Introduction
Environmental conditions affecting biodiversity are changing at unprecedented rates. Thus, it is important to improve our ability to detect and prevent the most pervasive effects (Pecl et al. Reference Pecl, Araújo, Bell, Blanchard, Bonebrake, Chen and Williams2017). This is key in conserving sensitive areas where the changes may affect the long-term survival of species and ecological systems (Watson et al. Reference Watson, Darling, Venter, Maron, Walston, Possingham, Dudley, Hockings, Barnes and Brooks2016). In this context, there are issues unique to mountains that need to be addressed if we are to make predictions about how bird communities will respond under future environmental scenarios (Chamberlain et al. Reference Chamberlain, Arlettaz, Caprio, Maggini, Pedrini, Rolando and Zbinden2012).
The Mediterranean region is a climate change hotspot due to the predicted strong changes in temperature and precipitation (Giorgi Reference Giorgi2006). This region is, in addition, a biodiversity hotspot where the altitudinal decrease in temperature and increase in precipitation allows the presence of some northern and montane endemic species, which could be negatively affected by climate change (Hampe and Petit Reference Hampe and Petit2005, Giménez‐Benavides et al. Reference Giménez‐Benavides, Escudero, García‐Camacho, García‐Fernández, Iriondo, Lara‐Romero and Morente‐López2018, Barredo et al. Reference Barredo, Mauri, Caudullo, Dosio, Vilibić, Horvath and Palau2019). A common explanation for this pervasive effect is that, as temperature increases, ecological conditions at high elevations become similar to the conditions at lower sites forcing the species to move upwards (Sekercioglu et al. Reference Sekercioglu, Schneider, Fay and Loarie2008, Reif and Flousek Reference Reif and Flousek2012). This shift may affect birds since it will lead populations to retreat to increasingly smaller areas on mountaintops, from which they may ultimately disappear (La Sorte and Jetz Reference La Sorte and Jetz2010, Flousek et al. Reference Flousek, Telenský, Hanzelka and Reif2015, Lehikoinen et al. Reference Lehikoinen, Brotons, Calladine, Campedelli, Escandell, Flousek, Grueneberg, Haas, Harris, Herrando, Husby, Jiguet, Kålås, Lindstrom, Lorrillière, Molina, Pladevall, Calvi, Sattler, Schmid, Sirkia, Teufelbauer and Trautmann2019). However, these altitudinal trends may not be the rule if climate effects on species distribution are buffered or affected by other environmental changes (Lenoir et al. Reference Lenoir, Gégout, Guisan, Vittoz, Wohlgemuth, Zimmermann, Dullinger, Pauli, Willner and Svenning2010, Peters et al. Reference Peters, Hemp, Appelhans, Becker, Behler, Classen, Detsch, Ensslin, Ferger, Frederiksen, Gebert, Gerschlauer, Gütlein, Helbig-Bonitz, Hemp, Kindeketa, Kühnel, Mayr, Mwangomo, Ngereza, Njovu, Otte, Pabst, Renner, Röder, Rutten, Costa, Sierra-Cornejo, Vollstädt, Dulle, Eardley, Howell, Keller, Peters, Ssymank, Kakengi, Zhang, Bogner, Böhning-Gaese, Brand, Hertel, Huwe, Kiese, Kleyer, Kuzyakov, Nauss, Schleuning, Tschapka, Fischer and Steffan-Dewenter2019). Thus, since the final outcome of these interactions can be very idiosyncratic (Sanders and Rahbek Reference Sanders and Rahbek2013), it is important to assess the actual consequences of climate change on species distribution at the scale of each mountain range.
This paper studies the altitudinal shifts of forest birds in the Guadarrama Mountains (central Spain) over the last 35 years, a period in which the temperatures have increased significantly (Giménez‐Benavides et al. Reference Giménez‐Benavides, Escudero and Iriondo2007, Gonzalez‐Hidalgo et al. Reference Gonzalez‐Hidalgo, Peña‐Angulo, Brunetti and Cortesi2016) and some plants and animals have experienced a shift upwards (Sanz-Elorza et al. Reference Sanz-Elorza, Dana, González and Sobrino2003, Wilson et al. Reference Wilson, Gutiérrez, Gutiérrez, Martínez, Agudo and Monserrat2005, Reference Wilson, Gutiérrez, Gutiérrez and Monserrat2007, García-Romero et al. Reference García-Romero, Muñoz, Andrés and Palacios2010). Forest birds are well distributed in central Europe but tend to be constrained in the dry Mediterranean region (Tellería et al. Reference Tellería, Baquero and Santos2003), thriving mainly in moist, mountain woodlands (Tellería and Santos Reference Tellería and Santos1994). These environmental preferences make these birds and the woodlands they inhabit sensitive to upward shifts as a response to climate warming (Ruiz-Labourdette et al. Reference Ruiz‐Labourdette, Nogués‐Bravo, Ollero, Schmitz and Pineda2013). However, it can be conjectured that other environmental changes over the last few decades have also affected the potential changes in the altitudinal distribution of birds (Lenoir et al. Reference Lenoir, Gégout, Guisan, Vittoz, Wohlgemuth, Zimmermann, Dullinger, Pauli, Willner and Svenning2010). Since birds are very sensitive to habitat change, it can be predicted that alterations in woodland structure could affect their response to climate warming (Reif and Flousek Reference Reif and Flousek2012). This is the case in the ongoing process of forest recovery in central Spain (Sociedad Española de Ciencias Forestales 2010), a region under the effect of rural abandonment and the concomitant encroachment of woodlands and scrublands (Kuemmerle et al. Reference Kuemmerle, Levers, Erb, Estel, Jepsen, Müller, Plutzar, Stürck, Verkerk, Verburg and Reenberg2016).
This study explores whether altitudinal shifts of forest birds have occurred since 1976–1980 and if the altitudinal shifts are related to the effect of global warming or can be explained by changes in woodland structure. This is interesting because, if habitat structure is a main driver of bird distribution, forest management could contribute to mitigating the effect of climate change on forest birds within this Mediterranean range.
Materials and methods
Study area
The Guadarrama Mountains are distributed along a NE to SW 80-km-long belt within the Iberian Plateau, ranging from 600 to 2,400 m asl (Peñalara, the highest peak at 40.85ºN, 3.96ºW, reaches 2,430 m). They are, in turn, a minor segment of the Central Mountains System, which extends 600 km across the Iberian Peninsula (Figure 1). These mountains show decreasing temperatures and increasing precipitation with elevation (Gonzalez‐Hidalgo et al. Reference Gonzalez‐Hidalgo, Peña‐Angulo, Brunetti and Cortesi2016) and are covered by an altitudinal succession of woodland types. These woodlands are composed of sclerophyllous holm oak Quercus ilex in the piedmont (under 1,000 m) and Scots pine Pinus sylvestris in the upper parts of the mountains (above 1,400 m). Between 1,000 and 1,400 m, the woodlands are composed of Pyrenean oak Quercus pyrenaica on mountain slopes and ash Fraxinus excelsior in wet mountain valleys. All of these woodlands are managed as tree-covered pasturelands for extensive cattle breeding except in the case of pinewoods, which are managed for timber production. Most of the Guadarrama Mountains were designated a National Park in 2013 and today tourism is the main economic sector in the area (López and Pardo Reference López and Pardo2018).
a). Elevation map of the south-western Palearctic and location of the study area. Increasing dark tones show increasing elevations, with the darkest tone showing the areas over 1,500 m asl. b) Distribution of tree-covered areas. Al: Algeria, Mo: Morocco, Pt: Portugal, Sp: Spain. c) Distribution of sampling points in the elevation gradient of the Guadarrama Mountains.
Bird and habitat sampling
During May and June from 1976 to 1980, 135 circular sampling points distributed at different elevations among the four woodland types (holm oak, ash, Pyrenean oak and Scots pine) were carried out once by the author to record the presence/absence of passerines and other bird species (e.g. woodpeckers and doves; Tellería Reference Tellería1987). In this way, the presence/absence of species per sampling point detected by sight and song during a period of 10 minutes within a 100-m radius was recorded (Johnson Reference Johnson2008). In 2014–2015, 124 sampling points were repeated at the same times of the day, on the same fortnights, and at the same sites (a set of mountain tracks and woodland patches) and elevation ranges (mean ± SE, 1976–1980: 1,267.82 ± 30.64; 2014–2015: 1,263.83 ± 31.97, t = 0.09, P = 0.928) as the sampling points in 1976–1980. Unfortunately, many of the exact sampling points as in 1976–1980 were difficult to locate, given that they had not been properly georeferenced within forest patches. However, despite this uncertainty, the data used in this paper were collected by the same methodology so that they can provide strong inferences about bird distribution changes when compared directly (Tingley and Beissinger Reference Tingley and Beissinger2009).
Bird counts during 2014–2015 were combined with an assessment of habitat structure in 25-m-radius circles around each sampling point. In each circle, the cover (percentage) of grass, shrub (vegetation < 0.5 and between 0.5 and 2 m height), and tree layers (> 2 m) were recorded visually, as well as the mean height (m) of the canopy and the number of shrub and tree species. These data were used to carry out a principal components analysis to obtain a small number of orthogonal variables. Two principal components were retained, which were interpreted as gradients of increasing shrub cover and richness (PC1, eigenvalue: 2.09; explained variance: 29.84%; factor loadings, grass layer: -0.456; shrub cover under 0.5 m: 0.790; shrub cover 0.5–2 m: 0.818; tree cover > 2 m: -0.099; mean tree height: -0.307; shrub species: 0.577; tree species: 0.399) and increasing tree cover (PC2, eigenvalue: 1.91; explained variance: 27.21%; factor loadings, grass layer: -0.586; shrub cover under 0.5 m: 0.172; shrub cover 0.5–2 m: 0.126; tree cover >2 m: 0.875; mean tree height: 0.808; shrub species: -0.098; tree species: -0.297). The factor scores of each sampling point within these two components were used as comprehensive indexes of habitat structure (see below).
Elevation shifts of bird species
Elevation has been defined as a "topoclimate" variable useful to assess climate in mountains where it is difficult to interpolate climate conditions from the typically scarce meteorological stations (Gutiérrez-Illán et al. Reference Gutiérrez-Illán, Gutiérrez and Wilson2010). Thus, a shift upwards in the mid-range elevation of individual species could suggest a displacement upwards of their climate requirements due to the effect of warming. Sampling points for the two study periods provided two sets of elevations per individual species that were used to explore the potential shifts upwards of their mid-range elevation as a response to climate warming. In addition, to detect the combined altitudinal shift of bird species, the whole set of mean elevations per species were compared with a t-test for dependent samples. Finally, mean factor scores of individual species in PC1 and PC2 (SHRUB and TREE therein) were used to assess their habitat preferences and explore if the altitudinal shifts were related to their habitat preferences. In all cases, species with sample size < 5 presences in either period were discarded from analyses.
Habitat changes perceived bird communities
SHRUB and TREE scores of all individual species were averaged to obtain two indices of the mean habitat preferences by bird communities in the two study periods. These indices (SHRUB-community and TREE-community) are used as surrogates of the habitat structure perceived by birds to speculate on the habitat changes experienced by woodland structure. The use of these indices needs to be supported by the assumption that they are able to predict the actual woodland structure (see below). The between-period changes in these SHRUB and TREE community indices were studied by general linear models in which elevation and elevation2 were included as covariates to detect a potential polynomial relationship, and period as factor. These and other analyses in this study were carried out with R Commander (http://CRAN.R-project.org).
Results
Elevational shifts of bird species
A total number of 73 bird species were detected in the Guadarrama Mountains, 65 in 1976–1980 and 68 in 2014–2015. Considering the 42 species with sample sizes over five in each study period (Table S1 in the online supplementary material), a significant decrease in the mean elevation range was detected by using Student’s t-tests in 13 species (Common Linnet Linaria cannabina, Short-toed Treecreeper Certhia brachydactyla, European Pied Flycatcher Ficedula hypoleuca, European Robin Erithacus rubecula, Eurasian Jay Garrulus glandarius, Crested Tit Lophophanes cristatus, Woodlark Lullula arborea, Coal Tit Periparus ater, Goldcrest Regulus regulus, European Serin Serinus serinus, Garden Warbler Sylvia borin, Northern Wren Troglodytes troglodytes and Eurasian Blackbird Turdus merula) while the remaining 29 species showed no significant changes (Table S1). Combining the 42 species, the mean (± SE) elevation of the whole set of forest birds experienced a significant decrease (from 1,306 ± 50.40 m to 1,212 ± 37.52 m; t-test for dependent samples t = 3.98, P < 0.001; Figure 2). Birds related to tree-developed woodlands (TREE scores) were particularly involved in the altitudinal shifts (r: -0.40, P < 0.01, n =42; Figure 2), while the distribution of species along the shrub gradient (SHRUB scores) did not fit this pattern (r = 0.06, P = 0.710, n = 42). Thus, at the scale of individual species, the reported trends do not support a shift upwards of forest birds but rather a shift downwards of the most tree-dependent species.
a). Relationship between the mean elevation of bird species between the two study periods along the elevation gradient of the Guadarrama Mountains. b) Relationship between the mean elevation shifts of individual species from 1976–1980 to 2014–2015 and their mean factor scores along the principal component interpreted as a tree cover gradient (TREE). Discontinuous lines indicate no changes in bird distribution.
Habitat changes perceived by bird communities
Observed tree cover in 2014–2015 fitted well to the TREE-community index (the main variable involved in PC2). This supports the usefulness of this index as a predictor of woodland structure (Figure 3). Consequently, the changes in this index were compared between periods along the elevation gradient to explore the changes in woodland structure. The results showed that the TREE-community index was higher in the most elevated woodlands (pinewoods) within the mountains (altitude: F1,254 = 20.51, P < 0.001, altitude2: F1,254 = 1.88, P = 0.171) and that it had increased from 1976–1980 to 2014–2015 (period: F1,254 = 15.95, P < 0.001) along the elevation range (whole model R2 = 0.71, F3,254 = 208.72, P < 0.001). Thus, the habitat changes perceived by bird communities support an increase in tree cover during the last few decades, a process particularly evident at mid-altitudes (over 1,200 m; Figure 3). This means that the availability of tree-covered sites has increased downwards, a change that agrees with the reported shift downwards of tree-dependent birds (Figure 2).
a). Relationship between the TREE-community index and the actual tree cover reported by bird and habitat sampling in 2014–2015. b) Altitudinal shift of TREE-community index between periods.
Discussion
The results in this paper show that the mean range position of birds associated with tree cover shifted downwards, a trend that suggests that an increase in tree cover has buffered the altitudinal shifts of forest birds predicted by climate warming.
It is commonly agreed that climatic variables affect species distribution at larger scales while land cover, habitat structure and biotic interactions are increasingly important at lower scales (Pearson and Dawson Reference Pearson and Dawson2003, Seoane et al. Reference Seoane, Bustamante and Díaz-Delgado2004, Thuiller et al. Reference Thuiller, Araújo and Lavorel2004). However, this balance may change in mountain ranges where sharp variations in climate, due to elevation and/or orientation, affect the distribution of species at lower scales (Sanders and Rahbek Reference Sanders and Rahbek2013). This effect of climate is supported by the presence of northern species in many Mediterranean ranges south of the Palaearctic and the withdrawal of species to the upper areas of mountains as a response to climate warming (Hampe and Petit Reference Hampe and Petit2005, Engler et al. Reference Engler, Randin, Thuiller, Dullinger, Zimmermann, Araujo, Pearman, Le Lay, Piedallu, Albert, Choler, Coldea, De Lamo, Dirnböck, Gégout, Gomez-García, Grytnes, Heegaard, Hoistad, Nogués-Bravo, Normand, Puscas, Sebastià, Stanisci, Thurillat, Trivedi, Vittoz and Guisan2011). However, there is some evidence on the effect of other factors on these patterns since a species range does not always shift upwards as a response to climate change (Lenoir et al. Reference Lenoir, Gégout, Guisan, Vittoz, Wohlgemuth, Zimmermann, Dullinger, Pauli, Willner and Svenning2010). This can be related, for instance, to the structure of available habitats, since it has been shown that northern, cold climate birds in the Western Palaearctic tend to have niche positions in closed sectors of woodlands (Barnagaud et al. Reference Barnagaud, Devictor, Jiguet, Barbet-Massin, Le Viol and Archaux2012, Reference Barnagaud, Barbaro, Hampe, Jiguet and Archaux2013). From this, it follows that the altitudinal distribution of species will vary according to their habitat requirements and the distinctive arrangement of environmental traits in each mountain range.
Previous studies on the altitudinal shifts of birds support the idiosyncratic response of species to climate change. For instance, European birds have experienced a shift upward in some mountains (e.g. Popy et al. Reference Popy, Bordignon and Prodon2010, Maggini et al. Reference Maggini, Lehmann, Kéry, Schmid, Beniston, Jenni and Zbinden2011, Reif and Flousek Reference Reif and Flousek2012) but not in others (Archaux Reference Archaux2004), as has been reported in the Guadarrama Mountains. In this case, the mean altitudinal range of birds has shifted 90 m downwards during the last few decades in a process in which the most tree-dependent species have been particularly involved (Figure 2). In addition, the predicted trend in habitat structure shows that tree cover has increased in lower areas during the last few decades (Figure 3). Both pieces of evidence suggest that the most tree-dependent birds have tracked the increasing availability of tree cover in lower elevations, a process that has reversed the shift upwards predicted by climate warming in this mountain range (Wilson et al. Reference Wilson, Gutiérrez, Gutiérrez, Martínez, Agudo and Monserrat2005, Ruiz-Labourdette et al. Reference Ruiz‐Labourdette, Nogués‐Bravo, Ollero, Schmitz and Pineda2013, Gonzalez‐Hidalgo et al. Reference Gonzalez‐Hidalgo, Peña‐Angulo, Brunetti and Cortesi2016).
The processes involved in these trends are not difficult to explain considering two issues related to the habitat preferences of forest birds. First, tree cover increase improves the availability of some key resources for birds that rely on trunks and foliage (e.g. Short-toed Treecreeper, Crested Tit, Coal Tit, Goldcrest). Second, tree-covered sites are able to maintain colder and wetter conditions compared to open areas (Suggitt et al. Reference Suggitt, Gillingham, Hill, Huntley, Kunin, Roy and Thomas2011) favouring the local survival of organisms associated with moist, dense understorey (Barnagaud et al. Reference Barnagaud, Devictor, Jiguet, Barbet-Massin, Le Viol and Archaux2012, Reif and Flousek Reference Reif and Flousek2012). This is the case of some species that shifted downwards in the Guadarrama Mountains (e.g. Northern Wren, European Robin, Eurasian Blackbird; Table S1), and are usually found in moist forest understorey in the Mediterranean woodlands. Thus, reported trends in the mean range position of forest birds result from the downwards encroachment of tree cover in the Guadarrama Mountains. These results are congruent with the reported expansion of forest birds in other areas of the Iberian Peninsula, where tree recovery resulting from rural abandonment and/or reforestation has sharply increased the availability of suitable habitat for these species (Seoane and Carrascal Reference Seoane and Carrascal2008, Gil-Tena et al. Reference Gil-Tena, Brotons and Saura2009).
Conclusions
The results in this paper suggest three main conclusions that could be used as management guidelines aimed at improving the resilience of forest bird communities. First, the results support the idea that any practice-oriented conclusion directed at moderating the effects of climate change will be affected by context-dependent processes that need to be investigated (Lenoir et al. Reference Lenoir, Gégout, Guisan, Vittoz, Wohlgemuth, Zimmermann, Dullinger, Pauli, Willner and Svenning2010, Rocchia et al. Reference Rocchia, Luppi, Dondina, Orioli and Bani2018). This means that the resilience of forest birds to climate change in the Guadarrama Mountains does not necessarily apply to other taxonomic groups, nor to birds in open habitats that will probably shift upwards as detected in other European mountains (Reif and Flousek Reference Reif and Flousek2012). In this context, the altitudinal shift of forest birds observed in this paper is a striking outcome if compared to other animals within this mountain range. Butterflies, for instance, have shifted upwards by around 200 m in the last 35 years (Wilson et al. Reference Wilson, Gutiérrez, Gutiérrez and Monserrat2007), apparently because butterflies respond consistently to topoclimate across a wide range of habitats (Gutiérrez-Illán et al. Reference Gutiérrez-Illán, Gutiérrez and Wilson2010, Nieto-Sánchez et al. 2015). This uneven balance of climate vs. habitat change on the distribution of both groups could explain why butterflies outrun birds in the race to shift the range in response to climate change (Devictor et al. Reference Devictor, Van Swaay, Brereton, Brotons, Chamberlain, Heliölä, Herrando, Julliard, Kuussaari, Lindström, Reif, Roy, Schweiger, Settele, Stefanescu, Van Strien, Van Turnhout, Vermouzek, WallisDeVries, Wynhoff and Jiguet2012) and strongly suggests the need to consider species traits in the study of these range shifts (Estrada et al. Reference Estrada, Morales-Castilla, Caplat and Early2016).
Second, the idiosyncratic response of species to environmental changes raises the problem of the representativeness of focal species when assessing the effect of environment changes (Caro and O’Doherty Reference Caro and O’Doherty1999). This problem occurs even within the set of forest birds, many of which are not adapted to dense tree patches but thrive within the clearings resulting from natural and human disturbances (Bengtsson et al. Reference Bengtsson, Nilsson, Franc and Menozzi2000). The results in this paper suggest, for instance, a decrease in the occurrence of some endemics (e.g. Citril Finch Carduelis citrinella) or northern birds (e.g. Dunnock Prunella modularis; Table S1) related to forest clearings in the upper part of these mountains, a trend already detected in the Iberian mountains (Lehikoinen et al. Reference Lehikoinen, Brotons, Calladine, Campedelli, Escandell, Flousek, Grueneberg, Haas, Harris, Herrando, Husby, Jiguet, Kålås, Lindstrom, Lorrillière, Molina, Pladevall, Calvi, Sattler, Schmid, Sirkia, Teufelbauer and Trautmann2019). Thus, despite the fact that the conservation of forest birds within a context of climate warming implies the protection of tree-covered sectors, it is important to avoid the pervasive effect on biodiversity of widespread forest densification (Clavero and Brotons Reference Clavero and Brotons2010, Regos et al. Reference Regos, Domínguez, Gil-Tena, Brotons, Ninyerola and Pons2016). This can be a difficult task since forest management to conserve biodiversity must be reconciled with other approaches aimed at improving wood production and/or carbon storage and sequestration capacity as part of global mitigation efforts (Felton et al. Reference Felton, Gustafsson, Roberge, Ranius, Hjältén, Rudolphi, Lindbladha, Wesliend, Riste, Brunet and Felton2016).
Finally, the trends of forest birds in the Guadarrama Mountains support an opposite effect of climate vs. land use changes, the two main drivers of biodiversity in a context of global change (Sala et al. Reference Sala, Chapin, Armesto, Berlow, Bloomfield, Dirzo, Huber-Sanwald, Huennke, Kackson, Kinzig, Leemans, Lodge, Mooney, Oesterheld, Poff, Sykes, Walker and Wald2000, Clavero et al. Reference Clavero, Villero and Brotons2011, Sirami et al. Reference Sirami, Caplat, Popy, Clamens, Arlettaz, Jiguet, Btrotins and Martin2016) that also applies to mountain ranges (Peters et al. Reference Peters, Hemp, Appelhans, Becker, Behler, Classen, Detsch, Ensslin, Ferger, Frederiksen, Gebert, Gerschlauer, Gütlein, Helbig-Bonitz, Hemp, Kindeketa, Kühnel, Mayr, Mwangomo, Ngereza, Njovu, Otte, Pabst, Renner, Röder, Rutten, Costa, Sierra-Cornejo, Vollstädt, Dulle, Eardley, Howell, Keller, Peters, Ssymank, Kakengi, Zhang, Bogner, Böhning-Gaese, Brand, Hertel, Huwe, Kiese, Kleyer, Kuzyakov, Nauss, Schleuning, Tschapka, Fischer and Steffan-Dewenter2019). In this context, the Mediterranean climate change hotspot balances between the effect of woodland encroachment from the effect of rural abandonment (Navarro and Pereira Reference Navarro and Pereira2015, Kuemmerle et al. Reference Kuemmerle, Levers, Erb, Estel, Jepsen, Müller, Plutzar, Stürck, Verkerk, Verburg and Reenberg2016) and a persistent increase in tree mortality associated with climate induced physiological stress and interactions with other climate-mediated processes, such as insect outbreaks and wildfires (Allen et al. Reference Allen, Macalady, Chenchouni, Bachelet, McDowell, Vennetier, Kitzberger, Rigling, Breshears, Hogg, Gonzáleez, Fensham, Zhang, Castro, Demidova, Lim, Allard, Running, Semerci and Cobb2010). Since the results in this paper support the prevalence of habitat changes on the long-term dynamics of forest birds in the Guadarrama Mountains, it may be worth considering the potential of forest management to moderate, in the medium term, the regional effects of climate change on forest biodiversity. After all, tree cover could create localised microclimatic conditions that allow populations of forest species, in many cases adapted to colder, northern environments, to survive in adverse climate conditions (Barnagaud et al. Reference Barnagaud, Devictor, Jiguet, Barbet-Massin, Le Viol and Archaux2012, Reference Barnagaud, Barbaro, Hampe, Jiguet and Archaux2013, Suggitt et al. Reference Suggitt, Wilson, Isaac, Beale, Auffret, August, Bennie, Crick, Duffield, Fox, Hopkins, Macgregor, Morecroft, Walker and Maclean2018). This proactive reaction to global change could be particularly important in small Mediterranean mountain ranges, such as the Guadarrama Mountains, in which it will be increasingly difficult to maintain viable populations of many forest species under the effects of climate warming (Engler et al. Reference Engler, Randin, Thuiller, Dullinger, Zimmermann, Araujo, Pearman, Le Lay, Piedallu, Albert, Choler, Coldea, De Lamo, Dirnböck, Gégout, Gomez-García, Grytnes, Heegaard, Hoistad, Nogués-Bravo, Normand, Puscas, Sebastià, Stanisci, Thurillat, Trivedi, Vittoz and Guisan2011, Ruiz-Labourdette et al. Reference Ruiz‐Labourdette, Nogués‐Bravo, Ollero, Schmitz and Pineda2013).
Supplementary Materials
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S0959270919000455.
Acknowledgements
This paper is a contribution to the project CGL2017-85637-P (Life at the border: population differentiation of forest birds south of the Palearctic) granted by the Spanish Ministry of Economy and Competitiveness. Robert Wilson and an anonymous referee considerably improved an early version of this manuscript.
