Soil property evolution and profile differentiation in the waste dump of the Hequ Open-pit Coal Mine under different restoration pathways

The inability to disentangle restoration pathway effects from age effects has long constrained mechanistic understanding of soil recovery in post-mining landscapes. To address this methodological challenge, we established a quasi-experimental design at a waste dump in the Hequ Open-pit Coal Mine, comparing four areas that share identical parent material but differ in restoration history: newly abandoned (A1), natural restoration (A2, 5 years), artificial reclamation (A3, 5 years, same origin as A2), and mature reclamation (A4, >10 years). We measured 23 physicochemical and biological indicators across three depths (0–30 cm) and applied ANOVA with Benjamini-Hochberg correction and principal component analysis (PCA). Results revealed that (1) all indicators except mean weight diameter (MWD) and geometric mean diameter (GMD) of water-stable aggregates differed significantly among plots ( p < 0.001). In contrast, both aggregate metrics exhibited strong depth effects—a pattern consistent with the interpretation that structural recovery lags behind fertility recovery. (2) In A3, fertility indicators showed no significant vertical differences, indicating profile homogenization; deep-layer (20–30 cm) microbial biomass carbon was 117.36% higher than in A2, far exceeding the surface increase of 58.81%, thus revealing a distinct deep-layer activation effect. (3) PCA extracted two components explaining 88.13% of the variance: PC1 (78.30%) loaded heavily on fertility and basic physical properties; PC2 (9.82%) was defined by MWD, GMD, and silt content. Notably, this high cumulative variance reflects the full restoration gradient (<1 to >10 years)—a subset PCA of the two 5-year pathways (A2 and A3) explained substantially less variance, indicating that pathway effects alone account for a more modest proportion of soil variation. Nevertheless, the PC1 score gap between A3 and A2 (3.30) exceeded that between A1 and A2 (2.50), indicating substantial separation between the two 5-year pathways. (4) Four physical legacy types were identified based on aggregate size distribution and profile configuration. Collectively, these findings demonstrate that restoration pathways impose multidimensional, legacy-forming imprints on soil properties, offering a quantitative framework for site-specific soil quality assessment and adaptive remediation strategies. Moreover, our quasi-experimental design provides a methodological template for isolating pathway effects from age-related confounding—a persistent challenge in mine rehabilitation research—and the four legacy types offer a diagnostic tool for tailoring restoration interventions to site-specific soil conditions.