- Top of page
- Study area
1. The booted eagle Hieraaetus pennatus is a poorly known and scarce raptor that breeds in Spain. In Doñana National Park (south-west Spain) its population has increased from only six breeding pairs in the early 1980s to about 150 today.
2. In order to guide habitat management for this raptor in Doñana National Park, we related nesting habitat selection to breeding success.
3. Birds withstood some human disturbance when nesting, choosing sites closer to pastures besides marshes, footpaths and crops than would occur in a random distribution. Birds also selected areas near to marsh and stands of cork oak Quercus suber.
4. Trees used for nesting were wider and taller than would occur at random. They were usually in small groups or were large isolated trees, typically eucalyptus (Eucaliptus spp.).
5. The most productive nests were close to marshland and stone pine trees Pinus pinea.
6. Habitat management to improve the breeding success of booted eagles in Doñana should include: (i) retaining small groups of trees or large isolated trees, especially eucalyptus and cork oaks close to marshland, isolated buildings and crops; (ii) creating clearings in stone pine plantations; (iii) burying potentially dangerous power lines to reduce collision risks; (iv) clearing some areas of scrubland to increase the rabbit population; and (v) controlling forest activities, especially in the breeding season.
7. The increase in booted eagle populations in western Europe during recent decades may be a consequence of the species’ capacity to adapt to environmental change. Deforestation policies designed to favour agricultural use implemented during the second half of the 20th century have not had a detrimental effect on this raptor.
8. Our work demonstrates how scarce and important organisms can be favoured by sensitive management in forestry and agricultural habitats.
- Top of page
- Study area
Booted eagles arrive in Spain at the beginning of March and leave for Africa late in September. They nest in all parts of the country except the Canary Islands, but numbers vary in the different regions (De Juana 1989). The population is estimated at 2000–4000 breeding pairs in Spain, and 2800–6100 in the whole of Europe (Purroy 1997). In Doñana National Park, the population of this raptor has increased from only six breeding pairs in the early 1980s to about 150 today (L. García, unpublished data).
Booted eagles occupy territories mainly in four different biotopes within Doñana National Park, where the dominant tree species are eucalyptus Eucaliptus spp., stone pine Pinus pinea L. and cork oak Quercus suber L., which they use for nesting.
Doñana is a protected area where directed habitat management is used to enhance populations of endangered species (e.g. the Spanish imperial eagle Aquila adalberti G.L. Brehm). Thus traditional management techniques, such as burning or clearing of scrubland, have been used to increase rabbit populations. Forest management consists of re-afforestation with cork oak in some parts of the reserve. It is also planned to remove the eucalyptus trees, which were introduced into the area in 1969 and now occupy about 5000 ha in the northern part of the Park.
Habitat characterization has been applied frequently in ecology and has been useful in deriving conservation measures (Newton, Davis & Moss 1981; Ferrer & Harte 1997). The aim of this study was to build models to predict suitable breeding habitat for booted eagles in Doñana. We also tried to identify habitat differences between productive and less productive territories. Finally, we offer recommendations designed to increase the chances of this species occupying new territories and to avoid any loss of nesting habitat in areas where eucalyptus will be felled.
- Top of page
- Study area
Doñana National Park is located in south-western Spain (37°N, 6°30′W) and its area is about 50 000 ha. The climate is Mediterranean with Atlantic influences. Marshes, Mediterranean scrubland mixed with scattered cork oaks or stone pines and coastal sand dunes are the main habitats. Other habitats include streams with riparian vegetation, woodlots of small stone pines planted 30 years ago, and eucalyptus. A more detailed description is presented in Rogers & Myers (1980).
The nests that are studied in this paper were located in seven different parts of the National Park (Fig. 1). (1) Reserva Biológica de Doñana (RBD), La Algaida, Las Mogeas and El Acebuche are situated in the centre of the National Park. Here, booted eagle nests occur in stone pines or cork oaks that are scattered within scrubland (matorral) comprising mainly Halimium halimifolium L., Cistus libanotis L. and Erica scoparia L. Large parts of these areas are also occupied by forests of small stone pines. (2) La Dehesa is situated on the north side and consists of cork oak scattered with scrub of Pistacea lentiscus L. Here, most nests are located in smaller cork oaks than those found in RBD. (3) Pinar del Vicioso is located in the northernmost part of the area and comprises mature woodland of Pinus pinea L. (4) Los Sotos is a plantation of eucalyptus located between RBD and La Dehesa. (5) La Rocina is a stream with riparian vegetation (poplars Populus spp., Fraxinus angustifolia L.) located in the west of the Park. (6) La Pequeña Holanda is located in the west of the Park and its main habitats are matorral with small groups of eucalyptus trees. (7) Crops are also grown extensively in the area, comprising mainly rice and other irrigated and non-irrigated crops.
Figure 1. Map of study area showing the different areas analysed in the paper. (1) RBD, Las Mogeas and La Algaida; (2) La Dehesa; (3) Pinar del Vicioso; (4) Los Sotos; (5) La Rocina; (6) La Pequeña Holanda; and (7) crops.
Download figure to PowerPoint
- Top of page
- Study area
A total of 84 nest sites used by booted eagles during 1994 and 1996 was used in this study: 50 nests from 1994 and 34 nests from 1996. There were only four cases of nests so closely set in pairs (eight nests) that pseudoreplication problems could arise, whereby each pair of nests belonged to one or two pairs of birds.
Nests were marked on aerial photographs (Andalusian Cartographic Institute, scale 1 : 20 000, years 1991–92) and on maps (1 : 50 000 topographic map of Spain, IGN; 1 : 50 000 farming and land-use map, MAPA; 1 : 100 000 ecological impact map of Doñana, Castroviejo). An equal number of random points was marked on the photos and maps. As the booted eagle is a forest species, open areas lacking potential nest sites (such as marsh, crops and the buildings) were excluded from the random points, as were wooded locations where tree heights and diameter at breast height (d.b.h.) were less than the minimum value used for nesting (d.b.h. = 68 cm, height = 430 cm) (Howell et al. 1978; González, Bustamante & Hiraldo 1992).
For each nest site and random point we measured 41 variables in order to quantify the habitat (Table 1). Thirty-three macrovariables were measured, such as distance from the site to selected habitat features or the percentage cover of habitat within a radius of 530 m, which was half the mean distance between nests (following the methodology of Bednarz & Dinsmore 1981; Gilmer & Stewart 1984). The percentage cover of vegetation types was measured using sygmascan pro 4.0 image analysis software (Fox & Ulrich 1995). In addition, eight microvariables were measured in the field (Table 1). Tree height was measured using an optical height meter and orientation measurements were made using a compass.
Table 1. Variables used to characterize nest sites of the booted eagle compared with random sites
|DELPOW||Distance (km) to nearest electric power lines|
|KELPOW||Km of electric power lines in circular sampling area|
|DBUILD||Distance (km) to nearest isolated building|
|DURBAN||Distance (km) to nearest urban centre|
|DPAVRO||Distance (km) to nearest paved road|
|KPAVRO||Km of paved roads in circular sampling area|
|DASPHRO||Distance (km) to nearest asphalt road|
|KASPHRO||Km of asphalt roads in circular sampling area|
|DPILGRIM||Distance (km) to nearest pilgrim trail (camino rociero)|
|KPILGRIM||Km of pilgrim trail (camino rociero) in circular sampling area|
|DFIRBRE||Distance (km) to nearest fire break|
|KFIRBRE||Km of fire breaks in circular sampling area|
|DUNMADRO||Distance (km) to nearest unmade road|
|KUNMADRO||Km of unmade roads (non-paved roads and tracks) in circular sampling area|
|DNIRCRO||Distance (km) to nearest non-irrigated crop|
|DIRCRO||Distance (km) to nearest irrigated crop|
|DRICRO||Distance (km) to nearest rice crop|
|DPASTU||Distance (km) to nearest pasture|
|DSCRUB||Distance (km) to nearest scrubland|
|DMARS||Distance (km) to the border of marsh|
|DWATER||Distance (km) to nearest open water|
|MARSH||% surface covered by marsh in the circular sampling area|
|EUCAL||% eucalyptus (Eucaliptus spp.)|
|PINE||% stone pines (Pinus pinea)|
|POPUL||% poplars (Populus spp.)|
|NIRRCRO||% non-irrigated crops|
|IRRCRO||% irrigated crops|
|OTHER||% other open lands (e.g. abandoned crops)|
|OAK||% cork oaks (Quercus suber)|
|SAND||% sand dunes|
|HEIGHT||Height of tree (m)|
|NESHEIG||Height of nest in the tree (m)|
|DBH||Nest tree diameter (d.b.h.) (cm)|
|AZIMUT||Nest position in tree canopy (degrees)|
|OPEN||Nearest open land direction (degrees)|
|GROUP||Tree group size category (1, isolated tree; 2, < 10 trees; 3, row of trees; 4, small wood < 5 ha; 5, large wood > 5 ha)|
|APERT||Nest wood aperture angle (0° closed wood, 360° isolated wood)|
Nest sites from the 2 years were considered as independent samples, although some bias was possible due to different nests being used by the same pair. However, the use of nests from non-consecutive years (1994 and 1996) located in different places from the Park reduced this bias.
The macrovariables were checked for statistical normality using Lilliefors test and variables were square root-transformed (distances) and arcsine-transformed (percentage circle area) as appropriate. Mean values for nest site and random site variables were compared using t-tests with a Bonferroni correction to reduce the chance of type I errors.
We used logistic regression, through a generalized linear model (GLM) procedure, to identify the set of variables that best separated nest sites from random sites (Jongman, ter Braak & Van Tongeren 1995). Using a forward stepwise procedure, each variable was tested for significance in turn, and the variable contributing to the largest significant change in deviance from the null model was then selected and fitted. At each step the significance of the variables included in the model was tested and any falling below the criterion level of P = 0·05 was excluded. The final model was considered to have been identified when all the variables had a significant effect at P < 0·05.
For GLM, the data were used without transformations for normality as this is not a requirement of logistic regression. Independent variables were analysed in three groups (macrovariables within the 530-m circle, those outside, and the microvariables). Model fit was assessed by examining the coefficient of sensitivity, residuals (deviance and Pearson chi-square) and potential leverage (Nicholls 1989). Use of GLM with a logistic link function was considered more appropriate than the alternative of linear discriminant function analysis because the distributions of values were highly skewed (Green, Osborne & Sears 1994). We constructed similar models for the binary response variable nest productivity (two or more fledged chicks vs. one or fewer). All data were analysed using systat and spss (SPSS Inc., Chicago, IL, USA).