Contrasting effects of land‐use changes on herbivory and pollination networks

• Main Authors: Naoto Shinohara, Kei Uchida, Takehito Yoshida
• Format: Article
• Language: English
• Published: Wiley 2019-12-01
• Series: Ecology and Evolution
• Subjects: agroecosystems; biodiversity; herbivore; interaction network; pollination; seminatural grassland
• Online Access: https://doi.org/10.1002/ece3.5814

Abstract Land‐use changes, one of the greatest threats to global biodiversity, can cause underappreciated effects on ecosystems by altering the structures of interspecific interaction networks. These effects have typically been explored by evaluating interaction networks composed of a single type of interaction. Therefore, it remains unclear whether the different types of interaction networks sharing the same species respond to the same land‐use changes in a similar manner. To compare the responses of herbivory and pollination networks to land‐use changes, we investigated both types of interaction networks in seminatural grasslands categorized into three types of agricultural land‐use (abandoned, extensively managed, and intensively managed) in a Japanese agricultural landscape. We quantified the structures of the interaction networks using several indices (connectance, evenness, diversity, generality, network specialization, and robustness) and compared them among different land‐use types. We conducted piecewise SEM to differentiate the direct and indirect effects of land‐use changes on the network structures. Although both land‐use changes (abandonment and intensification) led to reduced plant and insect species richness, the structures of herbivory and pollination networks showed different responses to the land‐use changes. There was a marked contrast in network generality; while, herbivore species were less generalized (i.e., having fewer host plant species) in fields with land‐use intensification, pollinator species were less generalized in abandoned fields. Furthermore, the mechanisms behind the changes in interaction networks were also different between pollination and herbivory networks. The change in herbivory network generality was induced by the decrease in plant species richness, whereas the change in pollination network generality was mainly induced by the effect independent of changes in species richness and composition, which possibly reflect the less number of flowers in shaded environment. The present study demonstrates that agricultural land‐use changes affect herbivory and pollination networks in contrasting ways and suggests the importance of assessing multiple types of interaction networks for biodiversity conservation in plant–insect systems. Our results also highlight the underappreciated importance of maintaining habitats with an intermediate intensity of land‐use.