Protected areas and insect conservation: questioning the effectiveness of Natura 2000 network for saproxylic beetles in Italy

Authors


  • Editor: Nathalie Pettorelli

    Associate Editor: Jonathan Davies

Correspondence

Manuela D'Amen, CNR-IBAF, Via Salaria km 29 300 Monterotondo Scalo, Italy; Corpo Forestale dello Stato, Centro Nazionale Biodiversità Forestale ‘Bosco Fontana’, Via C. Ederle 16/a, Verona, Italy.

Email: manuela.damen@ibaf.cnr.it

Abstract

Up to now, global conservation priorities are far from incorporating megadiverse invertebrate taxa. Thus, an important emerging field in biological conservation is how we might manage landscape to preserve insects. In this study, we analyze the efficacy of Italian reserve network for protecting multiple saproxylic beetles, considering both nationally designated areas and Natura 2000 sites. We selected 150 species inhabiting the Italian territory from the European Red List for saproxylic beetles, on the basis of distribution data availability. For each species, a vulnerability score was assigned according to their Red List status, and the species' distributions data were used to perform an irreplaceability analysis. Our analyses show that conservation targets based on geographic range extent are achieved for only 7% of the considered species. We find that 13 species are not represented in any protected area: among these, two click beetle species (Elateridae) are listed in the International Union for Conservation of Nature threatened categories (i.e. Ampedus quadrisignatus EN and Ampedus brunnicornis VU). Our analyses on protected area effectiveness for the conservation of saproxylic beetles showed that nationally designated protected areas are more irreplaceable than a random selection of cells. Surprisingly, the addition of Natura 2000 sites did not improve the species representation. Moreover, these reserves include sites that are not more irreplaceable than a random selection of cells. We identify some currently unprotected areas that protection could prevent from future extinctions and ensure a favorable conservation status of saproxylic beetles. In particular, we find an important stronghold for beetle conservation, which obtained a high irreplaceability score, in the Adige river basin. We recommend the designation of new reserves in this area to complement the existing network and to help guarantee invertebrate saproxylic fauna protection.

Introduction

The International Union for Conservation of Nature (IUCN) Red List is widely recognized as the most comprehensive and authoritative resource detailing the global conservation risk of plants and animals worldwide, relying on a number of objective criteria (e.g. Butchart et al., 2010). Nevertheless, to date there is a lack of taxonomical coverage on the IUCN Red List: the conservation status of only about 0.3% of invertebrates has been assessed for the Red List, whereas 42% of vertebrates taxa have been evaluated (Clausnitzer et al., 2009). Wild and native insects provide vital ecosystem services, that is, dung burial, pest control, pollination and wildlife nutrition, for which a value of about $60 billion year−1 was estimated in the US (Losey & Vaughan, 2006). While the majority of insects are still un-described (Hammond, 1992), insect species are being lost at an enormous rate (Samways, 2005), and few scientific data are available about the processes driving their extinction (Dunn, 2005). Some principles are emerging from recent research on how we might manage landscape for insect conservation (Samways, 2007). Among the six principles for insect conservation recognized by Samways (2007), the first one is to maintain reserves as source habitats, particularly for specialists species that cannot survive in transformed landscapes (Fermon et al., 2000; Maeto, Sato & Miyata, 2002; Lien & Yuan, 2003).

The in situ conservation of viable populations is a fundamental requirement for the maintenance of biodiversity (Rodrigues et al., 2004). Reserves alone are not adequate for nature conservation, but they are the cornerstone on which protection strategies are built (Margules & Pressey, 2000). Up to now, global conservation priorities are far from incorporating megadiverse invertebrate taxa for a number of reasons related to lack of awareness of their importance and to the deficit in distribution data and ecology knowledge (e.g. Mace, 2000; Brummitt & Lughadha, 2003; Cardoso et al., 2011). Although the use of well-known taxa as surrogates for conservation planning is somewhat encouraging (Rodrigues & Brooks, 2007), some regional data show little overlap between priority areas for arthropods and those for plants and terrestrial vertebrate taxa (Dobson et al., 1997). These findings emphasize the need to use primary invertebrate data in global conservation prioritization as they become available (Brooks et al., 2006).

Focusing on Europe, member states developed a continental initiative aimed at the identification of specific local sites for conservation, regulated by the European Community Directives (i.e. Birds Protection Directive, 79/409/EEC; Habitat Protection Directive 92/43/EEC). A network of nature protection areas, the Natura 2000, has been established under the 1992 Habitats Directive with the aim of maintaining or restoring to a ‘favorable conservation status’ the natural habitats throughout Europe. Besides, Natura 2000 is an important opportunity for member states to increase their existing national networks of conservation areas. Red Lists are an important tool to assess the status of species. They can complement the sphere of action of the Habitats Directive as they address all species in a specific taxonomic group, not just those protected by the European Union nature legislation. Recently, the first assessment of the Red List status of saproxylic beetles in Europe has been compiled, and a selection of 436 species present in Europe was evaluated (Nieto & Alexander, 2010). Saproxylic insects have been identified as a highly threatened group (Read, 2000; Alexander, 2004). The outputs from the European Red List can be applied at the national and regional scale to prioritize sites for beetles' diversity.

Planning exercises of protected areas rarely accounted for insect conservation needs (Martín-Pínera, 2001; Cabeza et al., 2010) or tested a posteriori the effectiveness of the existing reserves for their conservation (Romo, Munguira & García-Barros, 2007; Sánchez-Fernández et al., 2008; Wilhere, Goering & Wang, 2008). In particular, to our knowledge, only few studies examined the requirements of saproxylic beetles (e.g. Stokland, 1997; Gaspar et al., 2011). The present study focus on geographic priorities for saproxylic beetles conservation in Italy, where several studies have addressed the effectiveness of protected areas for terrestrial vertebrates only (Maiorano, Falcucci & Boitani, 2006; Maiorano et al., 2007; D'Amen et al., 2011; Bombi et al., 2011a). These authors found that existing protected areas are often insufficient to conserve current patterns of biodiversity in Italy. Similar results were pointed out for other European countries (Hopkinson, Evans & Gregory, 2000; Dimitrakopoulos, Memtsas & Troumbis, 2004; Araújo, Lobo & Moreno, 2007). We determine how well represented the saproxylic beetles species are in the current Italian system of natural protected areas, considering separately two components: the national reserve system and the Natura 2000 sites. Then, we apply a systematic conservation planning approach (Margules & Pressey, 2000) to identify priority areas for saproxylic beetles under different scenarios that is, without restrictions in the selection of cells, and departing from the already established natural protected areas or Natura 2000. Doing this, we identified 64 geographical strongholds for the conservation of endangered insect species. These areas contain unique taxocoenoses represented by a combination of threatened species, thus their protection is fundamental to fulfill the reserve goals.

Materials and methods

We considered all species listed in the European Red List of Saproxylic Beetles (Nieto & Alexander, 2010) inhabiting the Italian territory and whose distribution data were available in the CKmap 5.3.8 software (Stoch, 2000–2005). This database reports species occurrence within the Universal Transverse Mercator grid (10 × 10 km), composed of 3557 cells and represents the largest, most authoritative and updated resource of zoological knowledge in Italy, being composed of more than 500 000 records of c. 10 000 terrestrial and freshwater species (Ruffo & Stoch, 2005, 2006). In particular, we used distribution data for 150 species belonging to seven coleopteran families: Bostrichidae, Buprestidae, Cerambycidae, Cucujidae, Elateridae, Lucanidae, Cetoniidae (Appendix S1).

We derived data on the location of existing protected areas in Italy from the web site of the national Ministry for the Environment (ftp://ftp.scn.minambiente.it/Cartografie). These maps comprise the boundaries of nationally designed protected areas (hereafter ‘NPAs’) and sites included in the European Natura 2000 network. NPAs are reserves created by national or local administrations before 2004, and have a total surface of c. 29 400 km2 (slightly less than 10% of the country's surface). The Natura 2000 network is composed of 2885 sites, largely overlapping with NPAs. The establishment of Natura 2000 sites increased the protected surface by 34 700 km2 so that the overall protected areas cover a total surface of 64 100 km2 in Italy (21% of the country' surface). We define Natura 2000 areas not overlapping NPAs as ‘EPAs’, while we call the overall protected areas as ‘OPAs’ (Gambino & Negrini, 2002). Because data on species and reserves are available at different resolutions (species data are represented on grid cells, while reserves are represented as polygons), thresholds should be used for deciding when reserves of varying size and position are considered present or absent from a particular grid cell (Hopkinson et al., 2000). We tested different thresholds (from 0 to 100% with intervals of 10) and chose the value that ensured the selection of a number of cells with a total surface equal to the total surface of Italian protected areas (considering independently NPAs and OPAs) (D'Amen et al., 2011). As a result, we set a threshold of 40% coverage to determine whether protected areas should be considered present in the 10-km grid cells. This value is consistent with results from Araújo (2004), who observed a plateau in species accumulation curves using a 40% coverage threshold to consider a grid cell as reserved.

We performed two separate but complementary analyses to assess the level of protection for saproxylic beetles in Italy: gap and irreplaceability analyses. Both the analyses require the identification of a representation target. Representation is the amount of area of a species' range included in protected areas. Following Rodrigues et al. (2004), we used a species-specific representation target depending on distribution extent for each species. In particular, species with a narrow distribution (less than 10 cells) should have 100% of the area protected, while species with the widest distribution (291 cells) should have at least 10% of the area protected. For all the other species, representation targets were calculated by interpolating the extreme range size targets using a linear regression on the log-transformed number of occupied cells (Rodrigues et al., 2004).

Gap analysis seeks to identify species not adequately represented within protected areas. A total gap occurs when a species is not represented in any protected area, while a partial gap occurs when a species representation target is only partially met. To measure relative conservation importance of different cells, we performed also an irreplaceability analysis using the C-Plan Systematic Conservation Planning System, Version 4 (Pressey et al., 2009). Irreplaceability is a measure of the potential contribution of a site to a particular conservation target (Pressey, Johnson & Wilson, 1994). We estimated the weighted summed irreplaceability for each cell, which is calculated by adding all the feature (species) irreplaceabilities of all features in that site. A high value of summed irreplaceability indicates that the site is crucial for achieving conservation targets for many species, while values close to zero indicate that the site is not important for any features. In the calculation of summed irreplaceability, species were weighted for vulnerability (reservation priority) according to their IUCN Red List status (values of vulnerability from 1 to 5 for categories of increasing risk). We define ‘strongholds’ for saproxylic beetles conservation the 5% most irreplaceable cells.

To investigate the effectiveness of the Italian system of reserves (OPAs) and the contribution of the Natura 2000 component (EPAs) to the preexisting network (NPA) for beetle conservation, we compared the summed weighted irreplaceability values of cells in conservation networks to those expected in cells randomly selected regardless of their conservation status (Araújo et al., 2007; D'Amen et al., 2011). We calculated the probability that the observed mean value for protected cells differed from a random mean value comparing the observed mean value with the distribution formed by 5000 random selections of a number of grid cells equal to the number of protected cells (Gotelli, 2000). We thus tested whether OPAs, NPAs and EPAs have higher irreplaceability than the remaining grid cells.

Results

The ranges of about two-thirds of the considered species are covered by NPAs for 20% of their extent (Fig. 1a). The percentage of protected range increases up to 40% by considering the entire network (Fig. 1b). Overall, 12.67% of species are totally not represented in NPAs, and the number of totally gap species decreases to 8.67% when including Natura 2000 sites (Table 1). Following the predefined targets, we found that 6% of beetle species meet their conservation requirements by virtue of the NPAs network. The contribution of EPAs to the initial NPAs produces only a small increase of the number of species that reach their targets (7.33%). The majority of species do not achieve their representation targets, both considering the national components and the OPAs (Table 1). Considering species conservation status (Nieto & Alexander, 2010), we found total-gap species in all Red List categories. In particular, two total-gap species (family Elateridae) are listed by IUCN in the threatened categories: Ampedus brunnicornis (VU) and Ampedus quadrisignatus (EN). Among the six species grouped in the VU category, none fulfills the representation target (Table 1).

Figure 1.

Species frequency distribution as a function of the percentage of range protected considering (a) NPAs (nationally designed protected areas) and (b) OPAs (overall protected areas). At the top of the bars, we report the absolute number of species.

Table 1. Number of totally and partially gap species in each IUCN category, and percentage of target met, both considering the national component alone (NPAs) and the overall network (OPAs)
 n° speciesNPAsOPAs
% total gap% partial gap% target met% total gap% partial gap% target met
  1. IUCN, International Union for Conservation of Nature; NPAs, nationally designed protected areas; OPAs, overall protected areas; DD, data deficient; LC, low concern; NT, near threatened; VU, vulnerable; EN, endangered.
Overall15012.6781.3368.67847.33
IUCN categoryDD16 (10.67%)18.7556.252518.755031.25
LC97 (64.67%)9.2890.7205.1584.5410.31
NT24 (16%)8.3379.1712.58.3379.1712.5
VU6 (4%)16.6683.33016.6683.330
EN7 (4.67%)14.2857.1428.5714.2842.8542.85

Considering no cells as protected, summed irreplaceability analysis identified the areas of highest values for the conservation of saproxylic beetles, which are located throughout the entire peninsula (Fig. 2a). The identification of the 5% most irreplaceable cells lets us define 64 strongholds of different size scattered in both mountain and coastal/plain areas (Fig. 2a). The largest stronghold is located in the Adige river basin; other medium-sized areas are scattered in Eastern Alps, northern and southern Apennines and in the central Tyrrhenian coast of the peninsula. When the existing NPAs are included in the analysis, many of the highly irreplaceable areas mainly locate on the Apennines lose their importance. The inclusion of EPAs to the irreplaceability analysis produced a positive effect in some areas, especially for the Alps protection, however, the pattern of high-important unprotected sites is similar to when no protected cells are included. Only about one-fourth (27%) of the strongholds for saproxylic beetles conservation are protected by NPAs (Fig. 2b), while the OPA system overall safeguards less than 40% of strongholds (Fig. 2c).

Figure 2.

Irreplaceability patterns considering (a) no cell as reserved, (b) NPAs (nationally designed protected areas) only and (c) OPAs (overall protected areas). Strongholds for saproxylic beetles conservation (the 5% most irreplaceable cells) are white-bordered.

The comparison of the mean values of summed weighted irreplaceability of cells in conservation networks (Mobs) with values calculated from 5000 sets of randomly selected grid cells (Msim) reveals that the entire network of Italian reserves (OPAs) protects sites with greater irreplaceability than that expected by chance (Mobs = 1.11, P(Mobs≥Msim) < 0.001). We also tested the OPAs component separately: the NPAs protect sites with greater irreplaceability than expected by chance (Mobs = 1.66, P(Mobs≥Msim) < 0.001), but surprisingly the Natura 2000 component (EPAs) covers sites that are not more irreplaceable for saproxylic beetles conservation than a random selection of cells (Mobs = 0.65, P(Mobs≥Msim) = 0.321).

Discussion

This model study represents a synthetic view of the conservation regime in relation to a network of existing protected areas of a European country for saproxylic beetles and evaluates the contribution of different components of the network. According to our knowledge, although the effectiveness of networks of protected areas has been the topic of many studies, the contribution of reserve networks to the conservation of an important group such as saproxylic insects has never been tested. Our analysis clearly indicates that the existing protected areas are inadequate for assuring the conservation of beetles in the region.

Important lack of protection

The number of total-gap species can be considered quite high [8.7% of all species considered in the analysis compared to the 12% found by Rodrigues et al. (2004) at a global level], and the number of partial-gap species is extremely high. This is a clear indication that the existing reserve network is not adequate to protect the Italian saproxylic diversity. The efficacy of the existing reserves could be substantially increased including further sites, and the strongholds we indicate should be considered as a guide to strategically place new reserves. More specifically, we find the most important stronghold for saproxylic beetles conservation represented by the Adige catchment. The most plausible hypothesis is based on the diffuse presence of large patch of ‘old-growth forest’ (cf. Blasi et al., 2010) left free to evolve in the last 100 years (from the end of the first World War), for the lack of economic convenience in cutting, with a general lack of forest roads, especially along the inaccessible slopes between 1200 and 1800 m a.s.l. Typically, the sylvatic mosaic of this marginal relict or ‘protective forest’ is formed by biostatic or near-to-collapsing eco-units (cf. Oldeman, 1990) with highest amount of dead wood [coarse woody debris (CWD), Peterken, 1996] and large living senescent trees (in many cases with a diameter-at-breast-height of more than 100 cm, cf. Mason, 1977, 1989). It is not a coincidence that Speight (1989), in his well-known report published by the European Council ‘Saproxylic Invertebrates and their conservation’, includes many forest falling in the southern part of Adige catchment, in the ‘list of the European forest identified of importance for their fauna of saproxylic invertebrates’. Among these, one of the most studied is the forest ‘Monti Lessini, Ala di Trento’ (Abies/Fagus/Picea, cf. Chemini, Daccordi & Mason, 1986; Mason, 1988). For conservation planners, this old-growth forest represents one very important hot spot for the ‘saproxylic’ biodiversity to be included in the Italian conservation network. Secondary strongholds, which are only partially protected by the existing reserves, are located, according to the list of Speight (1989, Annex 2), in northern Apennines and along the central Tyrrhenian coast, while further small strongholds are scattered all across Italy.

According to our analyses, the species in greatest need of conservation are: A. brunnicornis and A. quadrisignatus, which are threatened (VU and EN, respectively) and completely unrepresented in the Italian protected areas at the same time. These two species are very localized and deserve the creation of ad hoc reserves, specifically designed for protecting them. In particular, one site in the Po river plain and one along the Ionian Sea coast would give an important contribution to the protection of these species. Second, two other threatened Elaterid species [Ischnodes sanguinicollis (VU) and Podeonius acuticornis (EN)] meet less than one-half of their respective conservation targets (12.50 and 50.00%, respectively), both considering NPAs and OPAs.

The efficacy of the Natura 2000 network

The contribution of the Natura 2000 network to biodiversity conservation has been explored by several works (Dimitrakopoulos et al., 2004; Maiorano et al., 2007; D'Amen et al., 2011), but up to now the invertebrates have been considered in few assessments (Abellán et al., 2007; Sánchez-Fernández et al., 2008; Hernández-Manrique et al., 2012). The Natura 2000 network has increased the percentage of protected Italian territory to more than 20%. Despite this increase, our results show that the network does not provide the necessary coverage for saproxylic gap species. Surprisingly, our analyses showed that the Natura 2000 sites are distributed in areas not more irreplaceable than chance for the considered insects. This result highlights the huge lack of consideration for insects in the process of conservation planning at the European level. Many protected areas were established opportunistically under political, economic or aesthetic constraints (Scott et al., 2001; Maiorano et al., 2007). Invertebrates are rarely considered in the selection of protected areas. The selection of surrogates has many limits (relations were weak), and therefore the lack of congruency between invertebrate taxa supports the use of a multi-taxa approach for the incorporation of invertebrates into conservation plans (Panzer & Schwartz, 1998; Lovel et al., 2007).

Conservation priorities in Italy

There are very few concordances between the outcomes of irreplaceability analyses on the Italian territory based on vertebrates and invertebrates taxa. The previous prioritization studies carried out in Italy (Maiorano et al., 2006, 2007; D'Amen et al., 2011), which dealt with vertebrate species, found that the areas of highest conservation value are located in Sardinia and in the lowlands of the Po river valley and the north-eastern continental Italy; further areas were positioned in Sicily, on the Apennines and along the Tyrrhenian coast. These areas only partially match with the strongholds we identify for saproxylic beetles. Although the northern Apennine and the central Tyrrhenian coast are rather important for vertebrates and insects at the same time, the most important stronghold for beetles (i.e. the Adige river basin) does not represent a conservation priority for vertebrates. On the contrary, Sardinia, which emerged as particularly important from all the studies on vertebrates, has only two small strongholds for saproxylics. This incongruence suggests that cross-taxon surrogacy (i.e. in this case vertebrates used as surrogates of insects) should be used cautiously and encourage the adoption of multi-taxonomical approaches to conservation planning (Grand et al., 2004; Lovel et al., 2007).

Caveats

In the face of unprecedented global biodiversity loss, conservation planning must balance between refining and deepening knowledge versus acting on current information to preserve species and communities. Our dataset is biased by the non-homogeneous field effort across the country, which cannot be quantified, and by gaps in the knowledge on the distribution of the species considered. We acknowledge that this factor might potentially affect the outcome of the results and conservation plans. On the other hand, biodiversity data will never be complete because exhaustive surveys over broad regions are rarely feasible and take longer than conservation planners can afford to wait. The use of suboptimal datasets may approximate the levels of representativeness and efficiency achieved with complete data (Grand et al., 2007). Moreover, investing in additional data might not be the most cost-effective approach to conservation when implementation is gradual and accompanied by ongoing habitat loss (Grantham et al., 2008). If the delay is too long, it can sometimes be more effective by just using a readily available habitat map (Grantham et al., 2009). It is important to improve the choice of conservation priorities based on such data and guide efforts to collect additional data.

The method we used for defining the species ranges incorporates a certain degree of uncertainty. Considering the species ranges as the occupied cells in a reference grid, as we did, can drive to underestimate the real species distribution (Bombi, Luiselli & D'Amen, 2011b). Omission errors in range maps can influence prioritization analyses in complex ways (see Bombi et al., 2011b). The optimal approach would be adopting species distribution models for producing range maps and feeding prioritizations with the produced models (Bombi et al., 2011b). Nevertheless, in the case of saproxylic beetles, distribution models cannot be adopted because of the low number of records for several species. In such a situation, using a grid-based approach to distribution mapping can be considered a suboptimal but indispensable solution.

In the real world, uncertainty is part of any decision process (Araújo, 2004). When comprehensive distribution data are available only from coarse resolution atlases, a common problem is to select a threshold for assigning reserves to atlases cells (Araújo, 1999, 2004). Potentially, the choice of different thresholds can affect estimates of species representation in reserves. We selected a single threshold, relying on a reasonable criterion (the number which allowed to select a total reserved surface equal to the real reserves coverage in the Italian territory), and this threshold also matched with previous studies (Araújo, 2004). Anyway, we should acknowledge that this could have produced biases in our outcomes, because small protected areas were omitted. A follow-up for this study would be to define realistic boundaries for protected areas at finer resolutions coupled with more detailed distribution data. This improvement could be particularly relevant especially for the areas that proved to be conservation strongholds.

Conclusion and future perspectives

The spatial resolution of available distribution data and, therefore, of our analyses (i.e. 10 km) is too coarse for several practical applications at a local scale. The solution to this problem would require a huge effort to increment the quality and resolution of the data, or a less accurate, but faster analytical approach. In particular, the adoption of cross-scale modeling, which allows to produce species distribution maps at higher resolution than the initial data (see Bombi & D'Amen, 2011). This approach would provide practical tools for the definition of specific sites for conservation measures on the one hand and reduce the uncertainty in species representation linked to low-resolution distribution maps on the other (Bombi, Salvi & Bologna, 2012). In this perspective, our outcomes represent a first step within a multilevel prioritization approach, which utilizes low-resolution studies for targeting further analyses at a local scale.

To date, many studies worldwide showed that range modifications are expected for many species under future climatic conditions, and these changes will affect the degree to which species are represented in the current network of protected area (e.g. Araújo, 2004; Hannah et al., 2007; Coetzee et al., 2009). Spatially fixed reserves may not necessarily host the same species in the future, with specialists not adapted to move through the fragmented landscape likely to be the first to suffer. This scenario has been observed also for insect species. For example in British butterflies, 30 of 35 species have not tracked recent climate change, and this is attributable to lack of suitable habitat (Hill et al., 2002). Our results showed that under current conditions the existing network does not represent neither the entirety of saproxylic beetle diversity nor its geographic pattern. This inadequacy might aggravate in the long term, if range shifts and reductions because of climate change will occur, leading at decreasing species representation in protected areas. Further studies are required for anticipating how climate change will impact on saproxylic beetles and optimizing conservation strategies in the long run.

In an era of change, the role of the wider landscape surrounding protected areas is particularly crucial for biological adaptation. Thus, tackling saproxylic diversity loss at national scale should also involve conservation efforts in managed forests. The value of dead wood and the protection of decaying wood habitat is now listed as an environmental concern in forestry policy and procedures in North America, Australia and many European countries (Grove, 2001; Humphrey et al., 2002; MCPFE, 2003). Examples of intervention to favor dead wood microhabitats are the retention and/or creation of snags, stumps and CWD, inducing decay in mature trees and the protection of veteran trees within the landscape, in order to maintain a balanced age structure of trees (Cavalli & Mason, 2003). Also the spatial arrangement of these dead wood element is important, as many of these species are poor dispersers (Radu, 2007). The success of saproxylic conservation in the future years will depend in our capability to balance conflicts among carbon stocks and biodiversity features and human economy needs.

Acknowledgments

We are really thankful to all those provided helpful suggestions and assistance in the english revision. We are grateful to three anonymous reviewers for their detailed and careful comments. M.D. and L.Z. were supported in this study by the LIFE Project ManFor CBD (LIFE09 ENV/IT/000078).

Ancillary