Abundance-Stratified Responses to Island Geometry During Time-Driven Microbial Succession

Understanding the determinism of community succession trajectories and the factors that shape them is not only a central question in ecology, but it also bears directly on microbial management in food processing, industrial materials, and ecosystem management. Within this framework, potential differences in dispersal arising from habitat spatial structure, and how these differences vary among taxa of differing abundance, provide an important lens for examining the spatial heterogeneity of community succession. We constructed six artificial agar islands with identical area but different perimeters and boundary complexities. Using a shared microbial source and identical culture conditions, we sampled communities destructively at 16 time points from 1 to 96 h, with three independent replicates for each island × time combination (n = 288), and characterized bacterial community dynamics by 16S rRNA gene amplicon sequencing. Communities underwent continuous reorganization over 96 h, with an overall decline in ASV richness and increases in the Simpson index and Pielou evenness at later stages. Whole-community PERMANOVA showed that Time explained 51.27% of compositional variation (p < 0.001), whereas Island explained only 1.45% and was not significant (p = 0.0734); the Time × Island interaction was also not significant. Island perimeter showed no significant monotonic relationship with three successional-rate metrics or six trajectory metrics. Communities consistently retained a long-tailed abundance structure. Among 747 ASVs with detectable temporal signals, the island trajectory concordance was significantly negatively correlated with global ASV abundance rank (ρ = −0.534, p < 0.001), indicating that cross-island temporal concordance generally weakened as abundance decreased. After abundance stratification, Island was not significant for Dominant taxa (mean relative abundance ≥ 1%, p = 0.1532), but was significant for Intermediate taxa (≥0.1% and <1%, p = 0.0106) and Rare taxa (<0.1%, p < 0.001); the Time × Island interaction was significant only for the Rare stratum (p = 0.0014). Overall, time-related change constituted the major axis of whole-community variation, whereas boundary length did not generate a stable whole-community successional gradient. High-abundance members showed stronger temporal synchrony across islands, while low-abundance taxa showed greater cross-island trajectory heterogeneity. These results support a closer correspondence between shared time-related processes and synchronous changes in high-abundance core members; whether island-associated differences among low-abundance members were caused by boundary-mediated dispersal or resource exchange requires direct process measurements.