Data from: Using beta diversity to inform agricultural policies and conservation actions on Mediterranean farmland

1. Spatial variation in species composition (β-diversity) is an important component of farmland biodiversity, which together with local richness (α-diversity) drives the number of species in a region (γ-diversity). However, β-diversity is seldom used to inform conservation, due to limited understanding of its responses to agricultural management, and lack of clear links between β-diversity changes and conservation outcomes. 2. We explored the value of β-diversity to guide conservation on farmland, by quantifying the contribution of bird α- and β-diversity to γ-diversity variation in low- and high-intensity Mediterranean farmland, before (1995-1997) and after (2010-2012) the Common Agricultural Policy reform of 2003. We further related β-diversity to landscape heterogeneity, and assessed the conservation significance of β-diversity changes. 3. In 1995-1997, bird diversity was highest in low-intensity farmland, where it further increased in 2010-2012 due to a strong positive contribution of α-diversity to γ-diversity. In high-intensity farmland, diversity converged over time to much the same values of low-intensity farmland, with strong positive contributions of both α- and β-diversity. These patterns were largely consistent for total, farmland and species of European conservation concern assemblages, and less so for steppe birds. 4. Beta diversity increased with landscape heterogeneity, particularly related to spatial gradients from agricultural to natural habitats in low-intensity farmland, and from annual to permanent crops (olive groves) in high-intensity farmland. The first gradient was associated with the replacement of steppe birds of high conservation concern by more generalist species, while the second was associated with the replacement between species with lower or higher affinity for woodland/shrubland habitats. 5. Synthesis and applications. In low-intensity farmland, β-diversity was largely stable over time, reflecting a positive conservation outcome related to persistence of landscape heterogeneity patterns required by endangered steppe bird species. In contrast, β-diversity in high-intensity farmland was favoured by increases in landscape heterogeneity driven by olive grove expansion, contributing to enhancement of total bird diversity. Overall, our results stress the value of β-diversity to understand impacts of agricultural policies and conservation actions, but also highlight the need of evaluating β-diversity changes against specific conservation goals.Bird dataTotal number of individuals of each bird species recorded along each 250-m transect per year. Rowname - row identification including the transect code and the bird census year. Transect - transect code; Farmland_area - low- or high-intensity farmland area; Year - bird census year; Date - bird census date; Recorder - recorder's name.bird_data.csvBird categoriesBird species categorization according to their habitat affinities and conservation status. Species - species scientific name; All_species – All species: bird species recorded except aquatic birds because they were excluded from analyses; Farmland - Farmland birds: species associated with all farmland habitat types (e.g., arable fields, permanent crops, hedgerows); Steppe - Steppe birds: a subset of farmland birds that are rare or absent outside open grassland habitats; SPEC1-3 - Species of European Conservation Concern (SPEC1-3).bird_categories.csvHabitat dataRowname - row identification including the transect code and the time period (T0: 1995-1997, T1: 2010-2012). Transect - Transect code; Farmland_area - low- or high-intensity farmland area; Time_period - 1995-1997 or 2010-2012; woodland - percentage cover by holm woods (Quercus rotundifolia) with >30% tree cover and shrub understory, old eucalyptus plantations with understory, and tree lines; open_woodland - percentage cover by holm woods with >30% tree cover and grassy understory, young plantations without understory; shrubland - percentage cover by shrublands without trees or with scattered trees; streams - percentage cover by streams with vegetation (herbs, shrubs or trees); water_bodies - percentage cover by ponds and reservoirs; arable_land_with_scattered_trees - percentage cover by annual crops with scattered trees (