Data from: Patterns and predictors of β-diversity in the fragmented Brazilian Atlantic forest: a multiscale analysis of forest specialist and generalist birds

1. Biodiversity maintenance in human-altered landscapes (HALs) depends on the species turnover among localities, but the patterns and determinants of β-diversity in HALs are poorly known. In fact, declines, increases, and neutral shifts in β-diversity have all been documented, depending on the landscape, ecological group and spatial scale of analysis. 2. We shed some light on this controversy by assessing the patterns and predictors of bird β-diversity across multiple spatial scales considering forest specialist and habitat generalist bird assemblages. 3. We surveyed birds from 144 point counts in 36 different forest sites across two landscapes with different amount of forest cover in the Brazilian Atlantic forest. We analysed β-diversity among points, among sites, and between landscapes with multiplicative diversity partitioning of Hill numbers. We tested whether β-diversity among points was related to within-site variations in vegetation structure, and if β-diversity among sites was related to site location and/or to differences among sites in vegetation structure and landscape composition (i.e. percent forest and pasture cover surrounding each site). 4. β-diversity between landscapes was lower than among sites and among points in both bird assemblages. In forest specialist birds, the landscape with less forest cover showed the highest β-diversity among sites (bird differentiation among sites), but generalist birds showed the opposite pattern. At the local scale, however, the less forested landscape showed the lowest β-diversity among points (bird homogenisation within sites), independently of the bird assemblage. β-diversity among points was weakly related to vegetation structure, but higher β-diversity values were recorded among sites that were more isolated from each other, and among sites with higher differences in landscape composition, particularly in the less forested landscape. 5. Our findings indicate that patterns of bird β-diversity vary across scales and are strongly related to landscape composition. Bird assemblages are shaped by both environmental filtering and dispersal limitation, particularly in less forested landscapes. Conservation and management strategies should therefore prevent deforestation in this biodiversity hotspot.

1. 人类改造景观（human-altered landscapes, HALs）中的生物多样性维持依赖于不同局地间的物种更替，但目前学界对人类改造景观中β多样性（β-diversity）的格局及其驱动因子仍知之甚少。事实上，已有研究报道了β多样性的下降、上升以及中性变化三种模式，具体结果取决于研究的景观类型、生态类群以及空间尺度。2. 本研究针对森林专食性鸟类群落与栖息地广适性鸟类群落，通过多空间尺度下的鸟类β多样性格局及其预测因子分析，为厘清这一争议提供了新的见解。3. 研究团队在巴西大西洋森林的两类森林覆盖度不同的景观中，于36个不同森林样地内开展了144次定点鸟类计数调查。研究采用希尔数（Hill numbers）的乘法多样性分解方法，分别分析了样点间、样地间以及景观间的β多样性。我们检验了样点间β多样性是否与样地内植被结构的变异相关，同时也检验了样地间β多样性是否与样地位置、样地间植被结构差异以及景观组成（即各样地周边的森林与牧场占比）相关。4. 两类鸟类群落的景观间β多样性均低于样地间与样点间β多样性。在森林专食性鸟类中，森林覆盖度更低的景观其样地间β多样性（即样地间鸟类群落分化程度）更高，而广适性鸟类则呈现相反的格局。但在局地尺度上，无论鸟类类群如何，森林覆盖度更低的景观其样点间β多样性均更低（即样地内鸟类群落同质化程度更高）。样点间β多样性与植被结构的相关性较弱，而在彼此更为隔离的样地间，以及景观组成差异更大的样地间，β多样性值更高，这一现象在森林覆盖度更低的景观中尤为显著。5. 本研究结果表明，鸟类β多样性的格局随空间尺度发生变化，且与景观组成密切相关。鸟类群落的构建同时受环境过滤与扩散限制的影响，在森林覆盖度更低的景观中这一效应尤为突出。因此，该生物多样性热点地区的保护与管理策略应优先防止森林砍伐。