (i.e. five different wintering events). Locations obtained less than one hour after the previous one were excluded from the analyses to avoid spatial autocorrelation and, when more than one location was available within an hour, we used the one of highest quality (Limiñana et al. 2008). On average, each wintering event accounted for 235 locations (range: , n ¼ 5) spanning from November to April. Habitat preferences and kernel estimation To map the wintering areas in detail, we estimated the individuals home-range core area of every wintering event through fixed kernel density contours (sensu Worton 1989), using the Animal Movement Extension Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

87 Wintering Eleonora s Falcons in Madagascar 31 Figure 1. (a e) Wintering locations ( ) of three Eleonora s Falcons tracked by satellite-telemetry in northern Madagascar. Range sizes (km 2 ) according to kernel density contours for five wintering events are as follows: (a) individual #80399, wintering event 2008/09; (b) #80399, wintering event 2009/10; (c) #80402, wintering event 2008/09; (d) #80402, wintering event 2009/10; (e) #80400, wintering event 2008/09. Dark green area represents humid forest, while pale green area represents degraded humid forest (following Moat and Smith [2007]). (f) Location of Eleonora s Falcon s tagging sites in the western Mediterranean and the wintering areas in Madagascar. for ArcView, version 3.2 (Hooge & Eichenlaub 2000). We calculated the 25%, 50%, 75% and 95% fixed kernels using the least-squares cross-validation (LSCV) procedure (Silverman 1986). We chose the 50% kernel to represent the core areas after a detailed exploratory analysis, since it represents a reasonable trade-off between including a sufficient number of locations excluding outliers and, moreover, it allows comparisons with similar studies (Mellone, Yáñez et al. 2011). In our case, nearly 60% of overall locations were included within the core areas (range: 42 93%; Fig. 1). To assess individual fidelity to the wintering grounds, we calculated the inter-year overlap of the core areas of the two birds tracked during consecutive years. In order to describe habitat composition of wintering areas, we calculated the percentage of every habitat type within the core areas for each one of the five wintering events separately, using the geographic information system layers produced by the Madagascar Vegetation Mapping Project ( These layers are based on the combination of remote sensing technology and field data collected by expert botanists ( ), and are the most accurate vegetation maps of Madagascar that are readily available (Moat & Smith 2007). Despite the ongoing changes in land cover in the country, it is unlikely that it could alter our results, owing to the short time-lapse between map compilation and the collection of our falcon tracking data. In fact, the deforestation rate of Madagascar forests has decreased from the beginning of the 1990s, in comparison with previous decades (Harper et al. 2007). We performed a habitat selection analysis to check whether Eleonora s Falcons occurred more frequently in certain habitats than expected by chance. To this end, we calculated a minimum convex polygon (MCP) for each wintering event (thus, three for the first year and two for the second one), which represents the maximum extension of available habitat, from which Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

88 32 U. Mellone et al. birds actively select the preferred habitats (Worton 1995). We then generated 1000 randomly distributed points within each one of the five MCPs using Hawth s analysis tools for ArcMap 9.2 (Beyer 2004), and assigned a habitat class to every random point, as well to as every real recorded location. To determine whether a bird preferred any particular habitat we used Monte Carlo tests (Manly 1997) to compare the observed frequency of real locations in different habitat types with the expected ones (Cadahía et al. 2010). The expected frequencies were obtained by sampling 1000 times, the same number of real bird locations from the list of random locations (Soutullo et al. 2008). Each individual wintering event was analyzed separately to avoid statistical pseudoreplication and no analysis including all the five wintering events at the same time was conducted. In addition, the analyses were first carried out for diurnal and nocturnal locations separately. Since the results did not differ between diurnal and nocturnal locations, we then pooled day and night data for further analyses, and here we present the results based on the complete data set. Monte Carlo analyses were conducted using Excel s Pop- Tools 3.1 (Hood 2010). Pair-wise comparisons were twotailed and the critical alpha level was set at P, RESULTS Eleonora s Falcons wintering areas were located in northern Madagascar (Fig. 1), where birds spent about five months of the year. Taking into account the core areas (50% fixed kernel contours), birds used one or two areas each wintering season (Fig. 1), with a median overall area of 606 km 2 (range: km 2, n ¼ 5; Table 1). The two birds tracked during two consecutive wintering seasons used areas in their second year that they had used in the previous year, indicating site-fidelity. Bird #80399 used a smaller area during the second year that was completely within the one used in the first year (Fig. 1). On the other hand, bird #80402 used areas of similar size in both years (Table 1), with an overlap of 78.4% (Fig. 1). Humid forest was the most abundant habitat within the core areas (pooling all data: 52.4% of the total surface area; Table 1). Habitat selection analyses showed that within individuals MCPs, Eleonora s Falcons actively selected some habitat types, showing preferences for degraded humid forests and cultivations (P, in the five cases; Table 1) and avoiding grasslands and humid forest (P, in the five cases; Table 1). Furthermore, a detailed inspection of maps showed that birds selected degraded humid forests and cultivated areas close to pristine humid forests (Fig. 1). DISCUSSION Wintering habitats of Eleonora s Falcons Our results show that Eleonora s Falcons wintered in northern Madagascar, and were mostly located on the humid eastern slopes of the island, rather than on the dry western ones (Moat & Smith 2007) (Fig. 1). The eastern region of Madagascar is amongst the highest areas of the country, which accounts for the high rainfall levels that occur there. In tropical rainforests, higher rainfall levels may result in high numbers of insects (Wolda 1978) and consequently, falcons may select these sites rather than dry lower areas of the island owing to higher insect prey availability. Eleonora s Falcons showed clear winter site-fidelity in consecutive years. Although this finding is based on a small sample size (two birds in only two consecutive years), our results are in agreement with previous reports concerning other raptor species during the wintering season (Alerstam et al. 2006, Strandberg et al. 2008, García-Ripollés et al. 2010, Limiñana et al. 2011), which suggest that falcons select areas at least in part on the basis of knowledge acquired in previous years. In addition, all three individuals used approximately the same area during the first wintering season (Fig. 1), which might suggest some degree of bird association during winter. This is in agreement with Zefania (2001), who reported Eleonora s Falcons associating in groups of 3 21 in Madagascar. These results, together with those of individuals tracked from other populations (Sardinia and Greece) which spent the winter in northern Madagascar (Gschweng et al. 2008, Kassara et al. 2011), suggests that our findings are consistent across years and might be applied to a larger proportion of Eleonora s Falcons during the wintering season in Madagascar. These results also suggest that the strength of migratory linkages (i.e. migratory connectivity) between breeding ranges and wintering grounds among different European populations of Eleonora s Falcon should be the subject of future research. Within the core wintering areas roughly half of the habitat was pristine humid forest. However, results of habitat selection analyses showed an active preference by Eleonora s Falcons for degraded humid forest and cultivated areas (Table 1). Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

89 Wintering Eleonora s Falcons in Madagascar 33 Table 1. Extension of the distinct habitats (km 2 ) encompassed within the 50% kernel contours and significance levels of the habitat selection analysis (Monte Carlo tests) of three Eleonora s Falcons during five wintering events in Madagascar. Individual ID Wintering event 2008/ / / / /10 Total Habitat Area % P Area % P Area % P Area % P Area % P Area % Water x x x ns ns Bare soil/rock x x ns Cultivation (+) (+) (+) (+) (+) Western dry forest ns ns ns Plateau grassland wooded mosaic ( ) ( ) ( ) ( ) ( ) Wooded grassland bushland ( ) ( ) ( ) ( ) ( ) Wetlands ns x ns Humid forest ( ) ( ) ( ) ( ) ( ) Degraded humid forest (+) (+) (+) (+) Total , habitat is selected;, habitat is avoided; significances are neglected when the given habitat had less than ten observed locations; x, habitat was not represented in any random point of the corresponding wintering event. Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

90 34 U. Mellone et al. The pristine humid forests of Madagascar are multilayered evergreen woodlands with a canopy that can reach m high. Since the 1970s, 33.4% of this forest has been destroyed, and repeated clearing has often made regeneration impossible. Currently, 39% of the area of pristine forest is formally protected (Moat & Smith 2007). The structure and ecology of degraded humid forests varies enormously, ranging from nearly intact forest to wooded grassland open areas (Moat & Smith 2007). Our results showed that Eleonora s Falcons selected degraded humid forests close to pristine humid forests. Although we cannot rule out the possibility that falcons hunt within pristine forests, our results suggest that they prefer forest edges, where they could perhaps be benefiting from a spill-over edge effect of prey insects. This could make these ecotonal areas more profitable for prey capture, with pristine humid forests acting as a source of prey (Rand et al. 2006) and the adjacent areas acting as sinks (Gibson et al. 2011). In agreement with our findings, Bildstein (2006) stated that migratory raptors wintering in the neotropics prefer edges and second-growth forests, rather than large pristine ones (Morel & Morel 1992) and, even within Madagascar, Watson et al. (2004) observed a significantly greater number of raptors on the edge of the littoral forest than in the core. It is likely, therefore, that the patches of pristine forests surrounded by a matrix of open areas could be exploited by Eleonora s Falcons, especially the edges between these two types of habitat. Finally, the results also showed a slight preference for cultivated areas, probably rice fields, where flying insects are abundant and easy to catch (Zefania 2001). Conservation implications It has been suggested that the loss of wintering habitat could negatively affect a population even more than the loss of breeding habitat (Sutherland 1996). Thus, conservation efforts addressed only to one season or stage of the life-cycle of this long-distance migratory species may be inefficient, being jeopardized by threats occurring in ranges completely different from the breeding ones (Martin et al. 2007). Our results suggest that the primeval humid forests of Madagascar and the close-by ecotonal areas are the most important habitat for Eleonora s Falcons during the wintering season. The findings of the present study should be of concern in the light of the rapid decline of humid forests in Madagascar, which could disappear by 2067 according to current predictions (Moat & Smith 2007). Our findings, although based on a limited sample, fill an important gap in the knowledge of the ecology of Eleonora s Falcons and could be used to focus conservation efforts and further field studies. In a global context, pristine forests currently attract the strongest protection efforts (Bermingham et al. 2005, Gibson et al. 2011), but our results, which are in agreement with other recent research on the value of tropical secondary forest (Chazdon et al. 2009), show that second growth forests, peripheral and ecotonal areas, especially when close to more pristine areas, deserve attention for the conservation of Eleonora s Falcons. ACKNOWLEDGEMENTS The Terra Natura Foundation and the Servicio de Biodiversidad from the Conselleria de Medio Ambiente, Agua, Urbanismo y Vivienda (Generalitat Valenciana) funded this project. Special thanks are due to all the people who helped in bird trapping. We also thank C. García-Ripollés, an anonymous referee and Phil Whitfield for their useful comments on an earlier draft of the manuscript. U. Mellone is supported by FPU grant of the Spanish Ministry of Education (reference AP ). R. Limiñana has a postdoctoral grant (reference 10/12-C) co-funded by Consejería de Educación y Ciencia (JCCM) and the European Social Fund. P. López- López is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Science and Innovation (reference JCI ). This paper is part of the PhD dissertation of U. Mellone at the University of Alicante. This study complies with the current laws in Spain. REFERENCES Alerstam, T., Hake, M. & Kjellén, N Temporal and spatial patterns of repeated migratory journeys by ospreys. Anim. Behav. 71: Beyer, H.L Hawth s analysis tools for ArcGIS Available at: (accessed 20 April 2010). Bermingham, E., Dick, C.W. & Moritz, C Tropical Rain Forests: Past, Present and Future. Chicago, IL: University of Chicago Press. Berthold, P Bird Migration. A General Survey. New York: Oxford University Press. Bildstein, K.L Migrating Raptors of the World. Ithaca, NY: Cornell University Press. BirdLife International International Species Action Plan: Eleonoras s Falcon Falco eleonorae. BirdLife International Eleonora s falcon Species Action Plan implementation review. Available at: org/webx/.2cba720c (accessed 20 October 2010). Bowlin, M.S., Bisson, I-A., Shamoun-Baranes, J., Reichard, J.D., Sapir, N., Marra, P.P., Kunz, T.H., Wilcove, D.S., Hedenström, A., Guglielmo, C.G., Åkesson, S., Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

91 Wintering Eleonora s Falcons in Madagascar 35 Ramenofsky, M. & Wikelski, M Grand challenges in migration biology. Integr. Comp. Biol. 50: Cadahía, L., López-López,P.,Urios,V.&JoséNegro, J.J Satellite telemetry reveals individual variation in juvenile Bonelli s eagle dispersal areas. Eur. J. Wildl. Res. 56: Chazdon, R.L., Peres, C.A., Dent, D., Sheil, D., Lugo, A.E., Lamb, D., Stork, N.E. & Miller, S.E The potential for species conservation in tropical secondary forests. Conserv. Biol. 23: Del Moral, J.C El Halcón de Eleonora en España. Población en y método de censo. Madrid: SEO/BirdLife. Dimalexis, A., Xirouchakis, S., Portolou, D., Latsoudis, P., Karris, G., Fric, J., Georgiakakis, P., Barboutis, C., Bourdakis, S., Ivovic, M., Kominos, T. & Kakalis, E The status of Eleonora s Falcon (Falco eleonorae) in Greece. J. Ornithol. 149: García-Ripollés, C., López-López, P. & Urios, V First description of migration and wintering of adult Egyptian vultures Neophron percnopterus tracked by GPS satellite telemetry. Bird Study 57: Gerkmann, B. & Meyburg, B.-U Habitats used by Lesser Spotted Eagles (Aquila pomarina) during migration and wintering as revealed by Satellite tracking and remote sensing. Popul. Ecol. Raptors. Owls. 6: Gibson, L., Lee, T.M., Koh, L.P., Brook, B.W., Gardner, T.A., Barlow, J., Peres, Bradshaw C.J.A., Laurance, W.F., Lovejoy, T.E. & Sodhi, N.S Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478: Gschweng, M., Kalko, E.K.V., Querner, U., Fiedler, W. & Berthold, P All across Africa: highly individual migration routes of Eleonora s falcon. Proc. R. Soc. Lond. B. 275: Harper, G., Steininger, M., Tucker, C., Juhn, D. & Hawkins, F Fifty years of deforestation and forest fragmentation in Madagascar. Environ. Conserv. 34: Hobson, K.A Using endogenous and exogenous markers in bird conservation. Bird Conserv. Int. 18: Hood, G.M Poptools, version 3.1. Available at: (accessed 10 October 2010). Hooge, P.N. & Eichenlaub, B Animal Movement Extension to Arcview, version 2.0. Anchorage, AK: US Geological Survey. Irwin, M.T., Wright, P.C., Birkinshaw, C., Fisher, B., Gardner, C.J., Glos, J., Goodman, S.M., Loiselle, P., Rabeson, P., Raharison, J.L., Raherilalao, M.J., Rakotondravony, D., Raselimanana, A., Ratsimbazafy, J., Sparks, J., Wilmé, L. & Ganzhorn, J.U Patterns of species change in anthropogenically disturbed forests of Madagascar. Biol. Conserv. 143: Jiguet, F., Barbet-Massin, M. & Chevallier, D Predictive distribution models applied to satellite tracks: modelling the western African winter range of European migrant black storks. J. Ornithol. 152: Kassara, C., Fric, J. & Sfenthourakis, S Home-range and resource use by Eleonora s falcon (Falco eleonorae) in its wintering quarters inferred by satellite telemetry data. In Fusani L., Coppack T. & Strazd M. (eds), Proceedings of the 8th Conference of the European Ornithologists Union: 181. Latvian Ornithological Society, Riga, Latvia. Available at: (accessed 5 September 2011). Limiñana, R., Soutullo, A., López-López, P. & Urios, V Pre-migratory movements of adult Montagu s Harriers Circus pygargus. Ardea. 96: Limiñana, R., Soutullo, A., Urios, V. & Reig-Ferrer, A Migration and wintering areas of adult Montagu s Harriers (Circus pygargus) breeding in Spain. J. Ornithol. online first: DOI: /s x López-López, P., Limiñana, R. & Urios, V Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool. Stud. 48: López-López, P., Limiñana, R., Mellone, U. & Urios, V From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Landsc. Ecol. 25: Manly, B.F.J Randomization, Bootstrap and Monte Carlo Methods in Biology, 2nd edn. London: Chapman & Hall. Martin, T.G., Chadès, I., Arcese, P., Marra, P.P, Possingham, H.P. & Ryan Norris, D Optimal conservation of migratory species. PLoS ONE 2(8): e751. Mellone, U., López-López, P., Limiñana, R. & Urios, V Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int. J. Biometeorol. 55: Mellone, U., Yáñez, B., Limiñana, R., Muñoz, A.R., Pavón, D., González, J.M., Urios, V. & Ferrer, M Summer staging areas of non-breeding Short-toed Eagles. Bird Study ifirst: DOI: / Moat, J. & Smith, P Atlas of the Vegetation of Madagascar. London: Kew Publishing. Morel, G.J. & Morel, M.Y Habitat use by Palaearctic migrant passerine birds in West Africa. Ibis 134: Myers, N., Mittermeier, R.A., Mittermeier, C.G., de Fonseca, G.A.B. & Kent, J Biodiversity and hotspots for conservation priorities. Nature 403: Newton, I Population limitation in migrants. Ibis 146: Rand, T.A., Tylianakis, J.M. & Tscharntke, T Spillover edge effects: the dispersal of agriculturally subsidized insect natural enemies into adjacent natural habitats. Ecol. Lett. 9: Rodríguez, A., Negro, J.J., Bustamante, J., Fox, J.W. & Afanasyev, V Geolocators map the wintering grounds of threatened Lesser Kestrels in Africa. Divers. Distrib. 15: Sanderson, F.J., Donald, P.F., Pain, D.J., Burfield, I.J. & van Bommel, F.P.J Long-term population declines in Afro- Palearctic migrant birds. Biol. Conserv. 131: Sergio, F., Newton, I. & Marchesi, L Top predators and biodiversity. Nature 236: 192. Sergio, F., Caro, T., Brown, D., Clucas, B., Hunter, J., Ketchum, J., McHugh, K. & Hiraldo, F Top predators as conservation tools: ecological rationale, assumptions, and efficacy. Annu. Rev. Ecol. Syst. 39: Silverman, B.W Density Estimation for Statistics and Data Analysis. Chapman and Hall, London. Soutullo, A., Cadahía, L., Urios, V., Ferrer, M. & Negro, J.J Accuracy of lightweight satellite telemetry: a case study in the Iberian Peninsula. J. Wildl. Manage. 71: Soutullo, A., Urios, V., Ferrer, M. & López-López, P Habitat use by juvenile Golden Eagles Aquila chrysaetos in Spain. Bird Study 55: Strandberg, R., Klaassen, R.H.G., Hake, M., Olofsson, P., Thorup, K. & Alerstam, T Complex temporal pattern of Marsh Harrier Circus aeruginosus migration due to pre- and postmigratory movements. Ardea 96: Sutherland, W.J Predicting the consequences of habitat loss for migratory populations. Proc. R. Soc. Lond. B. 263: Thorstrom, R. & René de Roland, L.-A Status and conservation of raptors on the Masoala Peninsula, Madagascar. In Chancellor, Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

92 36 U. Mellone et al. R.D. & Meyburg, B.-U. (eds), Raptors at Risk: 35 41, World Working Group on Birds of Prey, Berlin; Hancock House, Blaine, WA. Walter, H Eleonora s Falcon. Adaptations to Prey and Habitat in a Social Raptor. London: The University of Chicago Press. Watson, J.E.M., Whittaker, R.J. & Dawson, T.P Habitat structure and proximity to forest edge affect the abundance and distribution of forest-dependent birds in tropical coastal forests of south-eastern Madagascar. Biol. Conserv. 120: Wikelski, M., Kays, R., Kasdin, J., Thorup, K., Smith, J.A., Cochran, W.W. & Swenson, G.W. Jr Going wild what a global small-animal tracking system could do for experimental biologists. J. Exp. Biol. 210: Wolda, H Seasonal fluctuations in rainfall, food and abundance of tropical insects. J. Anim. Ecol. 47: Worton, B.J Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70: Worton, B.J Using Monte Carlo simulation to evaluate kernel-based home range estimators. J. Wildl. Manage. 59: Zefania, S. (ed.) Observation of sooty and Eleonora s falcons in Madagascar. In Anonymous, Wings Over Africa. Proceedings of the International Seminar on Bird Migration: Research, Conservation, Education and Flight Safety: , Society for Protection of Nature, Tel Aviv, Israel. (MS received 24 June 2011; revised MS accepted 12 October 2011) Q 2012 British Trust for Ornithology, Bird Study, 59, 29 36

93 VII 91

94

95 Bird Conservation International, page 1 of 8. BirdLife International, 2013 doi: /s Summer pre-breeding movements of Eleonora s Falcon Falco eleonorae revealed by satellite telemetry: implications for conservation UGO MELLONE, PASCUAL LÓPEZ-LÓPEZ, RUBÉN LIMIÑANA and VICENTE URIOS Summary Recent advances in bird tracking technologies are revealing that migratory birds use temporal staging sites other than breeding and wintering areas, and these areas deserve conservation efforts. Eleonora s Falcon Falco eleonorae is a long-distance migratory raptor that breeds colonially on islands and is considered a priority species for conservation. Anecdotal observations indicate that during the pre-breeding period, Eleonora s Falcons stay in inland areas far away from the colonies, but, to date there are no detailed data concerning the connectivity between these areas and breeding colonies. Using satellite telemetry, we analysed data from four summering events belonging to three individuals breeding in two colonies in the Western Mediterranean (Spain). All of them made inland movements in areas up to c.400 km distant from the respective breeding colonies, visiting several habitats, from forests to arable lands, probably taking advantage of high densities of insects. Perturbations occurring in these areas could threaten Eleonora s Falcons with serious consequences at the population level. We suggest that conservation measures implemented at breeding and wintering grounds may not suffice and that temporary staging areas should be identified at a larger scale and deserve protection as well. Introduction In a traditional, simplified view, migratory birds alternate short-distance non-directional movements, occurring in breeding and wintering areas, with long-distance directional movements, namely spring and autumn migrations (Newton 2008). Recent advances in tracking technologies are revealing that this scenario is rather more complex (Klaassen et al. 2011), with species also performing pre-migratory movements, spending some weeks in staging areas away from breeding sites (Limiñana et al. 2008, Strandberg et al. 2008, Catry et al. 2011), using stopover areas for relatively long periods (Chevallier et al. 2010, Kochert et al. 2011, Limiñana et al. 2012), or using staging areas different to breeding or wintering ranges before reaching sexual maturity (Mellone et al. 2011a). To date, pre-breeding movements have been described only for some seabirds (Phillips et al. 2007, Guilford et al. 2009, Bogdanova et al. 2011). In this scenario, migratory species require conservation measures that are difficult to implement, since efforts carried out in one area can be jeopardised by threats occurring far away (Martin et al. 2007, Angelov et al. 2012). Therefore, the identification of the different areas used throughout the year is of utmost importance from a conservation point of view. Eleonora s Falcon Falco eleonorae is a long-distance migratory raptor, which breeds colonially almost exclusively in islands of the Mediterranean Sea (Walter 1979). The species is included in Annex I of the Directive 2009/147/EC on the conservation of wild birds and constitutes a priority species for conservation (BirdLife International 2004, 2010). The winter range of the species is

96 U. Mellone et al. 2 concentrated in Madagascar (Walter 1979, Mellone et al. 2012a) and migration routes have been recently mapped in detail (Gschweng et al. 2008, López-López et al. 2009, 2010, Mellone et al. 2011b, Kassara et al. 2012). This species presents a unique adaptation, since it delays reproduction to late summer (August- September) in order to take advantage of the passage of migratory passerines en route to Africa, which form its main prey during the breeding season (Walter 1979). Nevertheless, timing of spring migration is similar to that of other Palearctic species, since Eleonora s Falcons normally arrive in Europe in May (Bernis and Castroviejo 1968, Walter 1979). Therefore, for these birds there is a time lapse of two months between the end of the spring migration and the beginning of the breeding season. There are some data indicating that during this period Eleonora s Falcons stay in inland areas far away from the breeding colonies (González et al. 1984, Ristow and Wink 1995, Cano 2001, Belenguer et al. 2004, Gregory 2007), but satellite tracking data concerning this behaviour are practically absent and therefore it has been impossible to unravel the connectivity between those staging areas and the breeding colonies (but see Gregory 2007). Here we analyse the pre-breeding movements of Eleonora s Falcons belonging to two different colonies in the western Mediterranean using Argos-based satellite telemetry. Methods Eighteen Eleonora s Falcons were trapped in Balearic and Columbretes Islands (Spain) in autumn between 2007 and 2010 (Fig. 1; details in López-López et al. 2009, 2010). Birds were sexed by molecular methods and equipped with Microwave Telemetry Inc. 9.5-g solar-powered satellite transmitters using Teflon ribbon. During the summer season, transmitters were programmed to collect data on a duty cycle of 12 h on/58 h off, apart from bird #92532 that partly transmitted with a cycle of 12 h on/18 h off. Here, we only used summer data from those individuals for which we recorded the complete summering event (since arriving in the Mediterranean basin until the onset of the following autumn migration). Overall, we used data from four summering Figure 1. Locations outside the colonies of three Eleonora s Falcons during four summering events occurring between 2009 and Bird#80399: n 5 10 locations; bird #80402: n 5 49 (2009) / n 5 28 (2010); bird #92532: n

97 Pre-breeding movements of Eleonora s Falcon 3 events belonging to three individuals: two adult females captured in 2008 (#80399 from Dragonera, Mallorca and #80402 from Columbretes islands) and one immature (2 nd calendar-year) male born in Columbretes Islands in 2010 (#92532). Bird #80402 was tracked for two consecutive summers. For the remaining tagged individuals, the transmitters did not work properly or birds died during migration (López-López et al. 2010, Mellone et al. 2012a). Locations were collected using the Argos system. The locations are coded for accuracy and grouped into location classes (l.c. hereafter). In order of decreasing accuracy these classes are 3, 2, 1, 0, A, B and Z. Initially all data apart from those with a l.c. of were included. Data were then filtered to exclude locations obtained in the same hour (retaining the highest quality one) and locations resulting in unrealistic values of speed and direction (e.g.. 80 km/h during the day and. 5 km/h at night). In order to identify the centroid of the areas used during summer outside the breeding colonies and to calculate its distance from colonies, we selected all the locations occurring on the mainland and calculated 50% fixed kernel density contours (Worton 1989) for every summering event, using the Animal Movement Extension for ArcView 3.2 (Hooge and Eichenlaub 2000) and the least squares cross validation (LSCV) procedure (Silverman 1986). In order to describe habitat composition, we followed a conservative approach, excluding data with l.c. A and B, and applying to the remaining data (l.c. 3, 2, 1 and 0; n 5 86) a buffer whose radius was the nominal accuracy of each location (l.c.3: 250 m; l.c. 2: 500 m; l.c. 1: 1,500 m; Argos 1996). For data with l.c. 0, nominal accuracy is higher than 1,500 m, therefore we conservatively applied an arbitrary buffer of 5,000 m. In our experience, after filtering for direction and speed (see above), those data are reasonably accurate (Soutullo et al. 2007, López-López et al. 2010, Mellone et al. 2011b, Limiñana et al. 2012). Within each buffer, we calculated the area of the different habitat categories, using the GIS layers of Corine Land Cover ( Since the buffer area differed according to the l.c. (e.g. a buffer whose centre was a l.c. 3 wassmallerthan a buffer with l.c. 2), we weighted the obtained surfaces according to the location class, in order to avoid a bias towards more inaccurate data. We compiled those percentages in Table 1. Results All individuals performed pre-breeding inland movements in areas hundreds of kilometres from the respective breeding colonies (Figure 1). Bird #80399 first visited the breeding colony located in Dragonera (Mallorca, Balearic Islands), between 20 May 2009 and 13 June It then spent 16 days in southern Spain, in the Sierra de Segura mountain range (Sierras de Cazorla, Segura y Las Villas Natural Park), 444 km from its breeding colony. Afterwards, it came back to the colony, where it stayed until the beginning of the autumn migration. Bird #80402 showed more complex movements during both summers that it was tracked. Its inland movements ranged from southern France to north-eastern Spain. During 2009 the first location within the Mediterranean basin was a low quality one in the Tyrrhenian Sea, close to central Italy (15 May). Then the falcon visited southern France (18 May), and before finally settling in the breeding colony (17 July) it alternated short visits to the colony (twice, of one and four days each) with inland movements to France (maximum period away from the colony of 31 days: 18 May to 17 June; 395 km away from the colony). During 2010 it again first visited central Italy (16 18 May) and Corsica (20 24 May), then from 26 May, it stayed mainly in southern France, including a first short visit (15 17 June) to the breeding colony. It finally settled in the colony on 1 July, but between the 23 and 27 July stayed inland in Castellón province (eastern Spain), the closest mainland area to the colony. During this year two core areas were identified (Figure 1), completely overlapping with the core area of the previous year, and located 440 and 331 km respectively from the breeding colony in Columbretes Islands. The immature bird (#92532) reached Spain on 24 June 2011, after having spent 10 days in northern Algeria, at the end of spring migration. Apart from two trips of four days each, one to northern Spain (central Pyrenees) and the other to southern Spain (Sierra Morena), it stayed in

98 U. Mellone et al. 4 Table 1. Percentages of CORINE land cover classes included within the buffer areas of satellite fixes with l.c. 3 0 (see methods) for each summering event of Eleonora s Falcons. Others includes: beaches, bare rocks, agro-forestry, sparsely vegetated and burnt areas, rice fields, inland and coastal water bodies (2009) (2010) average Artificial and urban surfaces Non-irrigated arable land Permanently irrigated land Rice fields Vineyards Fruit trees and berry plantations Olive groves Pastures Annual crops associated with permanent crops Complex cultivation patterns Land principally occupied by agriculture, with significant areas of natural vegetation Agro-forestry areas Broad-leaved forest Coniferous forest Mixed forest Natural grasslands Moors and heathland Sclerophyllous vegetation Transitional woodland-shrub Others Sea Total eastern Spain until 9 August, when it began to move northwards, stopping for 10 days in the eastern pre-pyrenees and then for 12 days in a coastal areas of southern France. On 22 September, for the first time, it went back to its natal colony (Columbretes Islands), where it remained for roughly one month before beginning the autumn migration (25 October), including two short daily trips to inland areas (Castellón province). The centroid of its core area was located 173 km from the colony. Concerning land cover, the main habitats were coniferous and broad-leaved forests, followed by areas covered by sclerophyllous vegetation, transitional woodland-shrub, and cultivated areas such as fruit plantations and vineyards (Table 1). Discussion Overall, during the summer pre-breeding season Eleonora s Falcons visited a highly varying range of environments, spanning mountains (#80399) to lower hilly areas (#80402, #92532), to coastal areas (#92532), as well as the breeding colonies. The main habitats represented within the core areas were forests, shrubland, arable lands and vineyards (Table 1). After performing these large-range movements, adult individuals settled in the breeding colonies in July, and after that they no longer visited inland areas, beginning the southward migration from the colonies in October. Apart from the last wave of spring migrant passerines that occurs in May, in late spring-early summer, small Mediterranean islands cannot offer enough food to support hundreds of falcons, and therefore birds are forced to look for areas where feeding opportunities are more favourable (Ristow and Wink 1995). It is remarkable that the peak of insect

99 Pre-breeding movements of Eleonora s Falcon 5 abundance, particularly beetles (Order Coleoptera) as well as dragonflies and damselflies (Odonata), occurs in the inland areas visited by Eleonora s Falcons just during the pre-breeding period (Cano 2001, Belenguer et al. 2004), providing an adequate food supply until the beginning of the autumn migration of passerines that form their main resource during breeding (Walter 1979). It has been reported that Eleonora s Falcons prey particularly upon the Common Cockchafer Melolontha melolontha (Cano 2001, Belenguer et al. 2004), an abundant insect that shows demographic explosions in June in some years. Inland observations of Eleonora s Falcons occurring in August-September, during chick rearing, involve mainly immature, non-breeding individuals (Ristow and Wink 1995; authors unpubl. data). The pattern shown by our immature bird (#92532) fits well in this scenario, and its behaviour during the stay in the coastal area of southern France (movements over the open sea during early morning) might suggest that it was already chasing migrating passerines. Field observations confirm the importance of the areas identified in this study, suggesting that our findings may apply to a larger proportion of individuals. For example, within the mountain range of Sierra de Segura (southern Spain) a maximum of 11 Eleonora s Falcons were observed together in early July 2009, and six were recorded in the same period the previous year (C. Ruiz pers. comm. 2011). In addition, the species is regularly observed in southern France and Corsica (Carp and Cheylan 1979, Mayol 1996). Finally, bird #80402 briefly visited Tolfa hills (central Italy) in 2010 (and probably also in 2009, but not enough data were obtained). This area is the most important one in inland Italy concerning the number of Eleonora s Falcons regularly observed, with up to 18 individuals observed hunting insects together (Celletti and Meschini 1992). In inland areas of Mallorca (Spain), close to the colony of bird #80399, upto49 individuals have been observed hunting insects in the Albufera wetland in May (Ristow and Wink 1995), and the species is usually observed year after year in particular areas of the Serra de Tramuntana (Mayol 1977, 1996). Unfortunately no satellite locations have been obtained from our tracked individuals that could confirm movements to this area. Other areas where Eleonora s Falcons are regularly observed far away from the colonies occur elsewhere within its breeding range and in a wide range of habitats and altitudes (reviewed by Ristow and Wink 1995, see also Premuda and Mellone 2007, Díaz-Porteroet al. 2009). Not surprisingly, the highest numbers have been reported in mainland Greece and Crete (up to 120 individuals hunting insects together, Ristow and Wink 1995), the region where the breeding density is the highest worldwide (Dimalexis et al. 2007). The use of cultivated areas by Eleonora s Falcons could be a cause for concern, given the widespread use of pesticides in Spain and France (Geiger et al. 2010). Ristow (2001) showed that poison ingested by Eleonora s Falcons through drinking at polluted pools inland in Crete caused several deaths and a strong decline in a close-by breeding colony. It has been also shown that insect availability prior to egg-laying, even near the colonies, can affect clutch size (Xirouchakis et al. 2012). Therefore, events occurring in the prebreeding staging areas could have serious consequences at the population level and it is thus of paramount importance to identify and protect at least the areas where the highest concentrations of Eleonora s Falcons occur. Our results highlight that conservation schemes for migratory species should be focused not only at breeding and wintering grounds, but should also take into account temporary staging areas. Similar processes have been described for non-migratory species, when juveniles disperse and settle in temporary staging areas (see Penteriani et al. 2005, Soutullo et al. 2008, Cadahíaet al. 2010, Caroet al. 2011). However, it is very difficult to obtain these data and only recently, transmitters lasting more time than just one or two years, are allowing gain new insights into hitherto unknown processes such as the pre-breeding movements, as shown here. Acknowledgements The Terra Natura Foundation and the Servicio de Biodiversidad (Generalitat Valenciana) funded this project. Special thanks are due to all the people who helped in bird trapping. We thank Phil Atkinson for his useful comments on an earlier draft of the manuscript. We also thank

100 U. Mellone et al. 6 C. Ruiz for providing unpublished data. U. Mellone is supported by an FPU grant of the Spanish Ministry of Education (reference AP ). R. Limiñana has a postdoctoral grant (reference 10/12-C) co-funded by Consejería de Educación y Ciencia (JCCM) and the European Social Fund. P. López-López is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Economy and Competitiveness (reference JCI ). This paper is part of the Ph.D. dissertation of U. Mellone at the University of Alicante. This study complies with the current laws in Spain. References Angelov, I., Hashim, I. and Oppel, S. (2012) Persistent electrocution mortality of Egyptian Vultures Neophron percnopterus over 28 years in East Africa. Bird Conserv. Internatn. DOI: /S Argos (1996) User s manual. Toulouse, France: CLS/Service Argos. Belenguer, R., Tena, V. and Méndez, J. (2004) Halcón de Eleonora, el pirata de Columbretes. Quercus 224: Bernis, F. and Castroviejo, J. (1968) Aves de las islas Columbretes en primavera. Ardeola 12: BirdLife International (2004) Birds in Europe: population estimates, trends and conservation status. Cambridge, UK: BirdLife International. (BirdLife Conservation series No. 12.). BirdLife International (2010) Species factsheet: Falco eleonorae. Downloaded from on 03/11/2010 Bogdanova, M. I., Daunt, F., Newell, M., Phillips,R.A.,Harris,M.P.andWanless,S. (2011) Seasonal interactions in the blacklegged kittiwake, Rissa tridactyla: links between breeding performance and winter distribution. Proc. R. Soc. B 278: Cadahía, L., López-López, P., Urios, V. and Negro, J. J. (2010) Satellite telemetry reveals individual variation in juvenile Bonelli s eagle dispersal areas. Eur. J. Wildl. Res. 56: Cano, C. (2001) Halcones de Eleonor en la Serranía de Cuenca. Quercus 185: 36. Caro, J., Ontiveros, D., Pizarro, M. and Pleguezuelos, J. M. (2011) Habitat features of settlement areas used by floaters of Bonelli s and Golden Eagles. Bird Conserv. Internatn. 21: Carp, E., and Cheylan, G. (1979) Observations of the Eleonora s Falcon, Falco eleonorae,in southern France. Nos Oiseaux 35: Catry, I., Dias, M. P. Catry, T., Afanasyev, V., Fox, J., Franco, A. M. A. and Sutherland, W. J. (2011) Individual variation in migratory movements and winter behaviour of Iberian Lesser Kestrels Falco naumanni revealed by geolocators. Ibis 153: Celletti, S. and Meschini, A. (1992) Ulteriori dati sull estivazione del Falco della Regina Falco eleonorae nell alto Lazio. Alula I: Chevallier, D., Le Maho, Y., Brossault, P., Baillon, F. and Massemin, S. (2010) The use of stopover sites by black storks (Ciconia nigra) migrating between West Europe and West Africa as revealed by satellite telemetry. J. Orn. 152: Díaz-Portero, M. Á., Ramos, J. J., Paz de la Rocha, J. L., Jodar, P. A., Álvarez, B., Mancera, J. A., Delgado, G. and López Rondon, J. (2009) Observations of Eleonora s Falcon Falco eleonorae at Oukaimeden (High Atlas, Morocco). Go-South Bull. 6: Dimalexis, A., Xirouchakis, S., Portolou, D., Latsoudis, P., Karris, G., Fric, J., Georgiakakis, P., Barboutis, C., Bourdakis, S., Ivovic, M., Kominos T. and Kakali, s E. (2008) The status of Eleonora s falcon (Falco eleonorae) in Greece. J. Orn. 149: Geiger, F., Bengtsson, J., Berendse, F., Weisser, W. W., Emmerson, M., Morales, M. B., Ceryngier, P., Liira, J., Tscharntke, T., Winqvist, C., Eggers, S., Bommarco, R., Part, T.,Bretagnolle,V.,Plantegenest,M.,Clement, L. W., Dennis, C., Palmer, C., Onate, J. J., Guerrero, I., Hawro, V., Aavik, T., Thies, C., Flohre, A., Hanke, S., Fischer, C., Goedhart, P. W. and Inchausti, P. (2010) Persistent negative effects of pesticides on biodiversity and biological control potential on European farmland. Basic Appl. Ecol. 11: González, L. M., González, J. L. & Garzón, P. (1984) Observación de Halcón de Eleonor

101 Pre-breeding movements of Eleonora s Falcon 7 Falco eleonorae en el centro de España. Doñana Acta Vertebrata 11: Gregory, M. (2007) Nuevas citas de Halcón de Eleonor en el interior peninsular. Quercus 256: Gschweng, M., Kalko, E. K. V., Querner, U., Fiedler, W. and Berthold, P. (2008) All across Africa: highly individual migration routes of Eleonora s falcon. Proc. R. Soc. B 275: Guilford, T., Meade, J., Willis, J., Phillips, R. A., Boyle, D., Roberts, S., Collett, M., Freeman, R. and Perrins, C. M. (2009) Migration and stopover in a small pelagic seabird, the Manx shearwater Puffinus puffinus: insights from machine learning. Proc. R. Soc. B 276: Hooge, P. N. and Eichenlaub, B. (2000) Animal movement extension to Arcview. ver Anchorage: U.S. Geological Survey. Alaska Science Center - Biological Science Office. Kassara, C., Fric, J., Gschweng, M. and Sfenthourakis, S. (2012) Complementing the puzzle of Eleonora s Falcon (Falco eleonorae) migration: new evidence from an eastern colony in the Aegean Sea. J. Orn. 153: Klaassen, R. H. G., Ens, B. J., Shamoun- Baranes, J., Exo, K.-M., Bairlein, F. (2011) Migration strategy of a flight generalist, the Lesser Black-backed Gull Larus fuscus. Behav. Ecol. 23: Kochert, M. N., Fuller, M. R., Schueck, L. S., Bond, L., Bechard, M. J., Woodbridge, B., Holroyd, G. L., Martell, M. S., & Banasch, U. (2011) Migration patterns, use of stopover areas, and austral summer movements of Swainson s hawks. The Condor 113: Limiñana, R., Romero, M., Mellone, U. and Urios, V. (2012) Mapping the migratory routes and wintering areas of Lesser Kestrels Falco naumanni: new insights from satellite telemetry. Ibis 154: Limiñana, R., Soutullo, A., López-López, P. and Urios, V. (2008) Pre-migratory movements of adult Montagu s Harriers Circus pygargus. Ardea 96: López-López, P., Limiñana, R. and Urios, V. (2009) Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool. Stud. 48: López-López, P., Limiñana, R., Mellone, U. and Urios,V.(2010) From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Landscape Ecol. 25: Martin, T. G., Chades, I., Arcese, P., Marra, P. P, Possingham, H. P. and Ryan Norris, D. (2007) Optimal conservation of migratory species. PLoS ONE 2(8): e751. Mayol, J. (1977) Estudios sobre el Halcón de Eleonor, Falco eleonorae en las islas Baleares. Ardeola 23: Mayol, J. (1996) El Halcón de Eleonora (Falco eleonorae): Situación de la especie y de su conocimiento. Pp in J. Muntaner and J. Mayol, eds. Biología y conservación de las rapaces Mediterránea., Proceedings of the VI Congress on Biology and Conservation of Mediterranean Raptors, Palma de Mallorca, September Madrid: SEO/BirdLife. (Monografía nº 4). Mellone, U., López-López, P., Limiñana, R. and Urios, V. (2012a) Wintering habitats of Eleonora s Falcons Falco eleonorae in Madagascar. Bird Study 59: Mellone, U., Urios, V., Rguibi-Idrisi, H., Limiñana, R., Benhoussa, A. & López- López, P. (2012b) Ranging behaviour of Eleonora s Falcons Falco eleonorae during chick-rearing. Acta Orn. 47: Mellone, U., Yáñez, B., Limiñana, R., Muñoz, A. R., Pavón, D., González, J. M., Urios, V. and Ferrer, M. (2011a) Summer staging areas of non-breeding Short-toed Snake Eagles. Bird Study 58: Mellone, U., López-López, P., Limiñana, R. andurios,v.(2011b) Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int. J. Biometeorol. 55: Newton, I. (2008) The migration ecology of birds. London, UK: Academic Press, Elsevier. Penteriani, V., Otalora, F., Sergio, F. and Ferrer, M. (2005) Environmental stochasticity in dispersal areas can explain the mysterious disappearance of breeding populations. Proc. R. Soc. B 272: Phillips, R. A., Catry, P., Silk, J. R. D., Bearhop, S., McGill, R., Afanasyev, V. and Strange, I. J. (2007) Movements, winter distribution and activity patterns of Falkland

102 U. Mellone et al. 8 and brown skuas: insights from loggers and isotopes. Mar. Ecol. Prog. Ser. 345: Premuda, G. and Mellone, U. (2007) Presenza regolare del Falco della Regina, Falco eleonorae, sull isola di Marettimo (Egadi, Sicilia). Riv. Ital. Orn. 77: Ristow, D.(2001) Poison is causing the sudden population decline of Eleonora s Falcon,International Hawkwatcher 3: Ristow, D. and M. Wink (1995) Distribution of non-breeding Eleonora s Falcon (Falco eleonorae). Il Merill 28: Silverman, B. W. (1986) Density estimation for statistics and data analysis. London: Chapman and Hall.Soutullo, A., Cadahía, L., Urios, V., Ferrer, M. and Negro, J. J. (2007) Accuracy of lightweight satellite telemetry: a case study in the Iberian Peninsula. J. Wildlife Manage. 71: Soutullo, A., López-López, P. and Urios, V. (2008) Incorporating spatial structure and stochasticity in endangered Bonelli s eagle s population models: implications for conservation and management. Biol. Conserv. 141: Strandberg, R., Klaassen, R. H. G., Hake, M., Olofsson, P., Thorup, K. and Alerstam, T. (2008) Complex temporal pattern of Marsh Harrier Circus aeruginosus migration due to pre- and post-migratory movements. Ardea 96: Walter, H. (1979) Eleonora s falcon. Adaptations to prey and habitat in a social raptor. London, UK: University of Chicago Press. Worton, B. J. (1989) Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70: Xirouchakis, S. M., Fric, J., Kassara, C., Portolou, D., Dimalexis, A., Karris, G., Barboutis, C., Latsoudis, P., Bourdakis, S., Kakalis, E. and Sfenthourakis, S. (2012) Variation in breeding parameters of Eleonora s falcon (Falco eleonorae) and factors affecting its reproductive performance. Ecol. Res. 27: UGO MELLONE*, PASCUAL LÓPEZ-LÓPEZ, RUBÉN LIMIÑANA, VICENTE URIOS Vertebrates Zoology Research Group, CIBIO, University of Alicante, Edificio Ciencias III, Campus San Vicente del Raspeig s/n, Apdo. 99, E-03080, Alicante, Spain. RUBÉN LIMIÑANA Instituto de Investigación en Recursos Cinegéticos (IREC). CSIC-UCLM-JCCM. Ronda de Toledo, s/n. E Ciudad Real Spain. *Author for correspondence; ugomellone@libero.it Received 24 April 2012; revision accepted 6 October 2012

103 VIII 101

104

105 ACTA ORNITHOLOGICA Vol. 47 (2012) No. 2 SHORT NOTES Ranging behaviour of Eleonora s Falcons Falco eleonorae during chickrearing Ugo MELLONE 1, *, Vicente URIOS 1, Hamid RGUIBI-IDRISI 2, Rubén LIMIÑANA 1,3, Abdelaziz BENHOUSSA 4 & Pascual LÓPEZ-LÓPEZ 1 1 Terra Natura Biological Station, Vertebrates Zoology Research Group, CIBIO, University of Alicante, Apdo. 99, E Alicante, SPAIN 2 Faculty of Sciences, Department of Biology, Laboratory for the Evaluation of Natural Resources and Biodiversity, El Jadida, MOROCCO 3 Research Institute for Game Resources (IREC) CSIC-UCLM-JCCM, Ronda de Toledo, s/n. E Ciudad Real, SPAIN 4 Department of Biology, Faculty of Sciences, Mohammed V University, Avenue Ibn Battota B.P Rabat Agdal, MOROCCO *Corresponding author, ugomellone@libero.it Mellone U., Urios V., Rguibi-Idrisi H., Limiñana R., Benhoussa A., López-López P Ranging behaviour of Eleonora s Falcons Falco eleonorae during chick-rearing. Acta Ornithol. 47: DOI / X Abstract. The Eleonora s Falcon is a cliff-nesting raptor that breeds on isolated small islands adjusting its breeding season to coincide with the post-breeding autumn migration of its small passerine prey migrating over the sea, between late August and early October. Two adult female Eleonora s Falcons were equipped with Argos satellite transmitters during the chick-rearing period in Morocco giving the opportunity to study the ranging behaviour of the species during at least a part of the breeding season. Results showed that the falcons spent most of the time at sea during mornings, stayed mainly inland during afternoons, and rested in the colony during nights. Interestingly, although most distances were recorded shorter than 50 km away from the colony, movements took also place to areas located more than 100 km away. Locating and protecting these inland areas used for resting and foraging may be of interest for the conservation of the species in order to avoid perturbations such as poisoning and habitat destruction. Key words: raptors, time budget, conservation, satellite telemetry, Argos, Morocco Received June 2012, accepted Nov Knowledge of birds ranging behaviour during the breeding season is of great importance to assess conservation priorities. The Eleonora s Falcon Falco eleonorae is a long-distance migratory raptor that breeds colonially in small islands of the Mediterranean Sea, but also along the Atlantic Sea coast of Morocco and on the Canary Islands (Walter 1979). Eleonora s Falcons delay their breeding season to late summer in order to prey upon southbound migrating passerines flying over the open sea. Despite the species being currently globally listed as Least Concern (BirdLife International 2012), it is included in Annex I of the European Directive 2009/147/EC on the conservation of wild birds and hence, it constitutes a priority species for conservation along its whole distribution range (see also BirdLife International 2004). Many aspects of its biology have been studied to date, but there are no satellite tracking data concerning its ranging behaviour during the breeding season. This technology has recently been used to reveal the species migratory routes and main wintering habitats in Madagascar (Gschweng et al. 2008, López-López et al. 2009, 2010, Mellone et al. 2011, 2012, Kassara et al. 2012). However, transmission problems affecting the performance of the Argos system in the Mediterranean basin prevented the use of this method to study ranging behaviour of Eleonora s Falcons during the breeding season. Contrary to the GPS system, which collects location data always with a similar low spatial error, the Argos system collects data of varying quality, according to an estimated error (Soutullo et al. 2007). In the case of the Mediterranean basin, the amount of locations with high probability of large errors (> 10 km) is very high, thus precluding detailed ranging behaviour studies (Anonymous 2005).

106 196 SHORT NOTES On the other hand, previous studies carried out by visual observations (Walter 1979) or with an optical range finder (Rosén et al. 1999) did not clarify what are the flight distances that the falcons can cover to hunt during the breeding season: Rosén et al. (1999) provided detailed data on the hunting behaviour but only within a range of 4 km. Here we analyze movements of Eleonora s Falcons belonging to the breeding colony of Essa - ouira (Morocco), using satellite telemetry in order to describe time budgets and ranging distances. Two adult female Eleonora s Falcons were trapped on Essaouira Island (ca. 1 km of Essaouira, Morocco, N, 9.78 W) on the 20 September 2011, when chicks were days old (chicks leave the nest when days old; Wink & Ristow 2000). The birds were equipped with 9.5- gram Argos solar-powered satellite transmitters (Microwave Telemetry Inc.) affixed to their backs using a Teflon harness. The transmitters were programmed to collect data on a duty cycle of 12 on/18 off (for further methodological details see López-López et al. 2009, 2010). Here, the falcons were named by the number of the PTT placed on them (#80417 and #108376). Locations were collected using the Argos system and only high quality locations (location classes 3, 2, 1; maximum estimated error: 1 km; see Soutullo et al. 2007, Mellone et al. 2012) were considered for analyses. Then, data were filtered excluding locations obtained within the same hour (using only the highest quality one), to avoid temporal autocorrelation (López-López et al. 2010). In order to analyze habitat use in relation to the period of the day, data located within 1 km around the nest were classified as colony, while the other data were displayed on a Geographical Information System (GIS) and classified either as occurring at sea or inland (i.e., within mainland). Finally, locations were classified according to their UTC time as occurring during morning (6h 11h), afternoon (12h 18h) or night (19h 5h). Between tagging and the onset of autumn migration, a total of 131 high quality locations were collected and subsequently used for the analyses (69 and 62 for each bird, respectively). One of the two falcons (#80417) performed a long southbound trip of at least 181 km, staying away from the colony for almost two days (24 25 September) and visiting the Massa River (Morocco, Fig. 1). Mean distances of sea locations for the two falcons were 8.6 km for bird #80417 (SD = 13.2, max 52.9, N = 18) and 12.7 km for bird # (SD = 15.2, max 53.3, N = 29), Atlantic Ocean Fig. 1. Locations obtained for two Eleonora s Falcons breeding in the island of Essaouira. Individuals: stars #80417, circles #

107 SHORT NOTES 197 respectively. On the other hand, mean distances of inland locations were 14.1 km for bird #80417 (SD = 12.4, max 39.7, N = 12; excluding two locations during the long trip) and 14.3 km for bird # (SD = 8.7, max 28.5, N = 24), respectively. Overall, the birds spent more time at sea during the morning, while during the afternoon they stayed inland and during the night in the colony (χ 2 = 39.05, df = 4, p < 0.001; Fig. 2). The first of the two falcons departed on, #80417, began the autumn migration on the 23 October, while the second one first moved northwards on the 24 October and, after having spent eight days close to Rabat, 430 km away from the colony, began the southbound autumn migration on the 2 nd November. Both individuals ceased data transmission dur - ing the autumn migration for unknown causes. These are the first detailed data on the ranging behaviour of the Eleonora s Falcon during the breeding season. These results confirm that falcons can fly tens of kilometers away from the colony during the breeding period, as it was supposed according to the amount of time that individuals from the colony of San Pietro (Italy) spent outside the colony. Assuming continuous straight flight, on the basis of the period of absence from the nest, Rosén et al. (1999) estimated hunting excursions at an average distance of 24 km (maximum 70 km). Such long flights are not surprising given that Eleonora s Falcons are able to perform non-stop flights of hundreds of kilometers (Mellone et al. 2011). Long foraging distances during the chick rearing period have been rarely percentage (%) morning afternoon night colony sea inland Fig. 2. Time budget of two Eleonora s Falcons breeding in Essaouira island (Morocco). Locations were classified according to their UTC time as occurring during morning (6h 11h), afternoon (12h 18h) or night (19h 5h). Numbers given in the bars refer to the sample size of the locations. described for raptors (but see Meyburg & Meyburg 2009 and Houston et al. 2011), while it is well known that seabirds can make long foraging trips of hundreds of kilometers lasting for several days (e.g. Weimerskirch et al. 1993). During the breeding season Eleonora s Falcon prey mainly upon passerines in active migration (Walter 1979), hunting regularly at altitudes higher than 1000m, where passerines are more abundant and have lower chances of escape (Rosén et al. 1999). Since those passerines migrate mainly during night (Newton 2008), Eleonora s Falcon probably take advantage of the last wave of migrants early in the morning, as shown by the time budget reported here. Also Rosén et al. (1999) described that the colony is much more active before midday. On the other hand, inland locations, which occurred mostly over scrubland areas during the afternoon, were probably mostly due to search for alternative prey other than migrating passerines, such as insects or local/resting birds. Nevertheless, some locations that were recorded close to water bodies may suggest that falcons could perform inland trips also for drinking freshwater, as observed also on Mediterranean islands (U. Mellone pers. obs.). Similar to this, Walter (1979) reported sporadic observations of Eleonora s Falcons in inland Crete (20 30 km from Paximada colony) during windless days. Ristow (2001) showed that several deaths and a strong decline in a breeding colony close to Crete were caused by Eleonora s Falcons drinking water at inland polluted pools. Therefore, it would be important to identify and protect those areas where Eleonora s Falcons search for freshwater, especially considering the scarcity of this resource at the end of the summer (dry season). Spatial data on ranging behavior are particularly important for the conservation of the species. As shown here, movements during the chick rearing period may take place within areas ranging more than 100 km away from the nesting colonies, which should be considered when assessing the potential threats during the breeding period. Finally, additional data from different individuals and colonies are needed, preferably collected with a higher spatial (altitude) and temporal resolution, in order to assess the influence of hunting behaviour on breeding performance and population dynamics. ACKNOWLEDGEMENTS We thank the Haut commesariat aux Eaux et aux ForLts et B la lutte contre la desertification de Rabat for the the

108 198 SHORT NOTES permission and the help to perform the field work, the Iniative petites Iles méditérranéenne- PIM, projet Albatros" for their support and Fabrice Bernard, Michel Thevenot and Sebastian Renous for their help. We also thank R. Strandberg for the useful comments on the manuscript. U. Mellone is supported by FPU grant of the Spanish Ministry of Education (reference AP ). R. Limi - ñana had a postdoctoral grant (reference 10/12-C) co-funded by Consejería de Educación y Ciencia (JCCM) and the European Social Fund. P. López- López is supported by a Juan de la Cierva postdoctoral grant of the Spanish Ministry of Economy and Competitiveness (reference JCI ). This paper is part of the Ph.D. dissertation of U. Mellone at the University of Alicante REFERENCES Anonymous Argos performance in Europe. Tracker News 6: 8 newsletters/winter_05page8.pdf BirdLife International Birds in Europe: population estimates, trends and conservation status. BirdLife Conser - vation series No. 12, BirdLife International, Cambridge. BirdLife International Species factsheet: Falco eleonorae. (accessed 03/05/2012). Gschweng M., Kalko E. K. V., Querner U., Fiedler W., Berthold P All across Africa: highly individual migration routes of Eleonora s falcon. Proc. R. Soc. Lond. B 275: Houston C. S., McLoughlin P. D., Mandel J. Y., Bechard M. J., Stoffel M. J., Barber D. S., Bildstein K. L Breeding home ranges of migratory Turkey Vultures near their northern limit. Wilson J. Ornithol. 123: Kassara C., Fric J., Gschweng M., Sfenthourakis S Complementing the puzzle of Eleonora s Falcon (Falco eleonorae) migration: new evidence from an eastern colony in the Aegean Sea. J. Ornithol. 153: López-López P., Limiñana R., Mellone U., Urios V From the Mediterranean Sea to Madagascar. Are there ecological barriers for the long-distance migrant Eleonora s falcon? Landscape Ecol. 25: López-López P., Limiñana R., Urios V Autumn migration of Eleonora s falcon Falco eleonorae tracked by satellite telemetry. Zool. Stud. 48: Mellone U., López-López P., Limiñana R., Urios V Weather conditions promote route flexibility during open ocean crossing in a long-distance migratory raptor. Int. J. Biometeorol. 55: Mellone U., López-López P., Limiñana R., Urios V Wintering habitats of Eleonora's Falcons Falco eleonorae in Madagascar. Bird Study 59: Meyburg B. U., Meyburg C G.PS-Satelliten-Telemetrie bei einem adulten Schwarzmilan (Milvus migrans): Aufenthaltsraum während der Brutzeit, Zug und Überwinterung. Populationsökologie Greifvogel und Eulenarten 6: Newton I The migration ecology of birds. Academic Press, London. Ristow D Poison is causing the sudden population decline of Eleonora's Falcon. International Hawkwatcher 3: Rosén M., Hedenström A., Badami A., Spina F., Åkesson S Hunting flight behaviour of the Eleonora s falcon Falco eleonorae. J. Avian Biol. 30: Soutullo A., Cadahía L., Urios V., Ferrer M., Negro J. J Accuracy of lightweight satellite telemetry: a case study in the Iberian Peninsula. J Wildl. Manage. 71: Walter H Eleonora s falcon. Adaptations to prey and habitat in a social raptor. The University of Chicago Press. Weimerskirch H., Salamolard M., Sarrazin F., Jouventin P Foraging strategy of wandering albatrosses through the breeding season: A study using satellite telemetry. Auk 110: Wink M., Ristow D Biology and molecular genetics of Eleonora's Falcon (Falco eleonorae), a colonial raptor of Mediterranean Islands. In: Chancellor R. D., Meyburg B.- U. (eds). Raptors at Risk. World Working Group on Birds of Prey, Hancock House, pp STRESZCZENIE [Przemieszczanie się sokołów skalnych w okresie karmienia piskląt] Sokół skalny gniazduje na niewielkich wyspach Morza Śródziemnego i płn.-wsch. wybrzeżach Afryki. Jego późny okres rozrodu od końca sierpnia do początku października związany jest z terminem wędrówki jesiennej drobnych ptaków wróblowych. Sokoły chwytają przelatujące ptaki najczęściej nad otwartym morzem, ale dotychczasowe metody obserwacji nie pozwalały na określenie, na jakie odległości od gniazda mogą przemieszczać się polujące ptaki. Dwie samice sokoła skalnego, gniazdujące na wyspie Essaouira u wybrzeży Maroka (Fig. 1), wyposażono w nadajniki satelitarne. W analizach pominięto dane o słabej jakości (błąd większy niż 1 km) i te pochodzące z tej samej godziny, uzyskując łącznie 131 lokalizacji ptaków. Określoną pozycję ptaka przypisywano do jednej z trzech kategorii: nad morzem, w kolonii lęgowej, nad stałym lądem. Obserwacje podzielono na te z godzin porannych (godziny ), popołudniowych ( ) i nocnych ( ). Stwierdzono, że w godzinach porannych sokoły obserwowane są nad morzem, zaś w godzinach popołudniowych nad stałym lądem. Noc spędzają w kolonii (Fig. 2). Większość zanotowanych przemieszczeń w badanym okresie odbywała się w odległości do 50 km od gniazda, natomiast wykazano także takie na odległość ponad 100 km, w kierunku stałego lądu i nad stałym lądem (Fig. 1). Odnalezienie terenów będących celem tych dalekodystansowych przemieszczeń, używanych przez ptaki prawdopodobnie jako miejsca odpoczynku i dostarczające słodkiej wody do picia, jest szczególnie ważne z punktu widzenia ochrony gatunku.

109 CONCLUSIONS 107