A novel method for achieving ecological indicator based on vertical soil bacterial communities coupled with machine...

Global industrialization has resulted in severe contamination of soil with heavy metals (HMs). Nevertheless, it is unclear if it affects the depth-resolved bacterial communities. Herein, we collected soil samples at different depths from a typical HM-contaminated site and used amplicon sequencing to determine the differences in depth-resolved bacterial communities and to assess the thresholds and ecological impacts of HMs. Results revealed that HM levels reduced markedly with soil depth. The bacteria in upper soil exhibited higher community diversity and a more complex and stable ecological network structure. As depth increased, the proportion of negative interactions gradually elevated, indicating more competitive interspecies behavior. Threshold analyses based on machine learning revealed that arsenic (As) and copper (Cu) exhibited nonlinear impacts on ecosystems. Cu demonstrated a low-threshold effect, with its ecological consequences manifested at extremely low concentrations. Our results highlighted the utility of microbial monitoring in assessing the adverse effects of HMs on soil health to support environmental management and ecological restoration.