Pollinating species are in decline globally, with land use an important driver. However, most of the evidence on which these claims are made is patchy, based on studies with low taxonomic and geographic representativeness. Here, we model the effect of land-use type and intensity on global pollinator biodiversity, using a local-scale database covering 303 studies, 12,170 sites, and 4502 pollinating species. Relative to a primary vegetation baseline, we show that low levels of intensity can have beneficial effects on pollinator biodiversity. Within most anthropogenic land-use types however, increasing intensity is associated with significant reductions, particularly in urban (43% richness and 62% abundance reduction compared to the least intensive urban sites), and pasture (75% abundance reduction) areas. We further show that on cropland, the strongly negative response to intensity is restricted to tropical areas, and that the direction and magnitude of response differs among taxonomic groups. Our findings confirm widespread effects of land-use intensity on pollinators, most significantly in the tropics, where land use is predicted to change rapidly.

Interaction networks are sensitive to elevation gradients through changes in local distribution of interacting partners. Here, we use plant-pollinator interaction network metrics to assess the effect of elevation on flowers and flower-visiting insect assemblages on a sentinel mountain used for monitoring climate change in the flower- and insect-rich Cape Floristic Region. We also use these interaction metrics to explain the effect of environmental factors on the interaction networks. We did this over four vegetation zones <1640m asl, as determined by former botanical studies. Overall, bees were the dominant flower visitors, followed by monkey beetles, and far behind were wasps and flies. The middle elevation zone (650–744 m a.s.l), which is also an ecotone between two distinct botanical zones, had the highest species richness and abundance of interacting plants and insects. Interaction frequency and size of network were also greatest in the middle zone, as were network diversity, generality, and linkage density, while lowest in the peak zone. In sum, there was distinct elevation zoning of flower-visiting insects. The greatest zonal change was between species at the middle compared with peak zone. Large-sized monkey beetles, bees and flies characterized the unique assemblage in the peak zone (1576–1640 m a.s.l.). The insect zonation tracked that of plant assemblages, with air temperature (lapse rate) being the primary driver of bee distribution, with lowest levels in the peak zone. In contrast, beetle distribution was driven mostly by flower assemblages as well as air temperature. In turn, wasp and fly interaction networks were not affected by any of the measured environmental variables. We conclude that increased elevation stress from reduced temperatures, changing abiotic weather conditions (e.g. strong winds at high elevations),and decline in flowering plant composition causes breakdown of interaction networks involving bees and beetles but not that of flies and wasps.

Fire is a major disturbance factor in many terrestrial ecosystems, leading to landscape transformation in fire‐prone areas. Species in mutualistic interactions are often highly sensitive to disturbances like fire events, but the degree and complexity of their responses are unclear. We use bipartite insect–flower interaction networks across a recently burned landscape to explore how plant–pollinator interaction networks respond to a recent major fire event at the landscape level, and where fire refuges were present. We also investigate the effectiveness of these refuges at different elevations (valley to hilltop) for the conservation of displaced flower‐visiting insects during fire events. Then, we explore how the degree of specialization of flower‐visiting insects changes across habitats with different levels of fire impact. We did this in natural areas in the Greater Cape Floristic Region (GCFR) biodiversity hotspot, which is species rich in plants and pollinators. Bees and beetles were the most frequent pollinators in interactions, followed by wasps and flies. Highest interaction activity was in the fire refuges and least in burned areas. Interactions also tracked flower abundance, which was highest in fire refuges in the valley and lowest in burned areas. Interactions consisted mostly of specialized flower visitors, especially in refuge areas. The interaction network and species specialization were lowest in burned areas. However, species common to at least two fire classes showed no significant difference in species specialization. We conclude that flower‐rich fire refuges sustain plant–pollinator interactions, especially those involving specialized species, in fire‐disturbed landscape. This may be an important shelter for specialized pollinator species at the time that the burned landscape goes through regrowth and succession as part of ecosystem recovery process after a major fire event.