Carnivores are of high ecological importance and often serve as flagship species for conservation. Yet, the direct and indirect effects of trophic interactions and human disturbance on carnivore communities and the occurrence of their prey in temperate forests remain poorly understood. Here we evaluate the direct and indirect effects of net primary productivity, prey (Muntjac spp) occupancy and human disturbance in shaping carnivore richness and biomass based on camera‐trap data and path analysis. Path analysis indicated that muntjac occupancy and proximity of human settlements were the main factors structuring carnivore communities. Carnivore biomass was directly affected by muntjac occupancy, while carnivore richness increased with distance from the nearest residential site. Muntjac occupancy per se directly depended on net primary productivity. We also identified several indirect effects on carnivore biomass. Increasing distance to the nearest residential site had a positive indirect effect on carnivore biomass that was mediated by increasing carnivore richness. Net primary productivity was positively associated with carnivore biomass indirectly via increasing muntjac occupancy. There were no significant correlations between the occurrence probability of livestock and muntjac. Our results point to the importance of incorporating prey occurrence and human disturbance in modelling biodiversity patterns of carnivore communities. Carnivore conservation management practices should include efforts to stop hunting and to restrict human presence in the core zones of protected areas.

Understanding and managing pollination service is hindered by taxonomic impediments and paucity of data, particularly in the tropics. Herein we apply integrative species delineation and taxonomy to test impacts of land use on the diversity of bee communities within Xishuangbanna (Yunnan, south China), a highly biodiverse tropical region which has undergone extensive land conversion to rubber plantation. 128 Operational Taxonomic Units (OTU) were inferred by an iterative and integrative approach. Bee activity differed significantly across land use samples, although community composition corresponded more to level of vegetation density, when accounting for spatial structure. Species diversity was high in young rubber plantations, although composition overlapped with other species-rich habitats (natural forest edge and river banks), and older plantations (>8 years) showed very low diversity under all measures. Community structures were similar between the natural forest interior and edge, although analysis indicated contrasting drivers of diversity, with clustering in the interior and overdispersion in the forest edge. Further, phylogenetic diversity and derived indices were underestimated when reference data were omitted from analysis. The description of bee communities herein permits more informed choices in land management with respect to ensuring continuation of essential services by bees.