Optimizing long-term soil organic carbon (SOC) sequestration requires a mechanistic understanding of spatial variability and drivers of stable carbon pools. This study quantified the responses of stable mineral-associated (MAOM) and labile particulate organic matter (POM) to contrasting management systems in a commercial agricultural setting, identified key environmental drivers, and developed predictive spatial models. The study was conducted on Chernozem soils (291 ha) in the Russian forest-steppe ecotone, comparing a conventional tillage (CT) system (sunflower-wheat) with two no-tillage (NT) systems (soybean-sunflower-wheat) established for 5 and 8 years. NT systems received moderate nitrogen inputs (21–34 kg N ha⁻¹ annually), whereas CT received none. Ninety soil samples (0–10 cm) were analyzed using particle-size fractionation, microbial approach, X-ray diffraction and statistical modeling. NT had significantly increased MAOM (22–27 %) while decreasing POM (9–23 %) compared to CT. MAOM increased with MBC and dolomite content, but decreased with quartz content in silt-clay fraction (16 %, 8 % and 11 % of explained variance, respectively), underscoring microbial-mineral stabilization pathways. In contrast, POM variability was poorly predicted by soil microbial and mineral properties. Gradient boosting machine models integrating remote sensing indices with soil properties (SOC, MBC, quartz) achieved high predictive accuracy for both MAOM (R² = 0.77) and POM (R² = 0.73), enabling farm-scale mapping of these pools. Our results demonstrate that short-term NT, coupled with soybean inclusion and N fertilization can enhance SOC stability within a relatively short timeframe through microbial mediation. The integration of soil and remote sensing data offers a powerful framework for targeted SOC management and landscape-scale sequestration strategies in temperate agroecosystems, with potential relevance for other regions.
扫码关注我们
求助内容:
应助结果提醒方式:
