Effects of landscape features on the structure and function of bird seed dispersal networks in fragmented forests

Abstract

Forest fragmentation has resulted in reduced natural habitats and alterations in landscape structure. Although the effects of landscape features on the structure of seed dispersal networks are well known, the effects on functional consequences of these networks remain unclear. In this study, we compared the structures of seed dispersal networks of three fragmented habitats (high-, medium-, and low-fragmented habitats). We used an integrative index (seed dispersal potential, SDP) to evaluate the network functions of the three habitats. We also studied how the structure and function of seed dispersal networks respond to different landscape features. Approximately 23, 26, and 22 bird species were recorded foraging for 9, 25, and 22 plant species, resulting in 315, 665, and 546 foraging events in high-, medium-, and low-fragmentation habitats, respectively. The structure and function of seed dispersal networks of the three fragmented habitats significantly differ. For instance, the network in a highly fragmented habitat exhibits high connectance and nestedness, whereas the network in a low-fragmentation habitat exhibits high modularity and specialization. Furthermore, the network function (the SDP value for frugivorous species) varies for each habitat. Frugivorous species exhibit a lower seed dispersal potential in highly fragmented habitats, as indicated by their low total SDP values. Most importantly, the structure and function of the seed dispersal network are differently affected by landscape variables. Increasing habitat fragmentation leads to a reduction in the structure of seed dispersal networks. However, medium levels of fragmentation can enhance the network function. Overall, our study highlights the importance of considering both structural and functional aspects of mutualistic networks to understand the landscape effects of habitat fragmentation. The SDP index provides a comprehensive approach to study the functions of mutualistic networks and is a useful tool for future forest management.