Based on species occurrence records of museum collections, published literature, and unpublished records shared by mammalian experts, we compiled a distribution database for 59 terrestrial mammals populating the extensively protected Dobrogea Region of Romania. The spatial patterns of mammal distribution and diversity was evaluated and systematic conservation planning applied to identify priority areas for their conservation. The spatial analyses revealed that intensive sampling was not directly correlated to mammal diversity but rather to accessibility for inventory. The spatial prioritisation analysis indicated a relatively aggregated pattern of areas with a high or low conservation value with virtually no connecting corridors between them. The significant overlap between Natura 2000 sites and national protected areas induced an over-optimistic vision of the effectiveness and representativeness of existing Natura 2000 network for species found in Annexes II and IV of the Habitats Directive. These results represent a key step in identifying core areas for the protection of mammal diversity and dispersal corridors for improved connectivity, and to guide future conservation efforts in increasing the effectiveness of the existing protected areas in the context of environmental changes.

Dobrogea, Natura 2000, species distribution, species richness, systematic conservation planning, terrestrial mammals

Terrestrial mammals are well-studied taxa, yet their distribution and conservation status are not fully understood (Crooks et al. 2011). Mammalian population decline accelerates the loss of ecosystem services and poses a substantial threat to species diversity at the community level (Ceballos 2002, Rodrigues et al. 2004). Since mammals display diverse traits and can exploit a wide range of ecological niches, they are also effective focal species for conservation, and their population status might be a proxy for both fragmentation and connectivity across landscapes (Crooks et al. 2011).

A common conservation strategy to prevent the loss of biodiversity is the creation of protected areas (Margules and Pressey 2000, Williams et al. 2002). Protected areas must ensure the long-term persistence and viability of species and should ideally support many rare, threatened, or endemic taxa, particularly those with low mobility and high sensitivity to environmental alterations (Rodrigues et al. 2004, Possingham et al. 2006). However, typically, the effectiveness of protected areas is undermined by poor governance (Eklund et al. 2011, Manolache et al. 2018, Nita et al. 2018) and lack of funding and relevant resources (Sánchez-Fernández et al. 2017).

The Natura 2000 network of protected areas of European importance represents one of the most extensive networks of conservation areas worldwide (Nita et al. 2017). Scientists and policymakers often question the effectiveness of this network due to the Member States allocating fewer funds than needed to implement conservation programs (Nita et al. 2017, Sánchez-Fernández et al. 2017). Natura 2000 is more effective in protecting species listed in Birds Directive because of a better overlap between ancillary conservation investments such as Common Agricultural Policy and biodiversity value (Lung et al. 2014, Maiorano et al. 2015), and because birds are more intensely studied than other vertebrate groups. To be more effective, Natura 2000 network must incorporate potential changes in species distributions (Popescu et al. 2013, Kukkala et al. 2016). Failure to acknowledge changes in species ranges may lead to gaps in protecting species that are sensitive to climate change and other anthropogenic pressures (Araújo et al. 2011).

One of Romania’s legal obligations since joining the European Union in 2007 was to designate Natura 2000 sites in a short time (Ioja et al. 2010). Due to the lack of adequate species and habitat distribution data, regions that already benefited from protection under national laws were preferred for the first phase of the designation process. Consequently, the EU conservation goals were not met, which resulted in the designation of additional protected areas (Ioja et al. 2010, Popescu et al. 2013) and a disproportionate increase of land protected in some regions such as it is the case of Dobrogea (now 63% under protection, 9700 km²). The Natura 2000 network from Dobrogea includes 67 sites (35 Sites of Community Importance - SCI and 32 Special Protection Areas – SPA, most of the SCIs and SPAs spatially overlap). Within Dobrogea, highlands and floodplains gained extensive protection while lowlands occupied by arable lands remained largely unprotected. However, the latter areas are inhabited by endangered species such as the marbled polecat (Vormela peregusna) and the steppe polecat (Mustela eversmanii) (Murariu et al. 2009, 2010).

Due to the diverse landforms, climatic influences, and habitats, Dobrogea harbours a large number of mammal species (Murariu 1996, Murariu et al. 2010). To date, 59 mammal species have been documented in this region, three of which reach the outer limit of their geographic range (the marbled polecat Vormela peregusna, the stoat Mustela erminea, and the common hamster Cricetus cricetus), and two other species have their core range in Dobrogea (the Romanian hamster Mesocricetus newtoni and the Southern birch mouse Sicista nordmanni) (Bunescu 1959, 1961, Popescu and Murariu 2001, Murariu and Munteanu 2005). Of the 59 mammal species, 14 are protected by Habitats Directive. Despite the focus of many Natura 2000 sites within Dobrogea on protecting mammal species, limited and outdated distributional databases are available for individual species, e.g., the Romanian hamster Mesocricetus newtoni (Hamar and Schutowa 1966), the Eurasian beaver Castor fiber (Kiss et al. 2012, Kiss et al. 2014), the European mink Mustela lutreola (Cuzic and Marinov 2004), and the Southern birch mouse Sicista nordmanni (Ausländer and Hellwing 1957). Moreover, with few exceptions, (e.g., Murariu 1996, 2006, Murariu et al. 2009, Murariu et al. 2010) the Dobrogea Region lacks actual regional species distribution data.

One tool supporting management decisions and for investigating species population coverage within protected areas is spatial conservation prioritisation (Pouzols et al. 2014). As part of systematic conservation planning (Margules and Pressey 2000) and accounting for complementarity, spatial prioritisation can be an efficient instrument in identifying spatial priorities and in achieving conservation goals (Pressey et al. 2007) even in broadly protected and underfunded regions such as Dobrogea (Rozylowicz et al. 2017). In this study, we evaluate priority areas for mammal conservation in Dobrogea, Romania and assess the spatial patterns of distribution and diversity of terrestrial mammals by: (1) compiling mammal distribution records from published papers, museum records, and unpublished data, (2) analyzing spatial patterns of distribution data, and (3) using systematic conservation planning in identifying high priority areas for conservation of terrestrial mammal listed in Annexes II and IV of Habitats Directive within the regional Natura 2000 network.

By using an updated distribution of terrestrial mammals, we identified high priority areas for protecting mammal diversity to guide future conservation efforts in an extensively protected Romanian region. In the broader context of systematic conservation planning, the prioritisation analysis is a useful tool to identify key areas for biodiversity conservation, e.g., where species are more likely to survive (Ferrier and Wintle 2009, Wilson et al. 2009, Kukkala and Moilanen 2013).

The number of reported occurrences in Dobrogea varied among species. The largest number of records (20%) are for the wood mouse (Apodemus sylvaticus), mostly because they are widespread within the region, have a higher population abundance, and are evidently. The wood mouse may be easily misidentified as a yellow-necked mouse (Apodemus flavicollis) (Bartolommei et al. 2016), but it still retains the first rank because of their higher population in Dobrogea (Popescu and Murariu 2001).

The lowest number of records in Dobrogea is recorded for the hazel dormouse (Muscardinus avellanarius) and the bank vole (Myodes glareolus), with only one record per each species. Two other widespread species, but with an uncharacteristically low number of reported presences are the red fox (Vulpes vulpes) and the roe deer (Capreolus capreolus), possibly because they are common species, with minor interest for biologists. The research effort for species sampling was focused on selected areas: Măcin Mountains National Park, Danube Delta Biosphere Reserve, and the North Dobrogea Plateau. Here, we recorded a higher than expected number of species occurrences per grid cell, mostly because the long-term protection status attracted faunistic inventory projects over time. Typically, the sampling bias is higher in protected areas because they attract more conservation funds leading to greater efforts for biodiversity research (e.g., Botts et al. 2011). This process describes most of the biodiversity spatial databases (Lobo et al. 2007). A higher than expected sampling effort also was evident near cities, major roads, and research facilities, which are easily accessible to researchers (e.g., Valu lui Traian Biological Research Station - see Figure 4). We noted a lack of research interest in central and southern Dobrogea, where most of the agricultural landscape is located, and only small patches of natural habitats remain as wildlife refuges (Rey et al. 2007). However, agricultural landscapes are essential for many species protected by Habitats Directive, such as Vormela peregusna, Mustela eversmanii, Spermophilus citellus, Mesocricetus newtoni, and Cricetus cricetus (Popescu and Murariu 2001, Murariu and Munteanu 2005, Murariu et al. 2009, 2010). Those identified species have restrictive ecological requirements, and hence, the researchers should focus on increasing sampling effort in these neglected areas to draft appropriate conservation plans.

We found that high species richness did not match all the hotspots of sampling efforts, such as in Valu lui Traian Biological Research Station and Fântânița-Murfatlar (location 1 in Figure 4) and Letea Forest Natural Reserve in the Danube Delta (location 3 in Figure 4). This validates the finding that intensive sampling was not directly correlated with mammal diversity, but rather ease of access to the regions (Santos et al. 2017). To better understand the patterns of species richness at the regional level, we expanded the resolution to 10 × 10 km, but the spatial pattern did not change between the two mapping resolutions. In both cases, the highest diversity (35 species at 5 × 5 km resolution, see Figure 5); 45 mammal species at 10 × 10 km resolution, see Figure 6), is found in the northern part of Dobrogea, overlapping North Dobrogea Plateau Natura 2000 site, where habitat heterogeneity is high (Rey et al. 2007).

The spatial turnover index (S₂) suggests that areas with lower species richness are dissimilar compared to their neighbours. Notably, we observed some affinities of particular species towards low species richness areas (Lennon et al. 2001), e.g., species dependent on steppe or agricultural landscape (Popescu and Murariu 2001, Murariu and Munteanu 2005). Typically, the spatial turnover tends to be correlated with species richness (Gaston et al. 2007), but in our study, the variation in turnover is determined by the rarity of the species which then tend to have narrower habitat niches and drive turnover patterns more than widespread species. By analysing species richness and turnover index maps, we found low congruency between the Natura 2000 sites and areas with high species richness and areas occupied by species with a narrower range. Species with narrow ranges (e.g. Lutra lutra, Memedemin et al. 2017) are often underrepresented in protected areas, potentially resulting in suboptimal effectiveness of the Natura 2000 network in protecting such species in Dobrogea, despite the large area protected under conservation.

Our results regarding the identification of high-priority areas in Dobrogea for mammal species listed in Annexes II and IV of the Habitats Directive highlight a relatively aggregated pattern of the grid cells with high conservation value in the northeastern and northern Dobrogea Region, where there are extensive wetlands and forests (i.e., Danube Delta, Măcin Mountains National Park, and North Dobrogea Plateau) (Rey et al. 2007). Additionally, we identified small isolated high-priority areas in agricultural landscapes of southern and central Dobrogea, where biodiversity-friendly agricultural practices should be considered as a conservation method. Distribution of top spatial conservation priorities demonstrated a lack of connectivity between Natura 2000 sites with high conservation values from the northern part of Dobrogea, isolated priority areas in the center of the region, as well as those in the southern part of the region. The distribution of high-priority areas for conservation suggests the necessity of addressing the lack of connectivity, as non-priority areas are essential for the dispersal of species (Christie and Knowles 2015).

The significant overlap between Natura 2000 sites and the other protected areas statutes leads to misunderstandings in law enforcement and an over-optimistic vision of their effectiveness (Ioja et al. 2010). As an example, species whose distributions are limited to the EU Steppic Biogeographic Region or reach the boundaries of their geographic range in Dobrogea tend to be under-represented (Popescu and Murariu 2001, Murariu and Munteanu 2005) as in the case of reptiles and amphibians (Popescu et al. 2013). In Dobrogea, isolation of protected areas leads to low connectivity between habitat patches, which then need to be addressed in future conservation planning and protected area management plans. The lack of research in agricultural landscapes may potentially lead to the populational decline of certain species by not being aware of their distribution and by using flawed species range data (Grand et al. 2007).

The absence of buffer zones and corridors between Natura 2000 sites and small isolated protected areas (the area of the smallest Natura 2000 site in Dobrogea is 0.11 km²), are not beneficial in maintaining viable populations, causing the isolation of species with low mobility and specific habitat requirements (Christie and Knowles 2015). Establishing corridors between Natura 2000 sites, especially in the central and southern part of Dobrogea increases connectivity and promotes species dispersal.

Our study is limited by the lack of viable and current distribution data. Most records do not identify geographical coordinates, but localities or toponymies. This makes the niche modelling at a fine scale a challenge. Furthermore, elusive species such as Mesocricetus newtoni, Sicista nordmanni, and Vormela peregusna, are data deficient, and the lack of records (false absences) may influence the results of the analysis. Similarly, misidentification of sibling species may lead to over- or under- estimation of their range. Notably, a study analysing the distribution of amphibians in Dobrogea (Székely et al. 2009) indicates similar issues regarding biased and incomplete distribution data due to the lack of comprehensive surveys of areas with difficult accesses. Another similarity is that some amphibians (e.g., Bombina bombina, Bufo viridis, Hyla arborea), as well as some mammals (e.g., Capreolus capreolus, Apodemus sylvaticus) are considered widespread and highly detectable, while amphibians such as Pelobates fuscus and Pelobates syriacus, are cryptic and elusive species and therefore, have low detectability and incomplete distributions (Székely et al. 2013), and that includes species such as Vormela peregusna, Sicista nordmanni or Mesocricetus newtoni. However, biased data lead to more priority areas to protect fewer species (Grand et al. 2007), which is not a shortcoming. Furthermore, Rodrigues et al. (2011) concluded that decision-based on incomplete taxonomic and/or phylogenetic data (such as misidentified sibling species) are robust, and the researcher can safely make use of the best available systematic data.

Future research may focus on identifying buffer zones around Natura 2000 sites to minimise potential negative impacts, particularly in Natura 2000 sites that are adjacent to agricultural areas. From this assessment, we envisage further mapping of corridor networks between small isolated protected areas in southern and central Dobrogea. New research should focus on systematic surveys of agricultural landscapes in central and southern Dobrogea, where vegetation patches remain as refugees for some species listed in Annexes II and IV of the Habitats Directive (Vormela peregusna, Mustela eversmanii, Spermophilus citellus, Mesocricetus newtoni, Cricetus cricetus, and Sicista nordmanni).

IVM, VDP, AN, SM, and LR were supported by a grant from the Romanian National Authority for Scientific Research, PN-III-P4-ID-PCE-2016-0483. Editing and proofreading by Edward F. Rozylowicz streamlined the flow of the manuscript. We thank Maria Patroescu, Tiberiu Sahlean, and an anonymous reviewer for their comments that improved the manuscript. We are grateful to Dumitru Murariu, Nastase Radulet, Bogdan Bejenariu, Catalin Stanciu, Dragos Mantoiu, Ionut Stefan Iorgu, Daniyar Memedemin, Razvan Popescu Mirceni, Silviu Petrovan, Marius Skolka, Razvan Zaharia, Alexandra Telea, Dragos Balasoiu, Mihaela Stanescu, and Elena Buhaciuc-Ionita for sharing their data on species occurrences.