Table 3. Minimal adequate GLM using predictors related to land-use and human disturbance measured in circular sampling plots to predict the presence and absence of suitable breeding territories of Black-shouldered Kite in southwestern Spain.
positive relationship between the percentage of irrigated crops and the probability of finding a breeding pair, and above it the opposite trend occurs (Fig. 2b). The extent of scattered trees at medium densities (dehecereal L2) also correlated directly with the
Term Coefficient se
probability of classifying a given area as suitable for
breeding (Fig. 2c). The environmental model was
Null 126.1418 90
not affected by spatial autocorrelation; neither the
spatial coordinates (P > 0.05) nor the autocov
variable (P > 0.05) showed a significant increment of the residual deviance after removing them from our minimal adequate model.
After applying the cross-validation leave-one-out re-sampling technique, our final model showed a
residual deviance. Three of the environmental variables were retained by the model: the percentage of cultivated dehesa with low (dehecereal L1) or medium (dehecereal L2) densities of trees entered the model as main effects, and the percentage of irrigated crops entered the model in its quadratic form (Table 3). As univariate statistics suggested, dehecereal L1 had the greater effect on the response variable, explaining 44.6% of the residual deviance: the higher the area in the circular plots occupied by this type of habitat, the higher the probability of finding a breeding pair of Black-shouldered Kites. However, the probability of finding a breeding pair is 100% at about 40% of the circular sampling area occupied by this type of habitat (Fig. 2a). The Black- shouldered Kites showed a preference for a particular range of the area occupied by irrigated crops, with an optimum at around 50% of the circular sampling area occupied by this type of habitat. The response curve showed that below this threshold there is a
Cohen’s Kappa value (mean ± se) of 0.62 ± 0.08, suggesting a robust performance. When taking the cut-point at 0.4 (Fig. 3), the model correctly reclassified
80.4% of breeding sites and 82.2% of the random plots (overall = 81.3%). The model was also stable, showing 85.7% of coincidences when comparing the predictions done by the original model with those obtained using the cross-validation re-sampling technique. When using the data-splitting strategy with 40 random iterations, this gave an AUC output (mean ± se) of 0.79 ± 0.08, indicating that our model was sufficiently robust for predicting the presence or absence of breeding pairs outside our study area.
Changes in farming practices during the 20th century have influenced many bird populations in Europe (O’Connor & Shrubb 1986, Pain & Pienkowski 1997, Pain et al. 1997, Berg & Gustafson
Figure 3. The proportion of cases (random and nest-sites) classified correctly in the occurrence model is represented on the y-axis with the selected cut points varying between 0.0 and
1.0 on the x-axis. The cut-point that classified equally well the presence and absence of breeding pairs within our sampled circular sampling areas was optimum at a predicted cut point of
Figure 2. Response curves for three significant explanatory habitat variables entering the minimal adequate model. The curves show the effect of each variable (x-axis) on the probability of finding a breeding site in a given sampling area (y-axis): (a) dehecereal L1, (b) irrigated land and, (c) dehecereal L2.
2007) and this may be responsible in part for the fact that 81% of the bird species breeding in farmlands are classified as of European Conservation Concern and that 76% have shown recent population declines (Suárez et al. 1997). How these changes in land-use have affected individual species is still poorly understood. The case of the Black-shouldered Kite may represent a reversal of these trends, as it seems to be a natural and recent addition to Europe’s biodiversity in agricultural landscapes.
The recent transformation of densely forested traditional dehesas for livestock grazing into more open cultivated dehesas may have benefited Black-shouldered Kite populations in the Iberian Peninsula. We have demonstrated that this raptor shows a strong preference for cultivated dehesas for breeding, and this dehesa type has only become widely available in the last 50 years (Costa et al.
2005). The Elanus kites throughout the world are rodent specialists and their respective distribution patterns, population densities, movements and breeding performance are largely dependent on rodent abundance and availability (Mendelsohn & Jaksic
1989). In southwestern Spain the Black-shouldered Kites staple prey are mice of the genus Mus, specifically the Algerian Mouse Mus spretus that constitutes
44.5% of prey found in pellets during the breeding
season in Extremadura (J. J. Ferrero unpubl. data). Other small mammals, such as the Wood Mouse Apodemus sylvaticus, voles of the genus Microtus (M. duodecimcostatus, M. arvalis) and shrews (Crocidura spp.) have also been found among prey remains (Parejo et al. 2001, Mañosa et al. 2005). It has to be noted, however, that diversity and abundance of microtine rodents in Mediterranean areas of the Iberian Peninsula are among the lowest in Europe (Soriguer et al. 2003). Cultivated dehesas may be an exception to this general pattern, at least in terms of rodent abundance, as they harbour relatively large populations of Mus and Apodemus (authors’ unpubl. data).
Interestingly, our results showed that a range between 0% and 50% of the circular sampling area occupied by irrigated fields increases the probability of occurrence of a breeding pair. The reason may be that the Black-shouldered Kite frequently uses irrigated fields as temporary feeding habitats, both inside and outside the breeding season (Parejo et al. 2001).
Overall, the occurrence model explained a high percentage of the deviance and its predictive power was high when compared with other studies on raptors (Seoane et al. 2003 and references therein). The variance that remained unexplained could be due to overlooked factors, including for example interspecific interactions (Sergio et al. 2004), which might be important in this small-sized species sharing the same habitat with other birds of prey, including the Short-eared Owl Asio otus, the Com- mon Buzzard Buteo buteo, the Black Kite Milvus migrans, the Booted Eagle Hieraaetus pennatus and the Eurasian Kestrel Falco tinnunculus. Predation of both juvenile and adult Black-shouldered Kites by other raptors in our study area in the period 2003 –
07 was the highest observed mortality factor in this population (authors’ unpubl. data).
As for the applicability of our model to evaluate habitat availability at a broader scale, our bootstrap re-sampling procedures gave us confidence in the performance of the occurrence model and we believe it could provide a basis for suitable habitat conservation for the species outside our study area. However, the possibility of building predictive distribution maps is as yet limited, because raster cover considering cultivated dehesas as a land-use class has not been included in available maps and many databases. For instance, European CORINE land-use/land-cover digital maps cluster all types of dehesas into a single agro-forestry class.
The primary result of this study is that there is a
clear association between the presence of breeding pairs and cultivated dehesas. This kind of habitat, suitable for both breeding and foraging, has become increasingly available in the second half of the past century, in parallel to Black-shouldered Kites colon- ization. Other factors overlooked in the present study could also be responsible for colonization and subsequent expansion of this kite in southern Europe. Recent shifts in distribution ranges for many species have been related to climate change (e.g. Valiela & Bowen 2003, Crick 2004, Carrillo et al.
2007). Available climate data indicate an increase in mean global temperature in the past century (IPCC
2001) in southern Spain and northern Africa. It is possible that the increment in mean temperature has favoured the expansion of the Black-shouldered Kite in the Iberian Peninsula. However, the contribution of both mean temperatures and availability of preferred habitat are difficult to separate because they have increased concurrently.
Today, the traditional dehesa is protected by law (e.g. in Extremadura since 1986); tree clearing is illegal and tree planting is encouraged with subsidies. However, cultivated dehesas are unstable habitats because they constitute a transitional step between the traditional dehesa devised and managed for keeping livestock and an agricultural pseudo-steppe with no trees. Today and probably in the future, the transformation of cultivated dehesas into extensive cereal monocultures and the agricultural practices associated with them, such as the clearing of field margins and the use of pesticides, could negatively affect Black-shouldered Kite populations. This bird of prey needs trees or tall bushes to build its nests, to preen, to shelter from predators and to roost at night. If cultivated dehesas become scarce in the future a possible management practice to favour the presence of this raptor would be to plant trees on cereal field margins. Extensive cereal monocultures in Spain and Portugal may hold enough small mammals for the Kites, but only if vegetation on field margins is respected and pesticide use minimized. However, these habitats are not currently used by Black- shouldered Kites due to the absence of adequate substrates for nesting or perching. There is also a lack of hedges in most cereal-producing areas of Spain. This kind of vegetation should be promoted by government policies because it would bring benefits not only for the Black-shouldered Kite but also for other species including game and their predators. Furthermore, for an expanding species, if cereal-growing dehesas are regressed to their former condition of
grounds for livestock ranching, or new cultivations are introduced, the Black-shouldered Kite may face the risk of extinction in Europe, as this species seems to be highly selective in terms of breeding habitat.
We are grateful to J.M. Casas for searching and finding many nest-sites. We also thank J.M. López Caballero who supervised the Elanus research project in the Consejeria de Medio Ambiente of Extremadura. J.M. Sayago, M. Vázquez, M. De la Riva, O. González, C. Corbacho, J.C. Nuñez, F. Sánchez, J. Zalba and J. Hernández are thanked for their help with fieldwork. We would like to thank J. Dávalos and A. Anega, and the staff of the Extremadura Government (Junta de Extremadura, Consejería de Medio Ambiente) who gave access to the raster cartography used in this study. We also thank three anonymous reviewers, Simon Butler and Rauri Bowie for their constructive comments.
Albert, A. & Harris, E.K. 1987. Multivariate Interpretation of
Clinical Laboratory Data. New York: Marcel Dekker.
Alvarado, E. 1983. El Sector Forestal en Extremadura. Ecología y Economía. Institución Cultural El Brocense. Cáceres, Spain: Diputación Provincial de Cáceres.
Augustin, N.H., Mugglestone, M.A. & Buckland, S.T. 1996. An autologistic model for spatial distribution of wildlife. J. Appl. Ecol. 33: 339 – 347.
Balbontín, J. 2005. Identifying suitable habitat for dispersal in Bonelli’s eagle: an important issue in halting its decline in Europe. Biol. Conserv. 12: 74 – 83.
Beck, J.B. & Shultz, E.K. 1986. The use of relative operating characteristic ROC curve in test performance evaluation. Arch. Pathol. Lab. Med. 110: 13 – 20.
Berg, A. & Gustafson, T. 2007. Meadow management and occurrence of corncrake Crex crex. Agri. Ecosyst. Environ.
120: 139 –144.
Bergier, P. 1987. Les rapaces diurnes du Maroc. Statut, reparti- tion et Ecologie. Annales du CEEP 3.
BirdLife International/European Bird Census Council.
2000. European Bird Populations: Estimates and Trends. BirdLife Conservation Series No. 10. Cambridge: BirdLife International.
Brown, L. & Amadon, D. 1968. Eagles, Hawks and Falcons of the World. New York: McGraw-Hill.
Bustamante, J. & Seoane, J. 2004. Predicting the distribution of four species of raptors Aves: Accipitridae in southern Spain: statistical models work better than existing maps. J. Biogeogr.
31: 295 – 306.
Carbajo, F. & Ferrero, J.J. 1985. On the ecology and status of the Black-shouldered Kite (Elanus caeruleus) in Extremadura, SW Spain. In Newton, I. & Chancellor, R.D. (eds) World Con- ference on Birds of Prey: Conservation Studies on Raptors. Cambridge, MA: ICBP Tech. Publ. No. 5.
Carrete, M. & Donazar, J.A. 2005. Application of central-place foraging theory shows the importance of Mediterranean dehesas for the conservation of the cinereous vulture, Aegypius monachus. Biol. Conserv. 126: 582 – 590.
Carrillo, C., Barbosa, A., Valera, F., Barrientos, R. & Moreno, E.
2007. Northward expansion of a desert bird: effects of climate change. Ibis 149: 166 –169.
Cohen, J. 1960. A coefficient of agreement for nominal scales.
Educ. Psychol. Meas. 20: 37 – 46.
Costa, T.M., Morla, C. & Sáinz, H. 2005. Los Bosques Ibéricos.
Una Interpretación Geobotánica. Barcelona: Editorial Planeta.
Cramp, S. & Simmons, K.E.L. (eds) 1980. The Birds of the
Western Palearctic, Vol. 2. Oxford: Oxford University Press.
Crick, H.Q.P. 2004. The impact of climate change on birds.
Ibis 146: 48 – 56.
Del Hoyo, J., Elliot, A. & Sargatal, J. (eds) 1994. Handbook of the Birds of the World, Vol. 2. Barcelona: Lynx Edicions. Duchateau, S. & Delage, F. 2006. Évolution, paramètres
reproducteurs et facteurs limitants de l’Élanion blanc Elanus caeruleus dans le sud-ouest de la France. Alauda 74: 385 –
Elena, M., López, J.A., Casas, M. & Sánchez del Corral, A.
1987. El Carbón de Encina y la Dehesa. Madrid: INIA, Ministerio de Agricultura, Pesca y Alimentación.
England, M.D. 1963. Observations on the Black-winged Kite in
Portugal, with preliminary notes on its status. Br. Birds 56:
444 – 452.
Fernández, N., Delibes, M., Palomares, F. & Mladenoff, D.J.
2003. Identifying breeding habitat for the Iberian Lynx:
inferences from a fine-scale spatial analysis. Ecol. Appl. 13:
Ferrer, M. & Harte, M. 1997. Habitat selection by immature Spanish imperial eagles during the dispersal period. J. Appl. Ecol. 34: 1359 –1364.
Ferrero, J.J. & Onrubia, A. 1998. Expansión del área de cría y distribución actual del Elanio Común Elanus caeruleus en España. In Meyburg, B.-U., Chancellor, R.D. & Ferrero, J.J. (eds) Holarctic Birds of Prey: 159 –171. Berlin: WWGBP-Adenex.
Ferrero, J.J. & Onrubia, A. 2003. Elanio común Elanus caeruleus.
In Mar tí, R. & del Moral, J.C. (eds) Atlas de las Aves Repro- ductoras de España: 158 –159. Madrid: Dirección General de Conservación de la Naturaleza-Sociedad Española de Ornitología.
Ferrero, J.J., Grande, J.M. & Negro, J.J. 2003. Copulation behavior of a potentially double-brooded bird of prey, the Black- shouldered Kite Elanus caeruleus. J. Raptor Res. 37: 1– 7.
Fielding, A.H. & Bell, J.F. 1997. A review of methods for the assessment of prediction errors in conservation presence/ absence models. Environ. Conserv. 24: 38 – 49.
IPCC. 2001. Climate change 2001: the Scientific Basis. In Albritton, D.L. & Meira Filho, L.G. (eds) Third Assessment Report of Working Group I. Cambridge: Cambridge University Press.
Landis, J.R. & Koch, G.G. 1997. The measurement of observer agreement for categorical data. Biometrics 33: 159 –174. Lilford, L. 1865. Notes on the ornithology of Spain. Ibis (2) 1:
Manel, S., Williams, H.C. & Omerod, S.J. 2001. Evaluating presence–absence models in ecology: the need to account for prevalence. J. Appl. Ecol. 38: 921– 931.
Mañosa, S., Montés, G., Bota, G. & Bonfil, J. 2005. Black- shouldered Kite Elanus caeruleus diet in an area recent colonized in the north-east of the Iberian peninsula. Revista Catalana d’Ornitologia 21: 11–16.