Search

Biodiversity patterns are shaped by the combination of dispersal, environment, and stochasticity, but how the influence of these drivers changes in fragmented habitats remains poorly understood. We examined patterns and relationships among total (γ) and site-level (α) diversity, and site-to-site variation in composition (β-diversity) of tree communities in structurally contiguous and fragmented tropical rainforests within a human-modified landscape in India's Western Ghats. First, for the entire landscape, we assessed the extent to which habitat type (fragment or contiguous forest), space and environment explained variation in α-diversity and composition. Next, within fragments and contiguous forest, we assessed the relative contribution of spatial proximity, environmental similarity, and their joint effects in explaining β-diversity. We repeated these assessments with β-diversity values corrected for the confounding effects of α- and γ-diversity using null models (β-deviation). Lower γ-diversity of fragments resulted from both lower α- and β-diversity compared to contiguous forests. However, β-deviation did not differ between contiguous forests and fragments. Fragmented and contiguous forest clearly diverged in floristic composition, which was attributable to β-diversity being driven by differences in elevation and MAP. Within fragmented forest, neither space nor environment explained β-diversity, but β-deviation increased with greater elevational differences. In contiguous forests by comparison, environment alone (mainly elevation) explained the most variation in β-diversity and β-deviation of both species' occurrences and abundances. Spatial gradients in environmental conditions played a larger role than dispersal limitation in shaping diversity and composition of tree communities across forest fragments. Thus, location of remnant patches at different elevations was a key factor underlying site-to-site variability in species abundances across fragments. Understanding the environmental characteristics of remnant forests in human-modified landscapes, combined with knowledge of species-environment relationships across different functional groups, would therefore be important considerations for management and restoration planning in human-modified landscapes., (© 2020 by the Ecological Society of America.)

Edge effects can alter the spatial organization of diversity in fragmented habitats. For tropical forests, however, there has been large variation in the strength and direction of such effects reported by different studies. For long-lived organisms like trees, one reason for inconsistent patterns might be due to most studies having examined patterns of diversity and compositional variation in older life stages that bear the legacy of a forest past. Younger life stages can reveal ongoing processes of assembly, but multi-stage examinations are rare. For seedling, sapling, and adult life stages of trees in a human-modified wet tropical forest in the Western Ghats Biodiversity Hotspot (India), we examined how proximity to forest edges (edge influence) modified the spatial organization of diversity. Specifically, for each life stage we tested whether edge influence led to loss of α- and γ-diversity and decreased β-diversity in this landscape. We found lower α- and γ-diversity closer to forest edges, but only for seedlings. Seedling composition at 90-100 m from forest edges diverged from composition of sites within 60 m, suggesting that edge influence restricted the recruitment of some species to interior sites. In contrast, β-diversity was greater near edges than interior forest for all life stages and most prominently for seedlings. Furthermore, β-diversity at edges was primarily driven by species turnover, suggesting either marked species-environment associations or dispersal limitation. Low turnover at 90-100 m implies that β-diversity arose from stochastic fluctuations in occurrences and abundances of the same species set. Overall, we find that high β-diversity structured spatial patterns of diversity near edges, but recruitment bottlenecks are likely to reduce alpha diversity of forest fragments. Our results also corroborate the need to maintain sufficiently large areas of tropical forest free from edge effects to avoid the loss of interior forest species. To improve landscape-scale diversity of fragmented landscapes, restoration efforts should focus on recovery of species that are unable to regenerate near forest edges., (© 2019 by the Ecological Society of America.)

Environment and human land use both shape forest composition. Abiotic conditions sift tree species from a regional pool via functional traits that influence species' suitability to the local environment. In addition, human land use can modify species distributions and change functional diversity of forests. However, it is unclear how environment and land use simultaneously shape functional diversity of tree communities. Land-use legacies are especially prominent in temperate forest landscapes that have been extensively modified by humans in the last few centuries. Across a 900-ha temperate deciduous forest in the northeastern United States, comprising a mosaic of different-aged stands due to past human land use, we used four key functional traits-maximum height, rooting depth, wood density, and seed mass-to examine how multiple environmental and land-use variables influenced species distributions and functional diversity. We sampled ~40,000 trees >8 cm DBH within 485 plots totaling 137 ha. Species within plots were more functionally similar than expected by chance when we estimated functional diversity using all traits together (multi-trait), and to a lesser degree, with each trait separately. Multi-trait functional diversity was most strongly correlated with distance from the perennial stream, elevation, slope, and forest age. Environmental and land-use predictors varied in their correlation with functional diversities of the four individual traits. Landscape-wide change in abundances of individual species also correlated with both environment and land-use variables, but magnitudes of trait-environment interactions were generally stronger than trait interactions with land use. These findings can be applied for restoration and assisted regeneration of human-modified temperate forests by using traits to predict which tree species would establish well in relation to land-use history, topography, and soil conditions., (© 2018 by the Ecological Society of America.)