Biogeographia – The Journal of Integrative Biogeography Seasonal bird assemblages in Dehesas (substeppic prairies with Quercus suber) of North-Western Sardinia (Italy): A poorly studied landscape of high eco-biogeographic interest

Dehesas (i.e., substeppic prairies with Quercus suber ) represent a poorly studied Mediterranean habitat type of high eco-biogeographic interest. Here, we applied a point-transect method along a yearly cycle (from spring to autumn), to focus on seasonal differences in bird assemblages inhabiting a Dehesas landscape in north-western Sardinia (Italy). We confirmed the presence of Mediterranean species with restricted ranges (e.g. Sturnus unicolor, Sylvia sarda ). At the community level, seasonal patterns show significant differences among mean abundance, with a maximum in autumn and a minimum in summer. The highest values in autumn may be explained by the aggregations of wintering small passerines. Considering the univariate metrics of diversity, the number of species significantly differs among seasons, with the richest community in spring. Shannon-Wiener diversity and Margalef (normalized richness) were higher in spring, and summer compared to autumn. Habitat heterogeneity of these landscape mosaics, driven

This work has been focused on the analysis of seasonal patterns (spring-autumn) in uni-variate metrics of bird diversity of a Dehesas landscape located in north-western Sardinia. Our aims were: (i) to verify interseasonal differences in diversity metrics; (ii) to provide an arrangement of seasonal-related dominance at single species. To our knowledge, these analyses has been applied for the first time in these environmental mosaics of high eco-biogeographic interest. Moreover, since Dehesas in Sardinia are recently under significant anthropogenic pressures (building of new road infrastructures, increase of urban sprawl, location of wind power plants), we think that our data may have implications for conservation of these peculiar Mediterranean habitat types.

Study area
The study area is located in the municipalities of Sindia (Nuoro), Monte Sant'Antonio and Montiferru (Oristano) at an altitude ranging between approximately 400 and 800 m a.s.l. (Monte Sant'Antonio, 808 m a.s.l.; North-Western Sardinia; Fig. 1). In this landscape mosaic, heterogeneous habitat types occur, largely represented by a matrix with pseudosteppe environments of mainly annual plants with Quercus suber dominant ('Dehesas'). At higher altitude, mesophilous oak woods with dominance of Quercus pubescens Willd., and undergrowth with Crataegus monogyna Jacq. and Rubus sp. are present (see Bacchetta et al., 2009). The site is surrounding to the Special Protection Area (EU Directive) SPA ITB023050 "Piana di Semestene, Bonorva, Macomer and Bortigali".
The vegetation is represented by calciferous pastures of Poetea bulbosae with dominant Poa bulbosa L. and many ruderal species (Cichorium intybus L., Dipsacus fullonum L.). The tree layer is represented by scattered individuals of Quercus suber and, more rarely, of the endemic Quercus ichnusae Mossa, Bacch. & Brullo, a species of the cycle of Quercus pubescens, which has been recently described (Mossa et al., 1999). The Dehesa derives from the degradation of the Quercus suber forest, referable to the association Violo dehnhardtii-Quercetum suberis Bacchetta, Bagella, Biondi, Farris, Filigheddu & Mossa, typical of the lower mesomediterranean belt on basalts and other volcanic rocks (Bacchetta et al., 2004). This habitat has been completely altered by intense grazing, coppicing and fire in historical times. It is therefore a partially anthropogenic and cultural environment that nonetheless optimizes the human exploitation with very low energetic input (Guzman, 2016). In the last times, these habitats are under anthropogenic pressure (building of road infrastructures and wind power plants, urban sprawl and so on).
3 Figure 1. Map of the study area (North-Western Sardinia).

Protocol
We carried out a field sampling protocol allowing a quali-quantitative characterization of bird communities during the period May-October 2021, using the point-transect method (Bibby et al., 2000;modified), carrying out a number of sessions, standardized in space (1 km transect each one, with lateral belts of 50 m), geo-referenced on technical maps (scale 1:2,000). We could not carry out the study in the winter period: for this reason, the data refer only to the spring-autumn period.
For each transect/point session, we obtained the number of individual records subdivided for each species. The total values of species-specific records (n) and relative frequencies (fr = n/N, where N is the total number of all species-specific records per season) were obtained for each seasonal session. Species with fr > 0.05 were considered as dominant (Turcek, 1956).
At the community level, the following community univariate metrics were obtained: number of species obtained with the standard transect method (S); normalized species richness (Margalef index): Dm = S -1 / ln(N); Shannon-Wiener diversity index (Shannon & Weaver, 1963): H' = -Σ fr × ln(fr); evenness index: e = H' / lnS (Magurran, 2013, for a review of diversity metrics). Averaged values have been compared using the Kruskall-Wallis test for equal medians. Alpha was set at 0.05 level.
To compare species diversity among seasonal assemblages, we performed the 4 approach for diversity profiles, which the main purpose is to compare diversities in several samples using graphs (Tothmeresz, 1995). The bootstrapping option (giving a 95% confidence interval) is based on 2000 replicates, using PAST software (Hammer et al., 2001).

RESULTS
A total of 4963 records were obtained (1,386 in spring, 2,399 in summer and 1,178 in autumn) belonging to 60 species.
We observed a significant increase in total abundance especially in the autumn period, due to the high availability of resources (mainly for ground foraging species; Leal et al., 2019). The role of Mediterranean agroenvironments for the wintering of many species is widely known (e.g., Tellerìa, 2001;Gillings et al., 2005;Tellerìa et al., 2005). However, even forest environments locally host interesting aggregations of some species in winter (for example, Columba palumbus). Regarding species richness and diversity, we observed highest values in spring and summer seasons, as evidenced by univariate metrics and diversity profiles.
Environmental heterogeneity, together with the presence of 'key structures' can greatly increase the species richness and the individual abundance (Tews et al., 2000;Zamora et al., 2007), also in birds (Farina, 1997;Malavasi et al., 2009;Pickett and Siriwardena, 2011;Santana et al., 2017). In this regard, a large number of available niches allows the coexistence of many species with different ecology (i.e. high richness), and the abundance of resources can explain the large number of individuals (i.e. high abundance; Wiens, 1992), as largely observed in Mediterranean landscapes (e.g. Atauri and de Lucio, 2001).
However, most of the studies are concentrated in continental areas and data for macro-insular environments are still scanty. Our data evidenced a role of Sardinian Dehesas as habitats with a high seasonal dynamism (i.e. high bird richness and diversity in spring and summer and high abundance in autumn-winter).
The area has been poor studied in respect its eco-biogeographic value, excluding a scanty gray local literature (e.g., Schenk et al., 1995). The mosaic of Dehesas is characterized by a high habitat heterogeneity with including oak wood patches, embedded in landscape matrix with pseudo-steppes, pastures and isolated Quercus suber trees. Although this landscape shows a very low anthropization (limited mainly to extensive traditional agropastoral practices) and patchy disturbances (e.g., fires), recently many new infrastructures have been planned (road infrastructures, wind farm plants). Therefore, this lack of knowledge may have conservation implications in the light of these landscape use changes, moving from agro-forest and pastoral ecosystems to urbanization and infrastructuration.
Our study represents only a preliminary arrangement and further research is necessary (winter period has been not considered). In this regard, we stimulated an increase of research effort, around a whole yearly cycle, about structure and dynamics of bird communities inhabiting these still little known Dehesas of high ecological and biogeographical interest.