Purpose River sandbars are critical river-land interfaces that regulate hydrological exchange, sustain biodiversity, and maintain energy flow and material cycling. However, systematic evidence on how soil microbial communities in these dynamic ecosystems respond to land-use change through a cascade of taxonomic diversity, functional potential, and network topology is still lacking. To address this gap, we selected five typical land-use types (bare land, grassland, natural forest, artificial forest, and farmland) on three sandbars in the middle reaches of the Yangtze River. Methods Soil bacterial and fungal communities were profiled using high-throughput amplicon sequencing. Bacterial functional potential was predicted using FAPROTAX, and fungal functional guilds (trophic modes) were assigned using FUN-Guild. Microbial association networks were inferred from OTU abundance profiles using Pearson correlation. To assess stability, we derived sample-specific subnetworks by mapping each sample's OTU onto the aggregated meta-network, and quantified positive connectivity and robustness using node and edge removal simulations. Results Our results indicate that bacterial communities showed significant differences among land-use types in the Shannon and Simpson indices, whereas fungal communities differed significantly in the Chao index and phylogenetic diversity (PD). In addition, both bacterial and fungal beta-diversity and community composition were structured considerably by land use, with redundancy analysis (RDA) explaining 56.17% of the variance for bacteria and 39.84% for fungi. Soil pH emerged as the dominant environmental driver. Land use mainly affected the bacterial FAPROTAX-predicted sulfur-cycling function, which differed significantly among land-use types. By contrast, carbon, nitrogen, and iron-cycling functions were largely stable at the group level. The most pronounced response was observed in co-occurrence networks: disturbed ecosystems exhibited a simplified association topology, consistent with reduced positive connectivity, lower network complexity, and decreased simulated robustness. Conclusion In summary, this study reveals distinct response patterns of soil microbiomes to land-use change in river sandbar ecosystems. Compared with diversity metrics, co-occurrence networks were more sensitive to land-use disturbance, consistent with strong environmental filtering along major edaphic gradients and compositional convergence in disturbed sites. Importantly, these co-occurrence patterns represent statistical associations and may partially reflect shared responses to abiotic drivers rather than direct biotic interactions. These findings highlight that restoration strategies should move beyond species diversity metrics and instead prioritize the rebuilding of native vegetation and the stabilization of soil properties to support the recovery of microbial association network structure and its inferred robustness.
Copyright 2002 - 2023 Wuhan Botanical Garden,Chinese Academy Of
Sciences
Email: wbgoffice@wbgcas.cn ICP: 05004779-1
