A novel ecological risk assessment method of potentially toxic elements based on soil nematode communities

Abstract

Potentially toxic elements (PTEs) in soil near coal mines threaten soil biota, ecosystem stability, and human health. Soil nematodes, which quickly respond to environmental changes, are reliable biological indicators of PTEs contamination. However, research on establishing a systematic ecological risk assessment model for PTEs contamination using general community indices and nematode-based indices (NBIs) are limited. To address the research gap, we selected 7 cities in Shanxi Province, China, where coal mining is actively conducted. Bayesian kernel machine regression (BKMR) was used to analyze dose–response relationships of PTEs, general community indices, and NBIs. Additionally, based on the general community indices and NBIs, the study developed ecological risk assessment models of PTEs using machine learning techniques. The results showed moderate pollution with significant spatial and seasonal variations, and PTEs such as Pb, Hg, Mn, and Zn concentrations significantly exceeded (0.2 to 6.35 times) than background values. Structure index (SI), nematode channel ratio (NCR), and maturity index (MI) showed negative linear dose–response relationships with PTEs concentration. The ridge regression (Ridge) model performed the best for the nemerow synthetic pollution index (NSPI) and potential ecological risk index (RI) of comprehensive PTEs, while the random forest (RF) model performed the best for the pollution load index (PLI). NCR, MI, and Shannon-Weaver diversity index (H) were the most important factors in determining NSPI (NCR = 21.08 %, MI = 20.78 %, and H = 18.48 %) and RI (NCR = 20.90 %, MI = 20.90 %, and H = 19.50 %). The results highlight that PTEs contamination near coal mine areas was severe, leading to significant disturbances in nematode community structure. Applying general community indices and NBIs, Ridge and RF models can effectively predict the ecological risks of PTEs.