Intensive ground cover management and landscape simplification affect community-wide floral visitor-plant interactions in olive groves

Abstract

Plant-pollinator interaction networks can provide insights about the impacts of agricultural practices on biodiversity. However, studies on the agricultural impact on these networks most often focus on specific pollinator guilds (i.e., wild bees), hindering our understanding of the effects at the entire insect visitor community level. Here, we evaluated the impact of intensive agriculture and the moderation effects of landscape context on the topology of community-wide plant-visitor interaction networks (381 insect species). We also compared the responses of wild bee-based networks and community-wide plant-floral visitor networks to agricultural intensification to examine whether the former can be a good proxy of the latter. To this aim, we sampled plant-pollinator interaction networks in 24 paired olive farms with contrasting herb cover management (low-intensity management, where ground herb cover is not removed during the flowering season, and intensive management, with herbs persistently removed by herbicides and/or recurrent tillage) across a gradient of landscape complexity in Andalusia (Spain).Results show that the topology of whole floral visitor interaction networks was affected by the herb cover management and the landscape simplification. Furthermore, we found that the effects of the herb cover management were often moderated through the interaction with the landscape context at both floral stand and landscape scale. However, the landscape scale at which these interaction effects occurred varied with the network metric. Non-bee floral visitors were fundamentally redundant for the network topology.We concluded that the implementation of low-intensity management of ground covers in olive groves, together with the enhancement of the flower stands at the local level (e.g., floral strips) and the maintenance and/or restoration of semi-natural habitats at the large landscape scale, would be beneficial to the preservation of healthy and robust floral visitor communities. Our study further supports that wild-bee based interaction networks may adequately reflect the response of the entire networks of plant-floral visitor, at least in this study system.