Ecological Succession Drives the Structural Change of Seed-Rodent Interaction Networks in Fragmented Forests
Xifu Yang,Chuan Yan,Qingjian Zhao,Marcel Holyoak,Miguel A. Fortuna,Jordi Bascompte,Patrick A. Jansen,Zhibin Zhang
DOI: https://doi.org/10.1016/j.foreco.2018.03.023
IF: 3.7
2018-01-01
Forest Ecology and Management
Abstract:While deforestation and fragmentation can cause massive species loss in forest ecosystems, forest regeneration can also drive successional changes in species composition. Although studies have sometimes documented the effects of these compositional changes on interspecific interactions, few studies have investigated changes in the structure of plant-animal networks. We investigated how interaction networks of assemblages of rodents and tree seeds changed with forest fragmentation and succession in a subtropical region. We compared seed-rodent interactions between 14 secondary forest patches that ranged in area from 2 to 58 ha, and from 10 to at least 100 years old, representing a successional gradient. We expected that deforestation and fragmentation would reduce seed production and diversify rodent communities, resulting in higher interaction strengths and connectivity, but weak nestedness (i.e., specialists interact with subsets of the species interaction of generalists). We measured the frequency of rodents eating and removing seeds (interaction strength) in each patch during 3 successive years, using seed tagging and infrared camera trapping, and calculated the properties of the seed-rodent networks. We found that the relative abundances of seeds and rodents changed with stand age not patch size, as did seed-rodent interactions: older patches produced more seeds, contained fewer individuals and species of rodents, and had seed-rodent networks with lower connectance and interaction strength, but higher nestedness. Connectance and interaction strength decreased with metabolic per capita seed availability (as measured by seed energy value); nestedness increased with seed richness, but decreased with rodent abundance. At species level, we found stand age and patch size showed significant effects on seed or rodent abundance of a few species. We also found seed coat thickness and starch contents had significant effects on network metrics. Our results suggest that during succession after deforestation, seed-rodent interactions in these sub-tropical forests change from a state dominated by high seed removal and highly connected seed-rodent networks to a state with more seeds and highly nested networks. From a management perspective of our study region, succession age, not fragment size, and network structure should be paid more attention so as to facilitate the restoration processes of degraded forests. Rodent management should be applied to protect native forest species and exclude incursive ones from farmlands and human residences at early succession stage.