Quantification of sources and input-output pathways of heavy metals in soils from an abandoned mining watershed using Cd isotope tracing and inventory analysis

Quantitative analysis of regional-scale soil heavy metal (HM) sources presents significant challenges. The reliability of the widely used source apportionment model (positive matrix factorization, PMF) remains to be validated. Moreover, PMF are limited in their ability to dynamically assess source-sink changes and their impact on HM accumulation trends based solely on soil concentrations. Therefore, Cd isotopic data (analyzed using both the MixSIAR Bayesian model and a three-End-Member model), the PMF model, and input/output inventories were integrated to jointly quantify the fluxes and sources of HMs in the soils from an abandoned mining watershed. The total annual input flux of soil HMs (2160 g/ha/y) was significantly higher than the output flux (486 g/ha/y), resulting in net annual increase rates of Cd, Cu, and Zn of -0.495, 359, and 43.6 mu g/kg/y, respectively. Irrigation water (1776 g/ha/y) and leaching water (417 g/ha/y) were the main input and output pathways for all HMs. Soil Cr and Ni originate from natural sources, Cd, Cu, and Zn mainly from irrigation water, and Pb from atmospheric deposition. A strong source relationship of Cd among upstream water, compound fertilizers, and surface soils was found. Combination of the three models showed good consistency in quantitative and source specific data for influx of Cd, indicating that irrigation water affected by open pits is the major Cd source (79 %, 62 % and 77 % for PMF, three-End-Member, and MixSIAR-Bayesian models, respectively). This comprehensive analytical framework provides a robust and broadly applicable strategy for quantifying HM sources and elucidating their multi-phase transport dynamics within complex agroecosystems subject to dual geogenic-anthropogenic influences.