Non-linear effects of landscape on pollination service and plant species richness in a peri-urban territory with urban and agricultural land use

International audiencePollination contributes to both human food security and the reproduction of the majority of wild plant species, but pollinators are facing a rapid decline, a major cause of which is habitat conversion and degradation due to human activities. Urbanization is one of the major types of habitat conversion, but its influence on pollination has been surprisingly mixed, ranging from markedly negative to strongly positive effects. One hypothesis proposed to explain these discrepancies is that pollinator responses to urbanization are highly dependent on the nonurban control habitat, with negative effects when the controls are natural or semi-natural areas but positive when they are intensive agricultural areas. It was also proposed that the pollination response along an agricultural-tourban gradient is non-linear, with maximum pollination observed at an intermediate level of urbanization due to increased environmental heterogeneity. To test these two hypotheses, we selected a group of 38 sites in a periurban area near Paris, France, using a semi-stratified sampling strategy that ensured that all three of the urban, agricultural and semi-natural gradients were maximized. We then estimated pollination using two approaches: we evaluated the pollination success of Lotus corniculatus, a strictly entomogamous self-sterile plant species pollinated mainly by bees, and we measured the species richness of entomogamous and non-entomogamous plants, the difference in their response being expected to relate to the pollination service provided by the overall pollinator community. We found that in our study area, pollination success of L. corniculatus responds positively to the agricultural to urban gradient but not to the semi-natural to urban gradient. The diversity of both entomogamous and non-entomogamous plants is highest at sites surrounded by intermediate proportions of urban and agricultural areas. In addition, high proportions of urban areas have a negative effect on the diversity of non-entomogamous but not entomogamous plant species, suggesting that pollinators are able to partially buffer entomogamous plant species against the negative effect of urban development. Our results show the importance of urban areas in pollination conservation plans and demonstrate that the interaction between different anthropogenic land-use is an important factor for understanding pollination