Spatio-temporal variation in plant-pollinator interactions: a multilayer network approach

Abstract

[eng] Despite the dynamic nature of communities, most research typically treated interaction networks as static entities, and only a few analysed the spatial and the temporal scales simultaneously. Here, we used spatial and temporal multilayer networks to explore the persistence of species and interactions in space and time, as well as the variation of species role (centrality) according to biotic factors. We further investigated, for the first time, the spatio-temporal variation in multilayer network structure. Species exhibited substantial turnover across time mainly due to differences in species phenology. In contrast, interaction turnover was more pronounced across habitats, which seems to be a common strategy of pollinators to adjust interactions spatially to different ecological scenarios. Plant species were comparatively more important to the cohesiveness of spatial than temporal networks, whereas the centrality of pollinators correlated between scales. The importance of plant species fluctuated temporally due mainly to changes in flower density, whereas that of pollinators fluctuated in space and time according to their relative abundance. Module composition was highly unstable in time. Species were highly capable of changing partners and module affiliation across both scales. We conclude that plant-pollinator interactions are highly dynamic in space and time, and that there are differences between plants and pollinators in their use of resources across habitats and subseasons with implications for the understanding of functional connectivity and multilayer structure.