Comprehensive evaluation system for amendment selection: A case study of degraded black soil improvement

Soil degradation threatens food security and climate-related soil functions. Various amendments have been developed to restore degraded soils, while amendment selection remains largely empirical because multifunctional responses and strong within-field heterogeneity hinder fair comparison among options. Here, we conducted a three-year field trial on the restoration of degraded Mollisols due to compaction and fertility loss in Northeast China and developed a screening-evaluation system for amendment selection. The system integrates expert-based pre-screening with multi-indicator weighting and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to rank amendments based on three functions: relieving compaction, stabilizing aggregate stability, and increasing crop productivity. Crucially, a root-distribution-based spatial weighting scheme was incorporated to address ridge-furrow and depth heterogeneity, enabling a root-relevant synthesis of bulk density (BD), penetration resistance (PR), and soil porosity (SP). Six organic amendment options were evaluated: woody peat (WP), biochar (B), sheep manure (OM), and three sheep-manure-based microbial composites inoculated with Bacillus subtilis (BOM), Trichoderma viride (TOM), or Aspergillus niger (AOM), alongside an unamended control (CK). Over the three-year period, WP and BOM consistently improved 0-20 cm topsoil structure, reducing PR by 26-40% relative to CK, and increasing SP and the proportion of water-stable macroaggregates (WR_0.25). WP showed stronger increases in soil organic matter (SOM) and available potassium (AK), whereas BOM promoted root biomass (RB) and the root-to-shoot ratio (RS). Both treatments achieved higher compaction improvement scores (CIS) and aggregate improvement scores (AIS), with WP maintaining the highest soil comprehensive improvement score (SCIS) throughout the study. The maize yields were positively associated with the SCIS, CIS, and AIS over three consecutive years (P < 0.05), although yield was not used in the index construction, which supports the decision relevance of the framework under field conditions. This field-validated, transparent workflow provides a practical tool for optimizing amendment strategies and supporting soil restoration planning in spatially heterogeneous Mollisols.