Landscape connection and patch complexity explain plant community similarity in sandy grasslands better than habitat amount and landscape heterogeneity via network analysis

Abstract

Based on the island biogeography theory and the habitat amount hypothesis, species richness and alpha diversity increase with an increase in connectivity and habitat amount, and high environmental heterogeneity usually maintains high plant alpha diversity. However, the mechanism of landscape pattern on plant community similarity, another dimension of plant diversity, are still poorly understood. We sampled 420 plant communities at 28 sampling sites comprising mobile, semi-fixed, and fixed sandy dunes, and lowland grasslands. The plant community similarity index was modelled as a response variable in stepwise linear regression models to select the primary landscape metrics. A general linear model was then used to identify the contribution of selected landscape metrics to plant network parameters. The results demonstrate that patch shape complexity, rather than habitat amount or area, was positively associated with plant community similarity (p < 0.05), as well as connection or aggregation, and not landscape heterogeneity (and/or landscape diversity). Such influences are closely linked to plant community network characteristics, which are determined by the sandy fragmented landscape. These findings suggest that shape complexity and edge traits of patches influence plant community similarity; consequently, at the landscape level, patch connection and aggregation influence plant community networks, while habitat amounts or landscape heterogeneity were less affected. These various underlying mechanisms need to be considered in management plans and reserve design for sandy grassland ecosystems to protect grasslands in sandy landscapes.