Introduction

Human activities have developed a peculiar agrosilvopastoral ecosystem, which dominates the landscape in the south-western Iberian Peninsula. This ecosystem is called “dehesa” in Spain and “montado” in Portugal, and covers approximately 3.5–4.0 million hectares in western Spain (San Miguel 1994; Olea et al. 2005) and 0.7 million hectares in Portugal (Pereira et al. 2004). The Spanish term “dehesa” (from the Latin word defesa) appears for the first time in the 10th century, referring to areas used for pasture and rearing livestock, mainly bulls and/or Iberian pigs (Klein 1920) and these practices are currently still in use (García-González 1979; Gómez-Gutiérrez 1992). Despite the origin and maintenance of the “dehesa” being due to livestock and farming uses, hunting activities have also taken place, thus creating different areas with lower and higher shrub coverage in the “dehesa” (Cabo 1976).

This ecosystem is the result of the transformation of the original Mediterranean forest by clearance and brushwood removal to introduce cattle, providing a typical landscape of savannah-like open woodland with scattered trees (Campos et al. 2013). Traditional practices in the “dehesa”, such as clear-cut stands, produce different successional stages and an ever-changing land-use mosaic with high socioeconomic and conservation values (Díaz et al. 1997; Bugalho et al. 2011). According to its long-term history of management, the “dehesa” is considered as a well-established cultural landscape (Díaz et al. 1997), and is included in the Habitat Directive (Annex I 6310 “Dehesas” with evergreen Quercus spp) and also in Natura 2000 Network (www.natura.org/).

The dominant trees in “dehesas” are Quercus species (Hernández 1995; Joffre et al. 1999; López-Sáez et al. 2007; Llorente-Pinto 2011) that are distributed in low density, varying between 20 and 50 trees/ha, in order to obtain good acorn production for livestock (Vázquez et al. 1999; Pulido and Diaz 2005). Usually, trees in the “dehesa” have a large diameter related to the age and openness of the ecosystem (Plieninger et al. 2003). This scattered tree distribution is one of the main factors that determine the richness and abundance of animal diversity in the “dehesa” (Gómez-Gutiérrez 1992; Díaz et al. 1997). The trees are also considered to be “keystone-structures” (Manning et al. 2006), as they provide food resources and refuge for a wide range of animals (Tews et al. 2004).

Up until now, only some groups have been studied: mammals (Malo and Suárez 1995), birds (Pulido and Díaz 1992; Díaz 2003), lizards (Martín and López 2002), ants (Reyes-López et al. 2003), Calliphorid Diptera (Martinez-Sánchez et al. Martínez-Sánchez et al. 1998), dung- beetles (Galante et al. 1991, 1994, 1995) and Carabids (Taboada et al. 2006, 2010). However, saproxylic insects have been poorly studied in the “dehesa” (see Micó et al. 2010, 2011a; Viñolas 2012), in spite of their dependence on mature trees. Saproxylic insects are key organisms in the trophic chain of forests due to their high diversity and biological diversification (Stokland et al. 2012). In general, Coleoptera and Diptera (Syrphidae) are the best represented among saproxylic insects (Dajoz 1998) and they also include a high percentage of threatened species (Nieto and Alexander 2010; Radenković et al. 2013). Moreover, many of them (i.e. beetles) are involved in important ecosystem services, such as breaking down deadwood and recycling nutrients (Dajoz 1998; Thompson and Rotheray 1998; Brustel 2001; Micó et al. 2011b; Stokland et al. 2012).

Micó et al. (2011a) revealed that scattered trees in the “dehesa” with a long-term history of management were an important reservoir of endangered saproxylic beetles and syrphids. However, until now, no study has analyzed the wide saproxylic community of the “dehesa” and how its management affects its diversity. The main goal of this study was to analyze the distribution of saproxylic Coleoptera and Syrphidae assemblages in a protected “dehesa” ecosystem from the western Iberian Peninsula, and to assess the effects of tree density on the distribution of the richness and abundance of Coleoptera and Syrphidae saproxylic species. We also analyzed the possible influence of dominant trees species and vegetation coverage on species diversity and composition of both taxa. The use of two saproxylic taxa with different dispersal abilities, habitat and food requirements, was to provide a broader view of how landscape management could affect saproxylic diversity.

Methods

Study area

Field work was carried out at the Biological Reserve “Campanarios de Azaba” (LIFE-Nature “Conservación de la Biodiversidad en el Oeste Ibérico”, LIFE/E/NAT/000762), western Spain (Salamanca province) (40 29.769 N and 6 47.551 W) with an extension of 522 ha and an altitude of 800 m. The annual average precipitation is 563 mm and temperatures range from −3 to 35 °C (annual mean temperature around 12 °C) (Gómez-Gutiérrez 1992; Llorente-Pinto 2011). The landscape is a typical “dehesa”, savannah- like open woodland with scattered trees (Gómez-Gutiérrez 1992), dominated by Quercus pyrenaica Willd, and Q. rotundifolia Lam that coexist with some individuals of Q. faginea Lam. It has been a private protected area since 2009 and is managed by the “Fundación Naturaleza y Hombre” (FNYH: http://www.fnyh.org/proyectos-life/reserva-campanarios-azaba/) and has recently been catalogued as the first Entomological Reserve of the Spanish Association of Entomology (http://www.entomologica.es/).

This “dehesa” was subjected to intensive agricultural use, livestock farming and small game hunting before its protection. Tree management has been carried out in two ways: (a) pollarding that consists of aggressive pruning to strip the tree of many of its upper branches, maintaining the main branches in order to obtain a semispherical crown that is as broad as possible for greater acorn production. (b) Lighter pruning, which removes some of the excess branches (mainly small branches) or dead ones (Gómez-Gutiérrez 1992).

Despite tree species having been exposed to the same management, Q. pyrenaica trees show a stronger pollard compared to Q. rotundifolia (FNYH. pers. comm.; Fig. 1). Pollarding practices allow the formation of many scars and cavities in the trunk of the trees, which are colonized by a wide range of taxa, mainly saproxylic insect species. Nowadays, the reserve is still managed using both practices, pollarding and lighter pruning, but under a protocol of suitable management and with hunting prohibited (Sánchez-Martínez et al. 2012).

Fig. 1
figure 1

Dominant tree species from the Biological Reserve of Campanarios de Azaba exhibiting the structure that is modeled by tree management (pollarding) showing different kinds of cavities. a Quercus rotundifolia, b Q. pyrenaica

Full size image

Sampling design

We selected 13 sampling units covering the whole extension of the reserve. Adults of saproxylic insects were sampled using Emergence Traps (ET) and Window Traps (WT). ET consist of a black acrylic mesh that covers and seals the tree hollow, and a catcher pot containing ethylene–glycol as preservative is attached to the mesh (Micó et al. 2010, 2011a; Gouix and Brustel 2011; Quinto et al. 2013). On the other hand, WT consist of two crossed rectangular plates of transparent methacrylate, under which a funnel is located that is attached to a catcher pot containing ethylene–glycol (Quinto et al. 2013). Traps were placed in front of the tree hollows, at a height of 1–2 m from the ground (Økland 1996; Bouget et al. 2008a). Saproxylic syrphids species that inhabit tree hollows were only collected with ET because the adults’ flight and their behavior make WT inefficient for collecting them. Both WT and ET were useful for sampling saproxylic beetles, as ET collect beetles that inhabit tree hollows and WT collect saproxylic flying beetles associated with different microhabitats offered by the tree and its immediate surroundings (Quinto et al. 2013). Both methods (ET and WT) are complementary for the monitoring of the saproxylic Coleoptera (Quinto et al. 2013).

A total of 31 ET and 29 WT were distributed across the 13 sampling units. The traps were distributed according to the dominant tree in the following way: 33 traps for Q. rotundifolia, 22 for Q. pyrenaica and 5 for Q. faginea (Table 1). Traps were active from April 2010 to April 2011 and pots were replaced monthly.

Table 1 List of the sampling locations with their zones of uses, dominant tree species, and density of trees per hectare and distribution of traps related to tree species and type of trap
Full size table

“Dehesa” structure

In order to determine the spatial structure of the “dehesa”, the openness of the canopy was estimated as tree density (trees per hectare) using a map (scale 1:5,000) of the reserve downloaded from the WMS (Web Map Service) server (http://www.idee.es/wms/PNOA/PNOA/). Tree density estimation was computed with the free software gvSIG v1.11 (http://www.gvsig.org/) using the SEXTANTE tool to draw a quadrant of 100 × 100 m for each of the 13 units sampled and then all the trees were counted in each of the selected quadrants.

In order to describe the vegetation coverage (determined by the canopy, shrub and grass coverage) we selected two samples of 10 × 10 m in each sampling unit, where the canopy, shrub and grass coverage were visually estimated by a botanist (see acknowledgements) following Gómez-Sal et al. (1992).

To characterize the dominant tree species (Q. rotundifolia and Q. pyrenaica), we measured the tree diameter and the tree hollow volume of the trees where the traps were placed. In order to calculate the tree diameter, we measured the tree circumference at both 0.30 m and 1.30 m, and we performed the equation \(diam = \frac{P}{\pi }\) where P is the average circumference of the two considered heights. Tree hollow volume was measured as a cylinder with a circular base with the equation \(Volth = \pi r^{2} h\) where r is the radio of the tree hollow at opening level and h the total height of the tree hollow.

Species identification

Nomenclature is according to the Fauna Europaea (http://www.faunaeur.org/), Bouchard et al. (2011) and the Catalogues of Palaearctic Coleoptera (Löbl and Smetana 2004, 2006, 2007, 2008). Syrphids were identified using the van Veen (2004) and Speight (2010) keys as well other specialized taxonomic papers for each group. Several European specialists in saproxylic beetles collaborated in species identification (see acknowledgments). The specimens are deposited in the Entomological Collection of the University of Alicante (CEUA).

Diversity analysis

We evaluated the completeness of the species inventory as the percentage of observed species in relation to the number of species predicted by two richness estimators: an abundance-based one and an incidence-based one (Chao 1 and Chao 2, respectively) with the EstimateS 8.0 program (Colwell 2005, 50 random permutations).

In order to assess whether tree density affects the saproxylic insect assemblages, we performed the Spearman Rank Order Correlation test between tree density versus richness and abundance per unit sampled. Correlations were performed with STATISTICA software v7 (http://www.statsoft.com/).

Due to the low abundance of Q. faginea trees in the reserve, they were not considered for the other ecological analyses; however, the material collected in Q. faginea is listed in Supplementary Table S1. Individual-based rarefaction curves with 95 % confidence intervals were used to assess differences in species richness according to dominant tree species (Q. rotundifolia and Q. pyrenaica), and vegetation coverage. Significant differences were defined when curves were situated outside the confidence limits of the other curves at the point where the less sampled assemblage ended (Micó et al. 2013). Rarefaction uses probability theory to obtain the expected richness of species and their variance for a certain size of sample, allowing sample sizes of different ecological assembles to be standardized (Magurran 2004). Rarefaction curves were computed with EcoSim software v7 (Gotelli and Entsminger 2004).

In the same way, differences in species composition were estimated as an inverse measure of turnover by calculating the Jaccard Similarity Index (Jaccard 1912; Magurran 2004). This index is based on the relationship of presence-absence between the number of common species in two areas (or communities) and the total number of species (Kent and Coker 1992). A similarity analysis (ANOSIM; Faith et al. 1987) was performed to test the significance of these differences using PRIMER software v6 (Clarke and Gorley 2006).

Canonical Correspondence Analysis (CCA) was conducted on species-variable matrices to analyze the importance of the “dehesa” structure (tree density, dominant tree species and shrub coverage), identifying which factors are responsible for species composition in structuring saproxylic insect assemblages. We used the option for downweighting rare species and Hill’s scaling. Global permutation tests (carried out by Monte Carlo permutation tests with 199 permutations under the full model) were performed to judge the statistical significance of the relationship between beetles and syrphid assemblages and “dehesa” structure variables. CCA was computed with CANOCO software for Windows v4.5 (ter Braak and Šmilauer 2002).

Indicator species analysis

We looked for indicator species for both tree species and vegetation coverage following Dufrêne and Legendre (1997). This method gives indicator values (IndVal) for each species, and the Monte Carlo Test (Crowley 1992; Manly 1997) was then used for statistical support. The IndVal method combines measures of habitat fidelity (frequency within that habitat type) and specificity (uniqueness to a particular site) to identify the characteristic species of each habitat. IndVal was performed with PC-ORD software v4 (McCune and Mefford 1999). As with previous saproxylic studies (see Micó et al. 2013), species with significant IndVal greater than 60 % were considered as characteristic species for the studied variables.

Results

“Dehesa” structure

As a result of forest clearance, tree density is low in a “dehesa”. It varied from 16 to 69 trees/ha among sites, with an average 39 trees/ha (Table 1) in the study area. Canopy coverage was uniform in the whole “dehesa” (50 % coverage) and grass cover was over 75 %. In contrast, “dehesa” openness was mainly determined by differences in shrub coverage, which varied among sites depending on the use of each site in the past (livestock, hunting, etc.). Generally, an oak “dehesa” is considered to be dense when more than 50 % of shrub coverage is observed (Gómez-Sal et al. 1992), we considered values of 30 % of vegetation coverage, which also allowed us to include both dense and intermediate coverage areas under this category. In this way, in the studied “dehesa”, areas characterized by abandonment of livestock farming showed notable regeneration in the shrub stratum (between 30 and 90 % of shrub coverage) dominated by Cytisus multiflorus Sweet, 1826 and diverse grassland. In contrast, areas destined to extensive livestock farming (following Olea and San Miguel 2006) showed very low shrub coverage (between 5 and 25 %).

When the tree structure was analyzed, we found that the average of both diameter and tree hollow volume varied between the two dominant tree species, Q. rotundifolia and Q. pyrenaica. Q. rotundifolia was the tree species with a larger diameter (0.62 m) compared to Q. pyrenaica (0.52 m), but when tree hollow volumes were analyzed, Q. rotundifolia hosted smaller ones (0.014 m3) than Q. pyrenaica (0.057 m3).

Species richness and turnover

We collected a total of 3,735 individuals of saproxylic beetles belonging to 136 species (40 families) and 289 individuals of syrphids from 16 species (see Supplementary Table S1). The species richness estimators (Chao1, Chao2) suggested quite reliable inventories (Table 2).

Table 2 Observed species richness (Sobs) and percentages of inventory completeness
Full size table

As regards saproxylic beetles, we found that 7 of the recorded species are included in the European Red List of Saproxylic Beetles or/and in the Red Book of Invertebrates from Spain (See Supplementary Table 1). Among them we found Limoniscus violaceus (Muller, 1821), a species that is included in the European Habitat Directive Annex II. We collected 3 individuals in 2010 and 2 more individuals in a second sampling period in 2011 (unpublished data). Moreover, to this list of 7 species we have to add another remarkable saproxylic species from the Habitat Directive, Cerambyx cerdo mirbeckii (Lucas, 1842). This was also registered in this protected “dehesa” in a second sampling period; all associated with Q. pyrenaica species (Micó et al. 2011a). Among saproxylic syrphids, we recorded Mallota dusmeti Andréu, 1926, which is catalogued as vulnerable (following the IUCN criteria), in the Red List of Invertebrates of Spain (Marcos-García 2006) and endangered for Europe (Speight 2010).

Despite tree density (tree individual/ha) being highly variable in the study area, we did not find any statistical significance when we correlated tree density with beetle diversity (abundance rs −0.21; p 0.49; richness rs 0.29; p 0.32), or with syrphids (richness rs −0.19; p 0.51; abundance rs −0.27; p 0.36).

Differences in cumulative species richness were observed among the dominant tree species in the reserve (Q. pyrenaica and Q. rotundifolia). Comparing species rarefaction curves at the lower abundance value for beetles (n = 1,513 individuals) and syrphids (n = 106 individuals), we found that Q. pyrenaica showed the highest species richness for saproxylic beetles (Fig. 2a), despite the number of traps placed in this tree species (22 traps) being fewer than in Q. rotundifolia (33 traps), which is the predominant tree species in the reserve (Table 1). In contrast, the highest species richness was linked to Q. rotundifolia for syrphids (Fig. 3a). Regarding vegetation coverage, the absence of any overlap of confidence intervals at the lower abundance value for beetles (n = 1,634 individuals) showed significant differences between sites with low shrub coverage (5–25 %) and high shrub coverage (30–90 %). Sites with low shrub coverage exhibited the highest beetle species richness (Fig. 2b), while ones with high shrub coverage (30–90 %) were the richest for syrphids (lower abundance value of n = 84 individuals) (Fig. 3b).

Fig. 2
figure 2

Individual-based rarefaction curves (with 95 % confidence intervals) for the estimated richness of saproxylic beetles species sampled in a the dominant tree species and b low and high shrub coverage

Full size image
Fig. 3
figure 3

Individual-based rarefaction curves (with 95 % confidence intervals) for the estimated richness of saproxylic syrphid species sampled in a the dominant tree species and b low and high shrub coverage

Full size image

On the other hand, beetles species composition varied significantly according to tree species (ANOSIM R 0.085; p 0.01) but the same did not occur for shrub coverage (ANOSIM R −0.013; p 0.63), or for syrphids (tree species ANOSIM, R 0.01; p 0.38; shrub coverage ANOSIM, R 0.031; p 0.25). Species similarity ranged from 62.5 to 69.2 % between dominant tree species and from 56.2 to 62.6 % between high and low shrub coverage for both saproxylic syrphid and beetle species respectively (Table 3).

Table 3 Percentages of similarity in species composition between (a) the dominant tree species and (b) shrub coverage for both saproxylic beetles and syrphids species. In bold the value statistically supported (ANOSIM R 0.085; p 0.01)
Full size table

In general, the studied structural variables (tree density, dominant tree species and shrub coverage) explained a low proportion of the variance for both beetle (26.8 %) and syrphid (33.3 %) species in the “dehesa”. However, beetles species were more influenced by dominant tree species (F ratio = 1.44; p = 0.006), and for syrphids, shrub coverage was the most relevant variable (F ratio = 1.69; p = 0.04) (Table 4).

Table 4 List of indicator species with their IndVal percentage and p value
Full size table

Indicator species

In general, we found that 10 indicator beetles species were related to tree species and only one to shrub coverage (Table 5). Eight beetle species were related to Q. pyrenaica. Among these indicator species, the xylophagous Cerambyx welensii (Küster, 1846) and the predator Elater ferrugineus Linnaeus, 1758 recorded the highest indicator p values (IndVal 84.7 % p 0.003; IndVal 76.2 % p 0.01 respectively) (Table 5). Moreover, Gnathoncus nannetensis (Marseul, 1862) was detected as an indicator species for Q. pyrenaica (IndVal 81 %, p 0.01), as well as for open shrub coverage (IndVal 67.1 %, p 0.03). In contrast, we found that only Dasytes pauperculus Laporte de Castelnau, 1840 (IndVal 68.5 %, p 0.04) and Ptinus (Ptinus) spitzyi Villa & Villa, 1838 (IndVal 77 %, p 0.04) were indicator species for Q. rotundifolia, even though it was the most abundant tree species in the reserve. As regards syrphids, no indicator species were found for tree species or for shrub coverage.

Table 5 The explained variance of CCA for (a) beetle species was 26.8 % and for (b) syrphids speices explained 33 %
Full size table

Discussion

Previous studies carried out in central Spain (Cabañeros National Park) assessing saproxylic assemblages (Micó et al. 2013; Quinto et al. 2013) have shown similar values of species richness. The high diversity of saproxylic beetles and syrphids (136 and 16 species respectively) from the “dehesa” ecosystem is presumably related to the dynamism of the ecosystem, due to its long history of agrosilvopastoral management, which has generated high landscape heterogeneity and a rich mosaic of habitats (Montserrat and Fillat 1990; Gómez-Sal et al. 1992; Fahrig et al. 2011). It is well known that saproxylic insect diversity is favored by the presence of old and large trees (Lindhe et al. 2005; Stokland et al. 2012), because they offer microhabitats for a wide range of species (Speight 1989; Warren and Key 1991; Fayt et al. 2006; Ricarte et al. 2009), and old managed trees are abundant in the “dehesa” ecosystem (Joffre et al. 1999; Llorente-Pinto 2011). The long-term history of management in the “dehesa” has shaped a scattered tree landscape with a low density (about 20–50 trees/ha), thus allowing the formation of trees with a large diameter (Plieninger et al. 2003). Several studies have demonstrated that the diameter of trees is positively related to species richness (Jonsell et al. 1998; Ranius and Jansson 2000; Grove 2002). Moreover, traditional tree management (pollarding and lighter pruning) promotes the formation of hollows (Sebek et al. 2013), thus increasing microhabitats for saproxylic fauna (Alexander 2003). In the absence of primary forest, all of these factors could determine that these long-term managed landscapes act as saproxylic reservoirs to a large extent, even for red-listed species such as Limoniscus violaceus, which have shown the most western population in Europe here (Micó et al. 2011a) or Mallota dusmeti, which is restricted to Spain and the Tunisian Republic (Claussen and Hauser 1990) and was well represented in the studied area.

With a total of 3.5–4 million hectares in Spain, the “dehesa” ecosystem represents about 7 % of the territory in Castilla and León (Montero and Serrada 2013) and in general, 15 % of the total productive area in Spain (Pérez-Soba et al. 2007). “Dehesa” is a mosaic habitat with high spatial heterogeneity as a result of agricultural practices, livestock husbandry, forestry and hunting activities (Moreno-Marcos et al. 2007; Pulido and Picardo 2010; Campos et al. 2013). In this way, brushwood coverage changes among sites (i.e. areas destined for hunting usually have a higher value for brushwood coverage) and the dominant tree species also varies.

Saproxylic richness was significantly affected by dominant tree species and brushwood coverage, but not in the same way for Coleoptera and Diptera, due to differences in their ecological requirements. Saproxylic beetle assemblages were significantly affected by dominant tree species. Beetle assemblages were more linked to Q. pyrenaica than Q. rotundifolia, and this can be explained by the structural differences between both tree species, for example, in Q pyrenaica the tree hollow volume was greater (0.057 m3) than in Q rotundifolia (0.014 m3). These characteristics derived from the type of management (pollard and pruning) used, which has been stronger in the past for Q. pyrenaica and has generated many larger tree hollows (Fig. 1b), thus enhancing the microhabitat of the saproxylic beetles (Speight 1989; Bouget et al. 2005). In this way, Quinto et al. (in review) have assessed the response of saproxylic beetles to the microenvironmental variables according to trophic guilds, determining that beetle diversity is influenced by higher hollow volumes. Despite Q. rotundifolia exhibiting larger diameters, the tree hollows hosted were smaller (Fig. 1a) and normally filled with water (pers. obs.), thus representing a favorable habitat for many syrphid larvae that are aquatic or semiaquatic (Rotheray and MacGowan 2000; Fayt et al. 2006; Speight 2010). This could explain the higher richness of syrphids related to Q. rotundifolia in this area.

The response of Coleoptera and Diptera richness to shrub coverage was also different. It is well known that many species only exhibit saproxylic habits for part of their life. This is the case of all saproxylic syrphid species, whose larvae live in tree hollows but whose adults need to visit suitable sources of pollen and nectar for feeding (Thompson and Rotheray 1998). Our results show that the main variable affecting syrphids was shrub coverage, and indeed species richness of syrphids was favored by higher shrub coverage (30–90 %), where floristic richness was also higher. In a broad sense, the presence of shrubs also increases saproxylic beetle richness (Økland et al. 1996; Janssen et al. 2009). However, beetle richness was similar at both shrub coverage levels in the studied area, and was only slightly higher in areas with lower shrub coverage (5–25 %) (Fig. 2b). The trophic habits of adults are more diverse in Coleoptera than in Syrphids, which could be why it is possible that the response of all the saproxylic beetle assemblages would seem to be less affected by this variable. Differences in species turnover were only found between the dominant tree species for saproxylic beetles (Table 3). All of these results seem to indicate that, in general, the main factor driving saproxylic diversity is related to the tree species and its management. In this way, we also found that ten beetle species were indicators for tree species, eight for Q. pyrenaica species and only two for Q. rotundifolia (Table 5). Cerambyx welensii (IndVal 84.7 %; p 0.003) mainly attack sick and decrepit trees and, as explained before, Q. pyrenaica was the species most affected in this area by management. Another remarkable indicator species for Q. pyrenaica was the click beetle Elater ferrugineus (IndVal 76.2 %; p 0.01) a near threatened (NT) species in Europe. This species develops in hollows filled with wood mold generated by some species of Dorcus and Protaetia genera, and also preys on the larvae of both genera (Schimmel and Tarnawski 2010). In addition, the larvae of E. ferrugineus have been found preying on larvae of the saproxylophagous Prionychus ater (Murria and Murria 2001), also an indicator species for Q. pyrenaica in the “dehesa” studied. On the other hand, Dasytes pauperculus was recorded as an indicator species for Q. rotundifolia. According to studies by Liberti (2009), this is a rare species associated with oaks that he has collected several times by passive breeding from oak branches, or using WT. This species has also been considered as a good indicator for forest maturity (Liberti pers. comm.).

Conclusions

The “Dehesa” ecosystem, included in the Habitats Directive and also in Natura 2000 Network (www.natura.org/), has been proven to be a rich reservoir of saproxylic insects, due to the maintenance in this ecosystem of large managed trees that offer a wide range of microhabitats for saproxylic insects.

Traditional management of this ecosystem develops spatial heterogeneity and habitat diversification, thus creating a Mediterranean landscape that affects saproxylic diversity positively. A key point is that richness of different taxa responds in a different way to vegetation structure and tree species and could even determine species composition in Coleoptera. Conservation strategies for “dehesa” ecosystems should try to maintain traditional management in order to conserve habitat heterogeneity, and different saproxylic taxa (with different requirements) should also be used to monitor the effect of management on saproxylic diversity.