Factors Influencing Nest-Site Selection by Spanish Imperial Eagles

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Digital.CSIC Factors inﬂuencing nest-site selection by Spanish Imperial Eagles

Isabelle A. Bisson,1,3 Miguel Ferrer,2 and David M. Bird1

1 Department of Natural Resource Sciences, McGill University, 21111 Lakeshore Road, Ste. Anne-de-Bellevue, Quebec, H9X 3V9, Canada 2 Estacioń Biolo´gica de Donãna C.S.I.C., Avenida de Mar´ı a Luisa s/n. Pabelloń del Peru´, 41013 Sevilla, Spain

ABSTRACT. We examined nest-site selection by Spanish Imperial Eagles (Aquila adalberti)inDonãna National Park (southwestern Spain) from 1984–1994. Spanish Imperial Eagles are large birds of prey that defend and occupy a territory year-round. Identifying a territory’s nest-site characteristics represents a critical step in safeguarding this endangered species’ habitat. We measured seven habitat variables describing microhabitat, vegetative composition, land use, and degree of human influence for 73 nest sites averaged for 14 occupied territories and 73 randomly selected sites for 14 unoccupied territories. Our best logistic regression model included two habitat variables representing microhabitat (height of nest tree) and human influence (distance to the nearest urban centre). The logistic regression model correctly classified 78.6% of the occupied and 85.7% of unoccupied territories. This analysis suggests that the presently occupied Spanish Imperial Eagle territories represent the majority of the remaining suitable habitat for this endangered species in and around Donãna National Park. There is an urgent need to create suitable nest sites for Spanish Imperial Eagles that would include tall trees in small tree stands away from intense human activity.

SINOPSIS. Factores que influyen en la seleccion de lugares de anidamiento de parte de Aguila adarberti Examinamos la seleccioń de nidos por parte del a´guila imperial (Aguila adarberti) en el Parque Nacional Donãna. Este estudio se llevo´ a cabo de 1984–1994. Estas a´guilas son grandes aves de presas que ocupan y defienden un territorio a trave´s de todo el anõ. El identificar las caracter´ı sticas del lugar de anidamiento representan un paso cr´ı tico para salvaguardad el habitat. Esta especie se encuentra en peligro de extincioń. Medimos siete variables en el habitat que describen el microhabitat, composicioń de la vegetacioń, uso del terreno y grado de influencia del humano para 73 lugares de anidamiento, en 14 territorios ocupados y 73 localidades selecionadas al azar para 4 territorios no ocupados. Nuestro mejor modelo log´ı stico de regresioń incluyo dos variables de habitat reprentando microhabitat (altura del a´rbol de anidamiento) e influencia humana (distancia al centro urbano ma´s cercano). Este modelo clasifico´ correctamente el 78.6 y el 85.7 de los territorios ocupados y no ocupados, respectivamente. Este ana´lisis sugiere que los territorios, actualmente ocupados por esta a´guila representan la mayoria del habitat adecuado remanente para esta especie en el Parque Nacional Donãna y en los alrededores de este. Hay una necesidad urgente de proveerles lugares adecuados para el anidamiento a esta a´guila. Esto debe inclir a´rboles altos en rodales de menor tamanõ y que esteń distantes de centros de alta actividad humana. Key words: Aquila adalberti, eagle territories, endangered species, habitat, human disturbance, Spain

Habitat loss has been identified as the prin- to a decrease in the number of available nest cipal factor affecting avian population declines sites (Martin 1993), it is crucial to quantify globally (Greenway 1967; Newton 1979; Rob- nesting habitat used by endangered birds. bins 1979). Thus, quantifying habitat require- The Spanish Imperial Eagle (Aquila adalber- ments for avian populations is an important ti) is one of the rarest raptors in the world and first step towards conservation efforts. Because is the most endangered Aquila species in Eu- nest-site availability may be the critical factor rope (Bijleveld 1974; Collar and Andrews determining habitat suitability in many open- 1988). Formerly found in Morocco, Portugal, nesting species (Martin 1993), and because ex- and Spain, the species is now restricted to tirpation of a species from an area may be due southwestern Spain (Gonza´lez et al. 1989). Threats to the survival of Spanish Imperial Ea- 3 Corresponding author. Current address: Depart- gles are numerous and primarily human-relat- ment of Biology, York University, 4700 Keele Street, ed; human persecution and electrocution by Toronto, Ontario K7L 3N6, Canada. Email: Ͻibisson@ electric power lines are the principal causes of sympatico.caϾ death (Ferrer and Hiraldo 1991). However, o habitat destruction, particularly for the Donãna spp., Cistus libanotis, and Erica spp. with scat- National Park population, has been identified tered cork oaks (Quercus suber), small stone as the primary factor limiting the species’ abun- pine (Pinus pinea) stands, and Eucaliptus spp. dance and distribution (Ferrer 1993). Protect- plantations. The marsh, which is flooded during Spanish Imperial Eagle habitat is regarded ing the winter and dry during the summer, is as the most urgent conservation need for this mainly covered by Scirpus spp., and the coastal highly endangered species, particularly given sand dunes are characterized by Ammophilia that these eagles occupy all-purpose, year-round arenaria, Corema album, and Juniperus phoeni- territories (Meyburg 1989; Ferrer 1993). cea. The climate is Mediterranean with an At- Habitat selection by the Spanish Imperial lantic influence (i.e., relatively warm dry sum- Eagle has been described primarily on a quali- mers and temperate rainy winters; Rogers and tative basis (Garzoń 1974; Meyburg 1975). Myers 1980). Gonzalez et al. (1992) provided the first quan- Nest-site characteristics. We sampled 73 titative description of the species’ habitat. How- nest sites (trees) in 14 occupied territories and ever, their study was conducted on a large scale 73 randomly selected trees in 14 unoccupied across the species’ entire range and did not pro- territories over 10 yr (1984–1994). Nests were vide information on microhabitat characteris- found through repeated searches within the tics. Furthermore, the habitat in and around study area conducted at the beginning of every Donãna National Park, where our study was breeding season (February to April). Nest conducted, is distinct from the habitat occupied searches were conducted by two to three sur- by other Spanish Imperial Eagles in Spain. The veyors from Donãna who identified territories Donãna area is the only coastal wetland inhab- by observing male territorial aggression and ited by this species; all other populations are courtship flight displays. Spanish Imperial Eagle found inland at higher elevations. It is therefore males aggressively defend their territory important to evaluate nest-site selection for the through conspicuous aerial displays (Ferrer Donãna population separately from other pop- 1993). Courtship flights are equally conspicu- ulations of Spanish Imperial Eagles. Moreover, ous. Once a territory was identified, surveyors this population provides an important study returned to a territory three or more times to group because of the intense and ongoing search for nests by observing nesting activities threat to the species’ habitat and the lack of (nest building and copulations). Because Span- expansion outside park boundaries, making it ish Imperial Eagles have year-round territories the densest (15–16 pairs in 20,000 ha) and and sometimes will not breed within an occu- most isolated Spanish Imperial Eagle popula- pied site, we designated a nest site only if at tion. Natal philopatry is high and recruitment least one egg was laid during the breeding sea- in this population is primarily from young re- son. We visited each nest three times during the turning to Donãna at three to five years of age breeding season to monitor reproductive per- (Ferrer and Calderoń 1990). Our goals were to formance. assess the factors which influence nest-site se- Using randomly computer-generated coor- lection by Spanish Imperial Eagles using a dinates, we located 14 unoccupied territories modelling approach and to provide a quanti- and selected a total of 73 random trees within tative description of Spanish Imperial Eagle the unoccupied territories on a 1:50,000 land habitat use in and around Donãna National use map of the study area (Ministry of Agri- Park from 1984 to 1994. culture). Random sites were at least 3.5 km from any known nest site (average distance be- METHODS tween neighboring Spanish Imperial Eagle nests). Because these eagles only nest in stone Study area. The study area included Do- pines, cork oaks, eucalyptus trees, and poplars nãna National Park (75,000 ha) and a 20-km (Populus spp.; Ferrer 1993), we rejected any site zone around the park in southwestern Spain that did not fall near these tree species. We only (37ЊN, 6Њ30ЈW). Three main habitats can be used random trees that were Ն10 m in height distinguished in the area: Mediterranean scrub- corresponding to the minimum height of trees land, marsh, and coastal sand dunes. Mediter- used by Spanish Imperial Eagles to support the ranean scrubland is characterized by Halimium nest (Calderoń et al. 1987). Table 1. Comparison of 14 occupied Spanish Imperial Eagle territories (N ϭ 73 nest sites) to 14 randomly selected unoccupied territories (N ϭ 73 random sites) in southwestern Spain, 1984–1994.

Occupied Unoccupied 2 Habitat variables x¯ SD x¯ SD ␹ 1 P Tree height (m) 18 4.6 13.9 1.9 6.5 0.011 Distance to the nearest (m): Marsh border 2136.3 2198.4 9334 3809.8 15.2 Ͻ0.001 Paved road 2811 1594 561.2 386.7 18.6 Ͻ0.001 Urban center 7820.7 2805.6 5390.6 2342.8 4.9 0.027 Agricultural land 8499.8 6361.9 1507 2078.7 11.9 Ͻ0.001

We measured five habitat variables (Table 1) hood ␹2 statistics, classification accuracy (based for each nest and random site or tree, in ad- on a logistic cutpoint of 0.5 to classify sites as dition to the two categorical variables of tree an occupied or unoccupied territories), and the species and type of tree stand (STAND1 ϭ iso- Hosmer-Lemeshow Lack-of-Fit test. lated tree; STAND2 ϭ group of Յ10 trees; STAND3 ϭ row of trees; STAND4 ϭ small RESULTS tree stand [between 10 trees and 5 ha]; STAND5 ϭ large tree stand [Ͼ 5 ha]). At the Our univariate analyses showed that territo- end of the breeding season (July–October), we ries occupied by Spanish Imperial Eagles dif- visited each nest and random site to measure fered significantly from unoccupied territories tree height (microhabitat). Macrohabitat vari- for all of the habitat variables measured (Table ables representing vegetative composition, de- 1). Tree species and type of tree stand used at gree of human influence, and land use were nesting sites also differed significantly from ran- sampled within a 1.75-km radius (average dis- dom sites (␹2 ϭ 18.7, P Ͻ 0.01 and ␹2 ϭ 32.9, tance between neighboring territory boundar- P Ͻ 0.001, respectively; N ϭ 146). Nest sites ies) around the nest or random tree (total area were found more frequently in eucalyptus of 9.62 km2) using aerial photographs (Junta de (33.3%), cork oak (13.3%), and poplar (5.3%) Andalucia) and land use maps from 1984, and in smaller tree stands (STAND1 [18.7%], 1985, 1991, 1992, and 1994. STAND2 [10.7%], and STAND4 [20.0%]). Statistical analyses. To avoid pseudore- The logistic regression model that best dis- plication (because eagles are known to occupy criminated between occupied and unoccupied the same territory year after year), we used the territories included height of the nest or ran- mean of each habitat variable (excluding tree dom tree and distance from the nest or random species and type of tree stand) for each of the tree to nearest urban center (Table 2). The 14 occupied (N ϭ 73) and 14 unoccupied ter- model correctly classified 78.6% of the occu- ritories (N ϭ 73; Table 1). Using the means for pied and 85.7% of the unoccupied territories. each habitat variable, we first conducted uni- A Hosmer-Lemeshow Lack-of-Fit test (␹2 ϭ variate comparisons between occupied and un- 22.01, P ϭ 0.635) indicated a good fit of the occupied territories using Kruskal-Wallis tests data to the logistic model. for tree height and distance to marsh, paved road, agricultural land, and urban center. Tree DISCUSSION species and type of tree stand comparisons were conducted using chi-square tests. We then used The univariate analyses showed that Spanish logistic regression modeling to determine which Imperial Eagle nest sites differed significantly habitat variable best discriminated between oc- from random sites at both the microhabitat and cupied and unoccupied territories. A logistic re- macrohabitat levels for all of the variables mea- gression analysis investigates relations between sured. Eagles nested in taller trees closer to the binary-response probabilities (occupied and un- marsh border and farther from paved roads, ur- occupied) and explanatory variables. The over- ban centers, and agricultural lands when com- all model significance was based on log-likeli- pared to unoccupied sites. These results dem- Table 2. The logistic regression model that best discriminated territories occupied by Spanish Imperial Eagles from unoccupied territories in southwestern Spain, 1984–1994.

Habitat variable Parameter estimate SE P Constant Ϫ11.708 4.688 0.0125 Tree height 0.510 0.248 0.040 Distance to the nearest urban center 5.590 ϫ 10Ϫ4 2.608 ϫ 10Ϫ4 0.0321 2 2 Overall log-likelihood ␹ 2 ϭ 16.806, P ϭ 0.002. Hosmer-Lemeshow Lack-of-Fit test ␹ 2 ϭ 22.010, P ϭ 0.6351. onstrate a clear difference between occupied consequently abandoned by her mate (pers. and unoccupied sites and suggest a lack of avail- obs.). Our univariate analyses showed that oc- able suitable nest sites for Spanish Imperial Ea- cupied sites were significantly farther from ag- gles around Donãna National Park. ricultural lands that may also be indirectly as- Our logistic regression model included two sociated with human persecution. Landowners variables representing microhabitat (height of in the area believe that raptors are a threat to nest tree) and human influence (distance to the their land and will shoot eagles and other birds nearest urban center; Table 2). Spanish Imperial of prey on sight. Furthermore, power lines, Eagles chose tall trees that were either isolated which are mainly found near urban centers, or in small tree stands. These microhabitat may represent an indirect threat because elec- characteristics represent an accessible nest site. trocution by power lines remains the primary The importance of an accessible nest site has cause of death in adult Spanish Imperial Eagles, been documented for other species of raptors particularly in Donãna National Park (Janss (e.g., White-tailed Sea Eagles [Haliaeetus albi- and Ferrer 1999). Gonza´lez et al. (1992) cilla, Shiraki 1994] and Red-tailed Hawks [Bu- showed that eagles nested far from villages and teo jamaicensis, Bednarz and Dinsmore 1981]. other human-related structures (paved roads As suggested by Sva¨rdson (1949) and Hildeń and inhabited buildings) throughout the rest of (1965), habitat use is influenced by the mor- the species’ range. phological, physiological, and behavioral adap- Although distance from nest or random tree tations of a given species, such as hunting tech- to the nearest marsh border was not in our lo- niques in the case of raptors (Andrew and gistic regression model, our univariate analyses Mosher 1982; Janes 1985). Spanish Imperial showed that Spanish Imperial Eagles sites were Eagles are large aerial hunters and may avoid closer to the marsh border than unoccupied densely wooded areas for easy maneuverability sites (Table 1). Marshland is the most impor- in order to optimize foraging activities. Tall tant distinguishing factor between the eagle’s trees may be limited in the area, and the lack habitat in Donãna National Park and its habitat of tall nesting structures for Spanish Imperial elsewhere. The marsh in Donãna National Park Eagles are potentially restricting the expansion is bordered by Mediterranean scrubland (Rog- of the population. ers and Myers 1980), which alone represents The second variable in our logistic regression optimal habitat for the eagle’s main prey, the model was distance to urban center. The model rabbit (Oryctolagus cuniculus; Ferrer 1993). But showed that site occupancy by Spanish Imperial most importantly, the marsh/Mediterranean Eagles increases with distance to urban center. scrubland ecotone contains a wider variety of Urban centers represent the highest intensity of prey for Spanish Imperial Eagles, including human activity around Donãna National Park aquatic birds such as the Greylag Goose (Anser and hence the highest danger of persecution for anser) and the Eurasian Coot (Fulica atra). Ea- Spanish Imperial Eagles. Persecution by hu- gle pairs that nest closer to the marsh tend to mans still represents an important mortality have a more diversified diet and greater avail- factor for this species in and outside park ability of prey compared to those that nest far boundaries (Ferrer 1993). In 1994, a female from the marsh, which are mainly restricted to Spanish Imperial Eagle was shot at her nest near rabbits (Delibes 1978). Therefore, Spanish Im- one of the bordering towns and the nest was perial Eagles possibly nest near this ecotone because it represents an area with higher prey den- los pollos en el Coto de Donãna. Donãna Acta Ver- sity. In contrast, Gonzalez et al. (1992) did not tebrata 5: 35–60. FERRER, M. 1993. El A´guila Imperial. Quercus, Madrid. show any significant differences between occu- ———, and J. CALDERO´ N. 1990. The Spanish Imperial pied and unoccupied sites in terms of vegeta- Eagle Aquila adalberti C. L. Brehm 1861 in Do- tion structure. nãna National park (south west Spain): a study of Management implications. Our study population dynamics. Biological Conservation 51: 151–161. suggests that both human persecution or dis- ———, and F. HIRALDO. 1991. Evaluation of manage- turbance and lack of adequate nesting habitat ment techniques for the Spanish Imperial Eagle. (specifically tall trees in small tree stands) ap- Wildlife Society Bulletin 19: 436–442. pear to be important factors inhibiting the ex- GARZO´ N, J. 1974. Contribucioń al estudio del status, alimentacioń y proteccioń de las Falconiformes en pansion of Spanish Imperial Eagles. Since the Espanã Central. Ardeola 19: 270–330. end of this study in 1994, the eagles’ critical GONZA´LEZ, L. M., J. BUSTAMANTE, AND F. H IRALDO. situation in Donãna National Park has not im- 1992. Nesting habitat selection by the Spanish Im- proved, and few pairs bred in 2001 (J. Da´vila, perial Eagle Aquila adalberti. Biological Conserva- pers. comm.). Considering that the Donãna tion 59: 45–50. ———, F. HIRALDO,M.DELIBES, AND J. CALDERO´ N. National Park eagle population is the densest 1989. Reduction in the range of Spanish Imperial and most isolated and that the area outside park Eagle (Aquila adalberti) since 1850. Journal of Bio- boundaries is under constant threat of expan- geography 16: 305–315. GREENWAY, J. C. 1967. Extinct and vanishing birds of sion, there is an urgent need to create and en- the world. Dover, New York. force management policies for this area that HILDEŃ, O. 1965. Habitat selection in birds, a review. would provide a buffer for eagles nesting in the Annales Zoologici Fennici 2: 53–75. area and an area into which the species could JANES, W. J. 1985. Habitat selection in raptorial birds. In: Habitat selection in birds (M. L. Cody, ed.), expand in the future. We suggest that these pol- pp. 159–188. Academic Press, New York. icies be geared towards managing habitat to in- JANSS,G.F.E.,AND M. FERRER, 1999. Avian electro- clude tall trees in small tree stands safe from cution on power poles: European experiences. Birds human interference. and power lines (G. F. E. Janss, and M. Ferrer, eds.), pp. 145–164. Quercus, Madrid. ACKNOWLEDGMENTS MARTIN, T. E. 1993. Nest predation among vegetation layers and habitat types: revising the dogmas. Amer- Funds were provided by the Spanish National Insti- ican Naturalist 141: 897–913. tute for Nature Conservation (ICONA) and a North MEYBURG, B.-U. 1975. On the biology of the Spanish Atlantic Treaty Organization (NATO) research grant. Imperial Eagle (Aquila heliaca adalberti). Ardeola We are indebted to J.-J. Negro, F. Hiraldo, M.-L. Cha- 21: 245–283. coń, R. Cadenas, L. Garc´ı a, Antonio Jesus, and M. ———. 1989. The Spanish Imperial Eagle Aquila (he- Blanco for assistance with data collection. The manu- liaca) adalberti: its biology, status and conservation. script was improved by comments of D. R. Norris, B. In: Raptors of the modern world. Proceedings of J. M. Stutchbury, G. Janss, R. D. Titman, G. Bortolotti, the third world conference on birds of prey and and T. S. Jung. owls. Eilat, Israel, 22–27 March 1987 (B.-U. Mey- burg, and R. D. Chancellor, eds.), pp. 255–267. International Council for Bird Preservation (ICBP) LITERATURE CITED and International Union for the Conservation of Nature (IUCN)/Species Survival Commission ANDREW,J.M.,AND J. A. MOSHER. 1982. Bald Eagle (SSC). nest site selection and nesting habitat in Maryland. NEWTON, I. 1979. Population ecology of raptors. Journal of Wildlife Management 46: 383–390. T&AD Poyser, Berkhamsted, UK. BEDNARZ,J.C.,AND J. J. DINSMORE. 1981. Nest-sites ROBBINS, C. S. 1979. Effect of forest fragmentation on and habitat of Red-shouldered and Red-tailed bird populations. In: Workshop proceedings: man- hawks in Iowa. Wilson Bulletin 94: 31–45. agement of northcentral and northeastern forests BIJLEVELD, M. 1974. Birds of prey in Europe. Macmil- for nongame birds (R. M. DeGraaf, and K. E. lan, London. Evans, eds.), pp. 198–212. United States Depart- CALDERO´ N, J., J. CASTROVIEJO,L.GARCIÁ, AND M. FE- ment of Agriculture Technical Report 51. RRER. 1987. El Aguila´ Imperial (Aquila adalberti) ROGERS,P.M.,AND K. MYERS. 1980. Animal distribu- en Donãna: algunos aspectos de su reproduccioń. tions, landscape classification and wildlife manage- Alytes 5: 47–72. ment, Coto Donãna, Spain. Journal of Applied COLLAR,N.J.,AND P. A NDREWS. 1988. Birds to watch: Ecology 17: 545–565. the ICBP world check-list of threatened birds. In- SHIRAKI, S. 1994. Characteristics of White-tailed sea ea- ternational Council for Bird Preservation (ICBP) gle nest sites in Hokkaido, Japan. Condor 96: Technical Publications No. 8. Cambridge, UK. 1003–1008. DELIBES, M. 1978. Ecolog´ı a alimenticia del Agila Im- SVA¨RDSON, G. 1949. Competition and habitat selection perial Ibe´rica Aquila adalberti durante la crianza de in birds. Oikos 1: 157–174.