Sylvopastoral Management and Conservation of the Middle Spotted

Forest Ecology and Management 242 (2007) 343–352 www.elsevier.com/locate/foreco

Sylvopastoral management and conservation of the middle spotted woodpecker at the south-western edge of its distribution range Hugo Robles a,*, Carlos Ciudad a, Rubeń Vera b, Pedro P. Olea c, Francisco J. Purroy a, Erik Matthysen d a Department of Animal Biology, University of Leoń, Faculty of Biological and Environmental Sciences, Campus de Vegazana s/n, E-24071 Leoń, Spain b Department of Agro-forestry, Unit of Zoology, University of Valladolid, Av. Madrid 44, Campus La Yutera, E-34004 Palencia, Spain c Department of Biology and Environmental Sciences, Faculty of Experimental Sciences, SEK University, Campus de Santa Cruz La Real, Cardenal Zuń˜iga 12, E-40003 Segovia, Spain d Laboratory of Animal Ecology, Department of Biology, University of Antwerp (UIA), Universteitsplein 1, B-2610 Antwerp, Belgium Received 8 May 2006; received in revised form 29 November 2006; accepted 21 January 2007

Abstract Anthropogenic changes in forest structure threats the persistence of numerous organisms. In Spain, Pyrenean oak (Quercus pyrenaica) forests are mainly subjected to three sylvopastoral management systems which determine variations in the forest structure: (i) ‘‘young forests’’ subjected to frequent disturbance by fires and clear-cuttings; (ii) ‘‘Pyrenean oak dehesas’’, subjected to a traditional sylvopastoral system exclusive to Spain which favours the development of both old oaks and pasture; (iii) largely undisturbed ‘‘mature forests’’. We examined the effects of these sylvopastoral management systems on habitat use and reproduction of the middle spotted woodpecker (Dendrocopos medius), a forest specialist bird, at the south-western edge of its distribution range. Habitat use was examined at multiple spatial scales. Within the territories, woodpeckers preferred large diameter oaks for foraging and decay oaks with holes for nesting. Within the forest tracts, no significant relationships between territory density and habitat structure in the forest tracts were found. At the landscape scale, dehesas and mature forests, both with high densities of important foraging and nesting trees, were occupied by woodpeckers. Young forests, with low densities of those trees, were avoided. No significant differences in territory density or reproductive success were found between dehesas and mature forests. The maintenance of viable woodpecker populations requires conservation actions at several spatial scales. Within the forest tracts, suitable densities of important foraging and nesting trees must be maintained. At the landscape scale, management decisions must consider the maintenance of traditional sylvopastoral systems in dehesas and mature forests. At the regional scale, taking into account a habitat cover in the landscape of the 7% for the Cantabrian Mountains, we estimated an area of 285,000 ha to be included in the conservation planning unit for the middle spotted woodpecker. # 2007 Elsevier B.V. All rights reserved.

Keywords: Habitat use; Multiple spatial scales; Reproductive success; Habitat alteration; Dehesas; Dendrocopos medius

1. Introduction in the structure and composition of the forest and, in consequence, in habitat structure of forest organisms. Habitat Most forests in Europe have been or are being subjected to alteration due to anthropogenic activity threats the persistence some type of anthropogenic activity (Bradshaw, 2004). Forest of numerous forest organisms (Ranius and Jansson, 2000; Luck, uses by humans varies from hunting and extensive traditional 2002; Erickson and West, 2003; Bu¨tler et al., 2004). An grazing to intensive wood production in plantations (Luis- essential step towards the development of effective conserva- Calabuig et al., 2000; FAO, 2001; Plieninger et al., 2003; tion strategies is to examine responses in the habitat use and Duriez et al., 2005). These different uses determine variations reproductive success of species to changes in anthropogenic forest use. An important tool for maintaining biodiversity is the use of keystone and umbrella species (Mikusinski, 1997; Simberloff, * Corresponding author. Present address: Department of Biology, Laboratory of Animal Ecology, University of Antwerp (UIA), Universteitsplein 1, B-2610 1999; Roberge and Angelstam, 2004). In the temperate Antwerp, Belgium. Tel.: +32 3820 2261; fax: +32 3820 2271. deciduous forests of the western Palaearctic, the presence of E-mail address: [email protected] (H. Robles). the non-migratory middle spotted woodpecker (Dendrocopos

0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2007.01.052 344 H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 medius) is associated to a high richness of other forest bird (Wiens et al., 1987; Luck, 2002); (ii) habitat structures (e.g. species (Roberge and Angelstam, 2006), acting as an umbrella foraging and nesting substrates) selected on a spatial scale may species (Angelstam et al., 2004). As an excavator of holes in be driving habitat use decisions at a larger spatial scale (Orians trees, the middle spotted woodpecker is considered a keystone and Wittenberger, 1991; Walters et al., 2002). First, we species in forests, providing breeding opportunities for compared the distribution and abundance of the middle spotted secondary hole-nesting birds (e.g. Wesołowski, 1989). Its woodpecker in forest tracts subjected to different sylvopastoral habitat requirements are associated with characteristics of the management system (landscape scale). Second, we analysed mature forests, such as large diameter trees for foraging foraging and nesting tree use within the territories (territory (Pasinelli and Hegelbach, 1997; Weib, 2003) and dead wood for scale), as the densities of these trees may influence selection excavating breeding cavities (Wesołowski and Tomialojc, processes on a larger spatial scale (see Pasinelli, 2000). Based 1986; Pasinelli, 2003; Kosin˜ski and Winiecki, 2004). The on the results of tree use, we examined the relations between alteration and destruction of its habitat has caused a reduction in habitat structure and woodpecker distribution and abundance in numerous populations of this woodpecker (Pasinelli, 2003). the forest tracts (forest tract scale). Last, we compared Examples of such habitat destruction are a reduction in the tree reproductive success from territories with different habitat ages, the elimination of old oaks and of dead and decaying trees structure and sylvopastoral management. Based on the obtained (Pasinelli, 2003). Middle spotted woodpecker is therefore results, management measures for the conservation of the potentially sensitive to changes in forest structure, the effects of middle spotted woodpecker were proposed, with implications which may be representative for other forest organisms for other forest bird species. (Roberge and Angelstam, 2006). Several studies have been carried out on habitat use of the 2. Study area middle spotted woodpecker in central and northern Europe (e.g. Pettersson, 1984; Pasinelli and Hegelbach, 1997; Fauvel, 2001; The study was conducted in a landscape dominated by Michalek et al., 2001; Hertel, 2003; Weib, 2003; Kosin˜ski and Pyrenean oak forests (Q. pyrenaica) of approximately Winiecki, 2004), but there is little information on southern 24,800 ha situated on the southern slope of the Cantabrian populations. Spanish population of middle spotted woodpecker Mountains (NW Spain), between Leoń and Palencia provinces represents the south-western edge of its distribution range (centred on 428430N, 5810W). It is located in the transition area (Pasinelli, 2003). In Spain, they mainly occupy Pyrenean oak between the Eurosiberian and Mediterranean bioclimatic Quercus pyrenaica forests in the Cantabrian Mountains and the regions (Penas, 1995), at altitudes from 900 to 1 350 m a.s.l. Basque Mountains (Purroy et al., 1984; Arrambari and Approximately the 85% of the study area is formed by Rodrı´guez, 1996; Garcıá-Fernańdez et al., 2002; Robles and Pyrenean oak forest (i.e. the potential habitat for the middle Olea, 2003; Onrubia et al., 2004). Pyrenean oak forests are spotted woodpecker in the Cantabrian Mountains, see Onrubia endemic of the western Mediterranean area (Ta´rrega and Luis, et al., 2004). A small proportion of other Quercus species (Q. 1990; Calvo et al., 1991) and have high bird species richness faginea, Q. petraea and Q. paucirradiata, see Penas, 1995; (A´ lvarez, 1989). Although the Iberian Peninsula has the largest Penas et al., 1997) may also be mixed together with the Pyrenean oak forest area in its distribution range, it is the least Pyrenean oak, and less frequently (<1%) other non-oak species protected forest in Spain, with only 2.21% of its surface (e.g. Fagus sylvatica, Pinus sp., Populus sp., Salix sp.). Other protected (Maldonado et al., 2001). Moreover, it has been forest habitats (2%) are pine plantations (Pinus pinaster and P. historically altered by human activities such as charring, sylvestris), beech (F. sylvatica), riverside (Populus sp., Salix sp. grazing, fires and clear-cuttings (Calvo et al., 1991; Luis- Fraxinus sp.) and holm oak (Q. rotundifolia) forests. The Calabuig et al., 2000; Maldonado et al., 2001). Three types of remaining 13% is formed by non-forest habitats, such as sylvopastoral management determine variations in Pyrenean pastures, scrublands, cereal croplands, roads and urban areas. oak forest structure: (i) heavily disturbed ‘‘young forests’’ Three different sylvopastoral management systems are used subjected to frequent fires and clear-cuttings; (ii) ‘‘Pyrenean in Pyrenean oak forests (Table 1). Firstly, ‘‘young forests’’ oak dehesas’’, subjected to a traditional sylvopastoral system maintained in that seral stage by frequent disturbances. They exclusive to central and northern Spain which favours the have frequently been burnt and/or subjected to repeated clear- development of old trees and pasture interspersed (Luis- cuttings within the last decades, so the oaks are mostly young Calabuig et al., 1993; Blanco et al., 1997; Camprodoń, 2001); trees regenerated from shoot on the rhizome or the stem of (iii) ‘‘mature forests’’ with scarce human activities. Data are subterranean roots (Ta´rrega and Luis, 1990; Luis-Calabuig necessary to assess the potential impact of these management et al., 2000). They have high shrub cover (mainly Erica sp., systems on the conservation of the middle spotted woodpecker. Genistella tridentata and Halimium sp.) and low livestock During a 5-year study period (2000–2004), we examined the pressure. Secondly, the ‘‘dehesas’’ are associated with effects of sylvopastoral management in Pyrenean oak forests on traditional extensive grazing, mainly sheep and goats. They habitat use and reproductive success of the middle spotted have low shrub cover and are often cleared by selective cutting woodpecker in the south-western edge of its distribution area to create pastures and to favour the development of large oaks (Cantabrian Mountains in NW Spain). Habitat use was (Dıéz et al., 1991; Luis-Calabuig et al., 1993; Blanco et al., examined at multiple spatial scales for two reasons: (i) 1997). These oaks are pruned to make their crowns larger, different factors may influence habitat use at different scales providing shade in summer and abundant acorn production for H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 345

Table 1 Characteristics of sylvopastoral management systems in Pyrenean oak forests Characteristics Young forests Dehesas Mature forests Livestock pressure Low Medium. Extensive grazing, Low mainly by sheep and goats Cut type Clear-cutting Selective cutting Little selective cutting for firewood extraction Other activities Burning Pruning Tree ages Young Old or bimodal (old and A wide variety of tree ages young trees) in abandoned dehesas. Little intermediate stratum Undergrowth cover High Low (high in abandoned dehesas) Variable cover with high species richness livestock. The low used or abandoned dehesas may have greater (Robles and Olea, 2003). Territory density in each forest tract shrub cover (Dıéz et al., 1991) and a bimodal age structure, i.e. was estimated from the mean number of territories observed with both large old trees and small young ones originated from during the study period. resprouting. Finally, the largely undisturbed ‘‘mature forests’’ maintain a wide variety of tree ages, a high richness of 3.2. Foraging and nesting tree use undergrowth species (Prunus spinosa, Ilex aquifolium, Sorbus aria, S. aucuparia, Crataegus monogyna, Rosa sp., etc.), low In the next step, we analysed foraging and nesting tree use livestock pressure and little firewood extraction by selective within the territories. The densities of these trees interact to cutting. These three sylvopastoral management systems make influence home range sizes in the middle spotted woodpecker up a mosaic landscape with patches subjected to different (Pasinelli, 2000), and may influence distribution and abundance degrees of antropogenic forest use and forest maturity. at a forest tract scale.

3. Methods 3.2.1. Trees used for foraging We collected data on foraging tree use based on visits to 3.1. Sylvopastoral management and woodpecker middle spotted woodpecker territories in March and April distribution and abundance 2002–2004, coinciding with the peak of territorial activity during the pre-breeding period (Pasinelli et al., 2001). We From 2000 to 2004, the distribution and abundance of the recorded the species, diameter (dbh, measured at 1.3 m from woodpeckers were determined in 49 randomly chosen forest the ground), condition (alive or dead) and height of each tree tracts (n = 17 young forest tracts, 9 dehesa tracts and 23 mature used, and the condition of the foraging substrate (alive or dead). forest tracts) of 21–40 ha, as most of dehesa patches had this In order to reduce dependence between observations only one size range. Forest tract size did not differ significantly between observation per tree and those distanced at least 10 min were the three-sylvopastoral systems (Kruskal–Wallis test: considered (Rudolph et al., 2002). We collected information in x2 = 1.667, d.f. = 2, p = 0.435). Woodpecker distribution 10 breeding territories, 5 in dehesas and 5 in mature forests, (presence/absence) was determined in each tract during three with a total of 160 trees (10–24 trees per territory). No data visits suitably distributed between March and the beginning of were collected in young forests as they were not occupied by May each year, when intra-specific aggressive interactions (i.e. woodpeckers (see Section 4.1). territorial activity) are most common (Pasinelli et al., 2001). Foraging tree availability was determined by establishing a During each visit the entire forest tract was covered network of circular plots of 0.04 ha (11.2 m radius) 100 m apart approximately at 20 ha/h. We stopped every 100 m to use a in each territory (Fig. 1). The species, diameter and condition of playback technique, alternating 30 s of species calls (kweek and all the trees were recorded in each plot. Only oaks measuring rattle calls, see Pasinelli, 2003) with 45 s listening, and 13 cm dbh were classed as available, as other species and repeating one more time this operation when the species was smaller oaks were not used by the woodpeckers. We also not detected (Robles and Olea, 2003). Given that not all the measured the mean height of 1–3 randomly-chosen trees forest tracts were sampled during the 5-year study period, we 20 cm dbh in each circle (96% of the trees used had 20 cm considered the species as being present when we detected birds dbh). for at least 2 years, and absent when no birds were detected in none of at least 4 years. 3.2.2. Trees used for nesting Forest tracts with woodpecker presence were visited five From May to July 2000–2004, we carried out intensive times per year to estimate woodpecker abundance. Once the searches to locate nests in 26 territories. We found 49 trees with presence of woodpeckers was detected we stopped using the breeding cavities. We defined breeding cavities as those playback technique, trying to avoid attracting birds from nearby excavated during the breeding season by territorial pairs if we territories (Kossenko and Kaygorodova, 2001). Woodpecker observed adults inside for longer than 20 min (i.e. incubating) were followed and signs of activity (visual and auditive) and/or feeding the nestlings. As multiple nesting trees used recorded on maps elaborated ad hoc to a 1:4000–1:10,000 scale within the same territories across years might be dependents, 346 H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352

Table 2 Description of the habitat variables Variables Description Density of small oaks Number of oaks <10 cm dbh and >1.5 m height per ha Density of middle oaks Number of oaks from 13 to <37 cm dbh per ha Density of large oaks Number of oaks 37 cm dbh per ha Density of potential cavity trees Number of oaks 22 cm dbh dead (with no living parts) or decay (partially debarked trunk and/or with polyporous fungi) and/or with holes >4 cm wide per ha Density of dead/decay oaks Number of dead and decay oaks >13 cm dbh per ha Altitude Metres above sea level

distribution and abundance of the middle spotted woodpecker in the forest tracts. For this aim, we recorded data on vegetation structure in the forest tracts using the same procedure as described in Section 3.2.1, that is, by establishing a network of Fig. 1. Network of circular plots of 0.04 ha (11.2 m radius) 100 m apart 0.04 ha circular plots 100 m apart (Fig. 1). Habitat variables are established to estimate foraging tree availability in the territories (white area) and habitat structure in the forest tracts. described in Table 2. We took two measurements of foraging tree availability: (1) the density of oaks with a diameter from 13 to 37 cm, the minimum diameter used and the lowest mean of we only considered one tree per territory randomly chosen. We the diameter used in a territory, respectively (see Section 4.2); measured 26 nest-trees and 26 non-used trees randomly- (2) the density of large oaks (dbh 37 cm). We calculated the situated within the territory limits. We only considered oaks density of potential cavity trees (i.e. oaks 22 cm dbh, dead or 22 cm dbh as available, because other tree species and smaller decay and/or with holes, see Section 4.3). As dead wood may be oaks were not used by the woodpeckers. We recorded the an important foraging substrate in winter for the middle spotted species, diameter (dbh), condition (alive, decay: trunk partially woodpecker (Pettersson, 1983), the density of dead and decay debarked and/or with polyporous fungi, or dead: with no living oaks was calculated. We also estimated the density of small parts), height and presence/absence of holes >4 cm wide (i.e. oaks (dbh < 10 cm and at least 1.5 m high) because dense the minimal size required by this woodpecker for nesting) for hardwood midstory has been shown to have a negative each tree, differentiating between natural holes and those influence on habitat use of other woodpecker species (e.g. excavated by woodpeckers. Conner et al., 1999; Rudolph et al., 2002). This was done by counting its number in four 2 m 2 m squares 5 m from the 3.2.3. Data analysis centre of each plot on each cardinal axis. Finally, altitude was Unless otherwise stated, statistical analyses in this paper measured in the centre of each plot. were conducted in SPSS 12.0. Comparisons between the characteristics of the trees used and available for foraging were 3.3.2. Data analysis carried out with paired and unpaired t-tests. We used unpaired t- Habitat variables that did not meet assumptions of normality tests and Mann–Whitney U-tests for the comparison of the trees were log10 or square root transformed (Kolmogorov–Smirnov used for foraging and nesting between the sylvopastoral normality test: all p > 0.05 after data transformation). The management systems. Comparisons between nesting trees and differences in habitat structure between the forest tracts randomly chosen trees were made with Wilcoxon singed-rank occupied and non-occupied by territories and between the tests for continuous variables, and x2-tests for categorical sylvopastoral management systems were tested using analysis variables. Fisher exact tests were used to compare categorical of variance (ANOVA). variables of nesting trees between the sylvopastoral manage- The relations between territory density of the forest tracts ment systems. Bonferroni correction was applied for multiple (n = 32) and habitat characteristics (Table 2), including also the tests performed on the same data set (Rice, 1989). forest tract size and the sylvopastoral system, were analysed by multiple linear regression. All continuous habitat variables, 3.3. Habitat structure and woodpecker distribution and with the exception of the density of small oaks and the forest abundance tract size, were significantly correlated (Pearson correlations: all p < 0.05). The highly correlated variables were reduced in 3.3.1. Field methods principal components analysis using varimax rotation to Based on the results of foraging and nesting tree use, we generate new variables non-intercorrelated. Composite vari- examine the relations between habitat structure and the ables of the principal components and the low correlated habitat H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 347 variables were analysed by backward stepwise multiple linear Table 3 regression. Diameter at breast height (cm) of used and available trees for foraging in 10 territories 3.4. Reproductive success Territory Used Available tp Mean S.E. n Mean S.E. n 3.4.1. Field methods Dehesa-1 43 12136 1562.187 0.006 From 2000 to 2004, breeding success was determined for 49 Dehesa-2 55 62031 3573.905 0.006 pairs in 26 territories. Pairs were scored as successful if they Dehesa-3 71 72432 3566.346 <0.001 produced fledglings or large nestlings (easily told from the Dehesa-4 68 51850 2 126 3.639 <0.001 changes in the begging calls of the young, see Pasinelli, 2001, Dehesa-5 58 71740 3513.041 <0.001 2003) or unsuccessful if they failed before producing large Mature forest-1 39 31333 1 185 1.927 0.055 Mature forest-2 43 31239 2580.949 0.346 nestlings. In a subsampling of accessible nests (n = 27 nests in Mature forest-3 37 51225 1 186 3.482 0.001 18 territories from 2001 to 2004), the number of nestlings 7–3 Mature forest-4 49 21030 1 199 6.485 <0.001 days before fledging were counted using a dentist mirror and a Mature forest-5 41 51326 1 151 2.923 0.004 lamp. We checked some nests for one more time before n refers to the number of trees. Comparisons were made with unpaired t-tests. fledging, but we did not find any differences in brood sizes. As we found no dead young inside the nest after fledging date (n = 18 nests), the number of large nestlings was assumed as an territory density was 0.52/10 ha (range = 0.15–0.92, adequate estimation of reproductive output. S.E. = 0.04, n = 32 forest tracts), with no significant differences between dehesas and mature forests (mean S.E. dehesas 3.4.2. Data analysis versus mature forests: 0.46 0.09 versus 0.54 0.04, The relations between brood size and territory habitat unpaired t-test: t = 1.050, p = 0.302, n = 9 and 23, respec- characteristics (Table 2, including also the sylvopastoral tively). management system) were analysed using generalized linear mixed models (GLMMs) with a Poisson error distribution in R 4.2. Trees used for foraging statistical software (available at URL http://www.R-projec- t.org). GLMMs allow both fixed and random terms to be fitted. All trees used for foraging were oaks. Non-oak trees (I. Random term in our models take into consideration multiple aquifolium, C. monogyna, Malus sp.) were very scarce in the measurements within territories (Shall, 1991). Thus, territory analysed territories (2.4/1000 trees). Diameter of used trees was identity was fitted as a random term. All continuous habitat variables, with the exception of the density of small oaks, were significantly correlated (Spearman rank correlations: all p < 0.05). The highly correlated variables were reduced in principal components analysis. Composite variables of the principal components, the density of small oaks, the sylvopastoral management system and year were fitted as fixed terms. Model selection was performed simplifying the maximum model by backward elimination of the least significant terms in order to get the minimal adequate model (Crawley, 2005). Because of the small sample size, no interaction or quadratic terms were added. Given the low number of breeding failures (n = 6), it was not possible to use GLMM analysis to examine the relationships between breeding success and habitat characteristics. Thus, the differences in breeding success between the dehesas and mature forests were analysed using the Fisher exact test. As the year had no significant effect on breeding success (Fisher exact test: p > 0.5 for all the possible combinations between years), the data of all the years were pooled.

4. Results

4.1. Sylvopastoral management and woodpecker distribution and abundance Fig. 2. Mean and standard error of the diameter (a) and height (b) of the used Middle spotted woodpecker was found in all dehesas and and available trees for foraging in dehesas (n = 5 territories) and mature forests mature forest tracts, but not in any of young forest tracts. Mean (n = 5). 348 H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352

Fig. 3. Characteristics of 26 trees used for nesting and 26 not used ones. The Fig. 4. Comparison of the characteristics of the nest trees between dehesas proportion of trees with woodpecker holes, natural holes, total holes (excavated (n = 9) and mature forests tracts (n = 17) by Fisher exact test. All p 0.1. The by woodpeckers and/or natural) and decay trees, were verified by x2-test. rest as in Fig. 3. Significance level p < 0.0125 when Bonferroni correction was applied ( p = 0.05/number of tests, Rice, 1989). All p 0.005. significantly greater than that of the available trees (mean 4.3. Trees used for nesting S.E.: 50 4 cm versus 34 2 cm, paired t-test: t = 4.882, p = 0.001, n = 10 territories), with significant All nest-trees were oaks. Non-oak trees (I. aquifolium, C. values for 8 of the 10 territories (Table 3). Diameter of the monogyna, Malus sp., Sorbus sp.) were very scarce (1.8/1000 used trees was greater in dehesas than in mature forests trees) in the analysed territories. Woodpeckers excavated (unpaired t-test; t = 3.157, p = 0.013, Fig. 2a), but the diameter cavities in oaks 22 cm diameter, using a significantly higher of available trees did not differ between both forest types proportion of decay trees (76.9% of the used trees) and trees (t = 1.690, p = 0.130, Fig. 2a). The height of the used trees did with holes (84.6%) than non-used ones (all p 0.005, see not differ significantly from that of the available ones Fig. 3). 88.5% of the used trees were decay and/or had holes (mean S.E. used versus available: 12.8 0.8 m versus (n = 26 nest-trees). Neither used trees nor random non-used 12.8 0.6 m, paired t-test: t = 0.215, p = 0.834, n =10 trees were dead. However, one nest-tree non-considered for the territories). The height of the used and available trees was analysis was dead (2.0%, n = 49 nest-trees). greater in mature forests than in dehesas, with differences being No significant differences between used and non-used trees marginally significant (unpaired t-test; used: t = 2.208, were found in the diameter (median and range, used versus non- p = 0.058; available: t = 2.239, p = 0.056) (Fig. 2b). used: 39 cm [22–87] versus 39 cm [22–75], Wilcoxon signed- Only a very small proportion of dead oaks was available rank test: Z = 0.794, p = 0.427, n = 26) and the height (1.8/1000 oaks), and was not used. Woodpeckers foraged (median and range, used versus non-used: 11 m [3–18] versus mainly in live parts of the trees (mean of observations in live 12 m [6–20], Wilcoxon signed-rank test: Z = 0.776, substrate in 10 territories: 81.4%, S.E. = 3.5, range = 66.7– p = 0.438) of the trees. In dehesas, birds nested in trees with 100.0%), with no significant differences between dehesas and larger diameter (median and range: 46 cm [35–87] versus mature forests (median dehesas versus mature forests: 76.5% 38 cm [22–66], Mann–Whitney U-test: U = 39.5, p = 0.046; versus 84.6%, Mann–Whitney U-test: U = 7.50, p = 0.295). n = 9 and 17 trees, respectively) and lower height (median and

Table 4 Results of the ANOVA for comparing the habitat variables (mean S.E.) between sylvopastoral management systems (n = 17 young forest tracts, 9 dehesa tracts and 23 mature forest tracts) Variables Woodpecker absence Woodpecker presence Fp Young forests Dehesas Mature forests Density of small oaksa 2945 432 a 855 249 b 2341 1168 ab 6.902 0.002 Density of middle oaks 152 33 a 57 12 a 389 47 b 15.292 <0.001 Density of large oaks 1 1a 54 11 b 57 8 b 21.061 <0.001 Density of potential cavity trees 2 1a 23 4b 29 3 b 26.330 <0.001 Density of dead and decay oaksb 21 9a 9 3a 48 8 b 12.055 <0.001 Altitude 1057 22 a 1015 15 a 1164 23 b 10.851 <0.001 The letters a and b indicate homogenous groups for each variable (post hoc analysis by Tukey’s HSD test). See text and Table 2 for a description of the habitat variables. a Square root transformed. b log10 transformed. H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 349 range: 11 m [7–12] versus 14 m [3–18], Mann–Whitney U-test: 5. Discussion U = 40.5, p = 0.050) than in mature forests. We found no significant differences between dehesas and mature forests with 5.1. Geographical variation in the habitat use by the regard to the remaining tree characteristics (all p 0.1, see middle spotted woodpecker Fig. 4). Middle spotted woodpecker has been considered as a habitat 4.4. Habitat structure and woodpecker distribution specialist mainly associated to mature deciduous forests rich in oaks (Pasinelli, 2000). In combination with other European Forest tracts occupied by woodpeckers (i.e. dehesas and studies, our results on habitat use support the idea that, in mature forests) had significantly higher density of large oaks landscapes dominated by oaks, the middle spotted woodpecker and of potential cavity trees than unoccupied ones (i.e. young exhibits similar habitat requirements in different areas of its forests) (Table 4). Density of middle oaks, of dead/decay oaks, distribution range. Woodpeckers selected oaks with greater and tree height were significantly greater in mature forests than diameter for foraging in our study area, in accordance with the in dehesas and young forests. Young forests had significantly only previous study carried out on this topic in oak forests by more small oak density than dehesas, whilst mature forests had Pasinelli and Hegelbach (1997) in Switzerland. Our results also intermediate values (Table 4). confirm observations of studies carried out in central Europe finding that middle spotted woodpeckers build their nests in 4.5. Habitat structure and woodpecker abundance dead parts of the tree, preferring those with old cavities and polyporous fungi (Wesołowski and Tomialojc, 1986; Kosin˜ski Principal components analysis of the continuous habitat and Winiecki, 2004). variables gave two independent components that accounted for Recent studies have found that middle spotted woodpecker 78.5% of cumulative variance in the data. Density of large oaks, also occupies non-oak forests with a high proportion of rough- of potential cavity trees and altitude had high factor loadings barked deciduous trees (Hertel, 2003; Pasinelli, 2003; Weib, with the first component (0.682, 0.850 and 0.760, respectively). 2003). Conditions in different parts of the range of a species This component was interpreted as a measurement of forest may, depending on its ability to adjust to changes, lead to maturity (older oaks are generally larger and more suitable for geographical variation in habitat selection (Rolstad et al., excavating), and was called OLDOAK. Density of middle oaks 2000). In old beech (F. sylvatica) forests in Germany, dead and of dead/decay oaks had high factor loadings with the wood plays an essential role as foraging substrate (Hertel, second component (0.904 and 0.914, respectively). This 2003), unlike the observations for populations in oak forests, component was difficult to interpret. where this woodpecker forages mainly in live parts of trees The final habitat variables (two composite variables, two (Pasinelli and Hegelbach, 1997; present study). Also, in a untransformed variables and the sylvopastoral management riverine-oak forest in Poland, Kosin˜ski and Winiecki (2004) system) were analysed using backward stepwise multiple lineal found that middle spotted woodpecker preferred to nest in dead regression. Regression analysis did not show any significant trees, most of which not oaks but ash (Fraxinus excelsior), alder effect of habitat variables on territory density of the forest tracts (Alnus glutinosa) and birch (Betula verrucosa). However, in our ( p > 0.2 for all models). study area, where the availability of other non-oak species was very limited, woodpeckers did not prefer dead trees for nesting. 4.6. Reproductive success So, the use and preference of the middle spotted woodpecker on the characteristics of trees for foraging and nesting vary Overall, 87.8% of the pairs (n = 49) were successful, with no geographically depending on the availability of the tree species significant differences between the dehesas and mature forests in the forest. This shows that this species exhibit certain (dehesas: 85.7% of 21 nests, mature forests: 89.3% of 28, Fisher plasticity in its habitat use on a regional scale (Pasinelli, 2003). exact test: p > 0.999). Despite this plasticity in habitat use, middle spotted Mean brood size was 2.96 (range = 1–5, S.E. = 0.22, n = 27). woodpeckers exhibit common characteristics for foraging Principal components analysis of the continuous habitat and nesting in different habitats of its distribution range. They variables gave two independent components that accounted normally forage on large diameter and rough bark trees, which for 75.8% of cumulative variance in the data. Similar to previous are thought to provide high food abundance (i.e. arthropods) for analyses, density of large oaks, of potential cavity trees and this woodpecker (Pasinelli, 2003). With regard to nesting, the altitude had high factor loadings with the first component (0.889, presence of soft and decaying wood seems to be more important 0.778 and 0.881, respectively). Density of middle oaks and of for cavity construction than tree species (Pasinelli, 2003) or tree dead/decay oaks had high factor loadings with the second condition. component (0.905 and 0.869, respectively). The final habitat variables (two composite variables, the density of small oaks and 5.2. Sylvopastoral management and conservation of the the sylvopastoral management system) were analysed using middle spotted woodpecker GLMMs. GLMM analysis did not show a significant relationship between brood size and habitat structure, including the Middle spotted woodpeckers avoided young forests, sylvopastoral management system (all p > 0.3). preferring dehesas and mature forests. We found no significant 350 H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 differences in both territory density and reproductive success dehesas and mature forests, which maintain high densities of between dehesas and mature forests. Thus, both dehesas and important old oaks for foraging and nesting, enable territorial low disturbed mature forests seem to be appropriate habitats for settlement and reproduction of the middle spotted woodpecker the middle spotted woodpecker, showing that a highly altered in this area. habitat (i.e. dehesas) can be also a suitable habitat for a forest An important consideration for habitat conservation of the specialist bird. Pyrenean oak dehesas also enable local middle spotted woodpecker is related to the even age economy to be maintained, and preserve a cultural heritage distribution of the trees in the highest intensively-used dehesas. exclusive to central and northern Spain (Luis-Calabuig et al., These dehesas are formed exclusively by very old oaks. Some 1993; Blanco et al., 1997; Camprodoń, 2001). This highlights old trees fall on the ground each year and, as no younger trees the conservation value of the traditional dehesa system in are available, long-term sustainability of the Pyrenean oak allowing the reconciliation of agroforestry and biodiversity. dehesas is limited. A similar situation threats the sustainability Availability of suitable foraging and nesting substrate limits of the holm oak (Q. ilex) dehesas, a habitat included within the the habitat use in many bird species (Newton, 1994; Walters EU Habitats Directive (see Pulido et al., 2001, and references et al., 2002; Luck, 2002). In sedentary birds that occupy all- therein). Following the recommendations of Pulido et al. purpose territories, such as the middle spotted woodpecker (2001), a sustainable dehesa system must include periods of (Pasinelli et al., 2001), the choice of territory is of great long-term exclusion of livestock to allow the tree regeneration importance, and must provide suitable foraging and nesting and its long-term sustainability. Once grazing pressure has been habitat (Pasinelli, 2000; Luck, 2002). Forest tracts occupied by limited, Pyrenean oaks are able to resprout at very high middle spotted woodpeckers (i.e. dehesas and mature forests) densities by vegetative regeneration (resprouting from shoot on had significantly higher densities of large oaks (dbh > 37 cm) the rhizome or the stem of subterranean roots) or regenerating and potential cavity trees densities than unoccupied ones (i.e. sexually (Ta´rrega et al., 2006). Indeed, the high density of small young forests). Large oaks and potential cavity trees provide oaks in abandoned (without livestock) dehesas indicates the the middle spotted woodpecker with suitable substrate for existence of regeneration. Also, a similar situation has been foraging and nesting, respectively (Pasinelli and Hegelbach, shown to limit habitat regeneration of species specialised on old 1997; Pasinelli, 2000; Kosin˜ski and Winiecki, 2004; present trees inhabiting forests subjected to past coppice-with-standard study). Thus, a low availability of suitable substrate for management practices in Europe and North America, including foraging and nesting could be limiting the distribution of the the middle spotted woodpecker (see Pasinelli, 2000, and middle spotted woodpecker in young forests. references therein). Conservation of species specialised on old Another potentially influential factor in habitat selection is trees requires maintaining an adequate proportion of all age competition (Mac Nally, 1990). Pasinelli (2000) did not find an strata of the trees, allowing in this way the future regeneration effect of great spotted woodpecker (Dendrocopos major) of its habitat (Pasinelli, 2000; Onrubia et al., 2004). Mature density on home range sizes of middle spotted woodpeckers. In forests, which maintain all age stages of the trees, represent a our study area, all the dehesas and mature forest tracts were suitable habitat for the persistence of such forest specialist occupied by great spotted woodpeckers, green woodpeckers species. (Picus viridis) and spotless starlings (Sturnus unicolor), Historically, oak forests in Cantabrian Mountains have been whereas such potential competitors were lacking in some cleared, felled, ploughed and burnt to increase the area used for young forests (pers. obs.), suggesting that inter-specific agriculture, livestock and mining (Purroy et al., 1984; Blanco competition is unlikely to influence the distribution pattern et al., 1997; Luis-Calabuig et al., 2000; Onrubia et al., 2004). of middle spotted woodpeckers. Large areas of oak forest have been replaced by exotic Territory quality has been shown to be one of the influencing plantations of conifers and eucalyptus (Purroy et al., 1990; factors of reproductive success in birds (Catchpole and Phillips, Luis-Calabuig et al., 1993). In recent decades, the intensifica- 1992; Luck, 2003). Similar to Pasinelli (2001), we did not find a tion of the livestock production systems promoted by the relationship between breeding success or brood size and the European Union’s Community Agrarian Policy (CAP) has structural habitat variables, which may be seen as measure- brought about a reduction in extensive sheep and goat grazing ments of territory quality (Pasinelli, 2001). As a possible systems (Serrano et al., 2002). All of these changes in land use explanation of this unexpected result, Pasinelli (2001) suggests have resulted in a decrease in the area occupied by dehesas and that, as middle spotted woodpeckers adjust the home range size mature oak forests, to such an extent that the largest surface of to the available resources (Pasinelli, 2000), differences in Pyrenean oak forest in Cantabrian Mountains is made up of habitat quality may already be compensated, so that these highly-disturbed young forests (Ta´rrega and Luis, 1990; Luis- differences do not translate into reproductive success. On the Calabuig et al., 2000; Calvo et al., 2003), which are not other hand, although no relationships between territory density occupied by this woodpecker. Thus, it is reasonable to expect and habitat structure in the forest tracts were found, habitat that distribution of the middle spotted woodpecker in Spain has quality does seem to be an important factor influencing been reduced over recent decades due to a decrease in habitat territorial occupation by middle spotted woodpecker (see availability (Onrubia et al., 2004). above). Thus, the factor limiting the reproductive output could The maintenance of viable populations requires conserva- be the choice of a habitat with sufficient quality to allow the tion actions at several spatial scales (Angelstam et al., 2004). birds to settle in a territory. Traditional management systems in Within the forest tracts, suitable densities of important foraging H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352 351 and nesting trees must be maintained (Pasinelli, 2000). As and Science (project REN 2002-03587). HR was financed by reference values, 51 territories in our study area had a mean the Junta de Castilla y Leoń and University of Leoń, and CC by density of 66 large oaks (37 cm dbh) and 30 potential cavity the Spanish Ministry of Education and Science. trees per ha, similar to the 63 oaks 36 cm dbh found in Switzerland (Pasinelli, 2000) and the 26 potential cavity trees in References Switzerland and Poland (Pasinelli, 2000; Kosin˜ski and Winiecki, 2004). At the landscape scale, management decisions A´ lvarez, A., 1989. Avifauna de los pisos de vegetacioń de la Cordillera must take into account the maintenance of traditional Canta´brica. Doctoral Thesis. University of Leoń. sylvopastoral systems in dehesas and mature forests, which Angelstam, P., Roberge, J.-M., Lo˜hmus, A., Bergmanis, M., Brazaitis, G., Do¨nz-Breuss, M., Kosinski, Z., La˜rmanis, V., Luˆkins, M., Mikusin´ski, G., support high densities of old trees essential for foraging and Racˇinskis, E., Strazds, M., Tryjanowski, P., 2004. Habitat modeling as a tool nesting. At the regional scale, the area requirements for a for landscape-scale conservation—a review of parameters for focal forest population of 100 middle spotted woodpecker females range birds. In: Angelstam, P., Do¨nz-Breuss, M., Roberge, J.-M. (Eds.), Targets from 20,000 to 130,000 ha, depending on whether it is a pure- and tools for the maintenance of forest biodiversity. Ecological Bulletin 51, habitat landscape or only 15% of the landscape is covered by 427–453. Arrambari, R., Rodrı´guez, A.F., 1996. Distribucioń y estima poblacional del habitat, respectively (see Angelstam et al., 2004). Nevertheless, Dendrocopos medius ´ 2 Pico Mediano ( )enAlava. Ardeola 43, 221–223. because only 7% of a 880 km landscape in a wider area of the Blanco, E., Casado, M.A., Costa, M., Escribano, R., Geńova, M., Go´mez, A., Cantabrian Mountains is covered by woodpecker habitat Go´mez, F., Moreno, J.C., Morla, C., Regato, P., Saínz, H., 1997. Los (unpublished data), we estimate an area requirement of Bosques Ibe´ricos. Una interpretacioń geobotańica. Planeta, Madrid, Espanã. 285,000 ha to be included in the conservation planning unit Bradshaw, R.H.W., 2004. Past anthropogenic influence on European forests and some possible genetic consequences. For. Ecol. Manage. 197, 203–212. for the middle spotted woodpecker. Moreover, as 7% is lower Bu¨tler, R., Angelstam, P., Ekelund, P., Schlaepfer, 2004. Dead wood threshold than the minimum proportion of suitable habitat patches in the values for the three-toed woodpecker presence in boreal and sub-Alpine landscape required for this woodpecker (15%, Angelstam et al., forest. Biol. Conserv. 119, 305–318. 2004), we expect habitat fragmentation, through an increment Calvo, L., Ta´rrega, R., Luis, E., 1991. Regeneration in Quercus pyrenaica of isolation among habitat patches and a reduction in patch ecosystems after surface fires. Int. J. Wildland Fire 1, 205–210. Calvo, L., Santalla, S., Marcos, E., Valbuena, L., Ta´rrega, R., Luis, E., 2003. sizes, to have negative effects (e.g. Pettersson, 1984) in this Regeneration after wildfire in communities dominated by Pinus pinaster,an population of the Cantabrian Mountains. obligate seeder, and in others dominated by Quercus pyrenaica, a typical The middle spotted woodpecker is one of the best indicators resprouter. For. Ecol. Manage. 184, 209–223. of the bird species richness in the temperate deciduous forests Camprodoń, J., 2001. Tratamientos forestales y conservacioń de la fauna of the western Palaearctic (Roberge and Angelstam, 2006) and vertebrada. In: Camprodoń, J., Plana, E. (Eds.), Conservacioń de la biodiversidad y gestioń forestal. pp. 135–179. is considered an umbrella species (Angelstam et al., 2004). This Catchpole, C.K., Phillips, J.F., 1992. Territory quality and reproductive success is because the middle spotted woodpecker is a high demanding in the Dartford Warbler Sylvia undata in Dorset, England. Biol. Conserv. 61, species very sensitive to changes in forest structure, the effects 209–215. of which may be representative for other forest birds (Roberge Conner, R.N., Rudolph, D.C., Schaefer, R.R., Saenz, D., Shackelford, C.E., and Angelstam, 2006). In fact, dehesas and mature forests (Q. 1999. Relationships among red-cockaded woodpecker group density, nest- ling provisioning rates, and habitat. Wilson Bull. 111, 494–498. pyrenaica, Q. faginea and Q. humilis), i.e. the potential habitat Crawley, M.J., 2005. Statistics: An Introduction Using. R. Wiley. for this woodpecker, maintain great bird species richness, Dıéz, C., Luis, E., Ta´rrega, R., Valbuena, L., 1991. Degradation process significantly higher than in oak forests with few large trees (i.e. in traditional systems of silvopastoral managements in dehesa young oak forests) (Camprodoń, 2001). Therefore, manage- systems dominated by Quercus pyrenaica. In: Proceedings of the Fourth ment and conservation measures previously proposed for the Congre`s International des Terres de Parcours, Montpellier, France, pp. 107–109. middle spotted woodpecker may have also implications for Duriez, O., Ferrand, Y., Binet, F., Corda, E., Gossmann, F., Fritz, H., 2005. many other forest bird species. Nevertheless, because a single Habitat selection of the Eurasian woodcock in winter in relation to earth- species would not suffice for ensuring the protection of all co- worms availability. Biol. Conserv. 122, 479–490. occurring species, combinations of species with complemen- Erickson, J.L., West, S.D., 2003. Associations of bats with local structure and tary habitat requirements should be used for biodiversity landscape features of forested stands in western Oregon and Washington. Biol. Conserv. 109, 95–102. conservation in the forests (Angelstam et al., 2004). FAO, 2001. Global Forest Resources Assessment 2000. Main Report. Forestry Paper 140. FAO, Rome, Italy. Acknowledgements Fauvel, B., 2001. E´ volution de la population du pic mar Dendrocopos medius dans le de´partement de l’aubre (Nord-est, France). Alauda 69, 162–163. ´ We thank the nearly 20 collaborators who helped us in the Garcıá-Fernańdez, J., Alvarez, E., Falagań, J., 2002. El Pico Mediano Den- drocopos medius en la Provincia de Leoń: cambios en la distribuciońy field work, especially to Marceliano Robles. Vittorio Baglione, tamanõ poblacional. Ecologıá 16, 335–342. Daniela Canestrari, Ziemowit Kosin˜ski, Tomasz D. Mazgajski, Hertel, F., 2003. Habitatnutzung und Nahrungserwerb von Buntspecht Picoides A´ ngela Taboada and two anonymous referees provided useful major, Mittlespecht Picoides medius und Kleiber Sitta europaea in comments on earlier versions of this paper, and DC also helped bewirtschafteten und unbewirtschafteten Buchenwa¨ldern des nordoust- with the statistical analyses. We are grateful to Antonio A´ lvarez deutschen Tieflandes. Vogelwelt 124, 111–132. Kosin˜ski, Z., Winiecki, A., 2004. Nest-site selection and niche partitioning who allowed us to work in the ‘‘Dehesa del Plumar’’. This study among the great spotted woodpecker Dendrocopos major and middle was partially financed by the Fundacioń MonteLeoń/Caja spotted woodpecker Dendrodopos medius in riverine forest of central Espanã, Diputacioń de Leoń and Spanish Ministry of Education Europe. Ornis Fenn. 81, 145–156. 352 H. Robles et al. / Forest Ecology and Management 242 (2007) 343–352

Kossenko, S.M., Kaygorodova, E.Y., 2001. Effect of habitat fragmentation on Pettersson, B., 1983. Foraging behaviour of the middle spotted woodpecker distribution, density, and breeding performance of the middle spotted Dendrocopos medius in Sweden. Holartic Ecol. 6, 263–269. woodpecker Dendrocopos medius in Nerussa–Desna woodland. Entomol. Pettersson, B., 1984. Ecology of an isolated population of the middle spotted Rev. 81 (Suppl. 1), S161–S166. woodpecker (Dendrocopos medius L.) in the extinction phase. Dissertation. Luck, G., 2002. The habitat requirements of the rufous treecreeper (Climacteris Swedish University of Agricultural Sciences, Uppsala. rufa). 1. Preferential habitat use demonstrated at multiple spatial scales. Plieninger, T., Pulido, F.J., Konold, W., 2003. Effects of land-use history on size Biol. Conserv. 105, 383–394. structure of holm oak stands in Spanish dehesas: implications for conserva- Luck, G., 2003. Differences in the reproductive success and survival of the tion and restoration. Environ. Conserv. 30, 61–70. rufous treecreeper (Climacteris rufa) between a fragmented and unfrag- Pulido, F.J., Dıáz, M., Hidalgo de Trucios, S., 2001. Size structure and mented landscape. Biol. Conserv. 109, 1–14. regeneration of Spanish holm oak Quercus ilex forests and dehesas: effect Luis-Calabuig, E., Ta´rrega, R., Calvo, L., Diez, C., Marcos, E., Valbuena, L., of the agroforestry use on their long-term sustainability. For. Ecol. Manage. Alonso, I., 1993. Ana´lisis fisiono´mico de los ecosistemas de Quercus 146, 1–13. pyrenaica Willd. en la cuenca del Duero. Congreso Forestal Espanõl, Purroy, F.J., Alvarez, A., Pettersson, B., 1984. La poblacioń de Pico Mediano, Lourizań, Pontevedra. Dendrocopos medius (L.), de la Cordillera Canta´brica. Ardeola 31, 81–90. Luis-Calabuig, E., Ta´rrega, R., Calvo, L., Marcos, E., Valbuena, L., 2000. Purroy, F.J., Alvarez, A., Clevenger, A.P., 1990. Bosque y fauna de vertebrados History of landscape changes in northwest Spain according to land use and terrestres en Espanã. Ecologıá (Fuera de Serie No. 1) 349–363. management. In: Trabaud, L. (Ed.), Life and Environment in the Medi- Ranius, T., Jansson, N., 2000. The influence of forest regrowth, original canopy terranean. Wit Press, Southampton, pp. 43–86. cover and tree size on saproxylic beetles associated with old oaks. Biol. Mac Nally, R., 1990. Modelling distributional patterns of woodland birds along Conserv. 95, 85–94. a continental gradient. Ecology 71, 360–374. Rice, W.R., 1989. Analyzing tables of statistical tests. Evolution 43, 223–225. Maldonado, J., Sainz, H., Sańchez, R., Xandri, P., 2001. Distribucioń y estado Roberge, J.M., Angelstam, P., 2004. Usefulness of the umbrella species concept de conservacioń de los bosques espanõles: un ana´lisis de las carencias en la as a conservation tool. Conserv. Biol. 18 (1), 76–85. red de territorios protegidos. In: Camprodoń, J., Plana, E. (Eds.), Con- Roberge, J.M., Angelstam, P., 2006. Indicator species between resident forest servacioń de la biodiversidad y gestioń forestal: su aplicacioń en la fauna birds—a cross-regional evaluation in northern Europe. Biol. Conserv. 130, vertebrada. Centre Tecnolo´gic Forestal de Catalunya. Universitat de Bar- 134–147. celona, Barcelona. Robles, H., Olea, P.P., 2003. Distribucioń y abundancia del Pico Mediano Michalek, K.G., Auer, J.A., Grobberger, H., Schmalzer, A., Winkler, H., 2001. (Dendrocopos medius) en una poblacioń meridional de la Cordillera Die Einflu¨sse von Lebensraum, Witterung, und Waldewirstchaftung auf die Canta´brica. Ardeola 50, 275–280. Brutdichte von Bunt-und Mittelspecht (Picoides major und P. medius)im Rolstad, J., Loken, B., Rolstad, E., 2000. Habitat selection as a hierarchical Wienerwald. Abhandlungen und Berichte aus dens Museum Heineanum 5 spatial process: the green woodpecker at the northern edge of its distribution (Sonderheft), 31–58. range. Oecologia 124, 116–129. Mikusinski, G., 1997. Woodpeckers in time and space: the role of natural and Rudolph, D.G., Conner, R.N., Schaefer, R.R., 2002. Red-cockaded woodpecker anthropogenic factors. Dissertation. Swedish University of Agricultural foraging behaviour in relation to midstory vegetation. Wilson Bull. 114, Sciences, Uppsala. 235–242. Newton, I., 1994. The role of nest sites in limiting the numbers of hole-nesting Serrano, E., Lavıń, M.P., Ruiz, A., 2002. Caracterizacioń de los sistemas de birds: a review. Biol. Conserv. 70, 265–276. produccioń de ganado vacuno de carne de la montanã de Leoń. CSIC-Valles Onrubia, A., Robles, H., Salas, M., Gonza´lez-Quiro´s, P., Olea, P.P., 2004. Pico del Esla, Leoń. Mediano, Dendrocopos medius. In: Madronõ, A., Gonza´lez, C., Atienza, Shall, R., 1991. Estimation of generalised linear models with random effects. J.C. (Eds.), Libro Rojo de las Aves de Espanã. Direccioń General para la Biometrika 78, 719–727. Biodiversidad-SEO/BirdLife, Madrid, pp. 304–307. Simberloff, D., 1999. The role of science in the preservation of forest diversity. Orians, G.H., Wittenberger, J.F., 1991. Spatial and temporal scales in habitat For. Ecol. Manage. 115, 101–111. selection. Am. Nat. (supplement) 137, S29–S49. Ta´rrega, R., Luis, E., 1990. La problema´tica de los incendios forestales y su Pasinelli, G., 2000. Oaks Quercus sp. and only oaks? Relations between habitat incidencia sobre los robledales de Quercus pyrenaica en la provincia de structure and home range size of the middle spotted woodpecker Dendro- Leoń. Ecologıá (Fuera de Serie No. 1) 223–237. copos medius. Biol. Conserv. 93, 227–235. Ta´rrega, R., Calvo, L., Marcos, E., Taboada, A., 2006. Forest structure and Pasinelli, G., 2001. Breeding performance of the middle spotted woodpecker understory diversity in Quercus pyrenaica communities with different Dendrocopos medius in relation to weather and territory quality. Ardea 89, human uses and disturbances. For. Ecol. Manage. 227, 50–58. 353–361. Walters, J.R., Daniels, S.J., Carter, J.H., Doerr, P.D., 2002. Defining habitat Pasinelli, G., 2003. Dendrocopos medius middle spotted woodpecker. BWP quality of red-cockaded woodpecker foraging habitat based on habitat use Update 5 (1), 49–99. and fitness. J. Wildlife Manage. 64, 1064–1082. Pasinelli, G., Hegelbach, J., 1997. Characteristics of tress preferred by foraging Weib, S., 2003. Erlenwa¨lder als bicher unbeachteter Lebensraum des Mittel- middle spotted woodpecker Dendrocopos medius in northern Switzerland. spechts Dendrocopos medius. Vogelwelt 124, 177–192. Ardea 85, 203–209. Wesołowski, T., 1989. Nest-sites of hole-nesters in a primaeval temperate forest Pasinelli, G., Hegelbach, J., Reyer, H.U., 2001. Spacing behavior of the middle (Białowiez˙a National Park, Poland). Acta Ornithol. 25, 321–351. spotted woodpeckers in central Europe. J. Wildlife Manage. 65, 432–441. Wesołowski, T., Tomialojc, L., 1986. The breeding ecology of woodpeckers in a Penas, A., 1995. Bioclimatologıá y vegetacioń. Atlas del Medio Natural de la temperate primaeval forest—preliminary data. Acta Ornithol. 22, 1–21. provincial de Leoń. I.T.G.E.-Diputacioń de Leoń, Leoń. Wiens, J.A., Rottenberry, J.T., Van Horne, B., 1987. Habitat occupancy patterns Penas, A., Llamas, F., Pe´rez-Morales, C., Acedo, C., 1997. A new species of of North American shrubsteppe birds: the effect of spatial scale. Oikos 48, Quercus L. (Fagaceae) from NW Spain. Bot. Helv. 107, 75–82. 132–147.