National-scale patterns of soil organic matter dynamics across agricultural and pasture systems of Mexico

Soil organic matter (SOM) plays a central role in soil fertility, climate regulation, and ecosystem resilience, yet its spatial variability across large and heterogeneous territories remains poorly quantified. This study presents a national-scale assessment of soil organic carbon (OC) as a proxy for SOM dynamics across agricultural lands and pastures in Mexico, covering approximately 51 million hectares. Using a harmonized dataset of ~7,000 georeferenced soil observations (0–30 cm depth) and a Digital Soil Mapping framework based on Quantile Regression Forests, we generated spatially explicit predictions of OC at 250 m resolution together with uncertainty estimates. Model performance for OC was moderate to high at the national scale (r ≈ 0.6–0.7; IOA ≈ 0.6–0.7), capturing major pedoclimatic and management-driven gradients. Predicted OC contents ranged from ~1.1–1.4% in arid and semi-arid northern regions to >2.0% in the Yucatán Peninsula, reflecting strong controls of climate, texture, and parent material. Across land-use systems, pasture systems and moisture-dependent agriculture exhibited significantly higher OC levels, whereas irrigated agriculture showed consistently lower OC, highlighting the influence of management intensity and disturbance on SOM stocks. Beyond mapping spatial patterns, the integration of uncertainty metrics allows identification of areas where SOM estimates are more reliable and regions where additional data are required. This work provides the first uncertainty-informed national baseline of SOM-related carbon for Mexico, supporting region-specific soil management strategies and contributing to monitoring, reporting, and verification efforts related to soil carbon and soil health under contrasting pedoclimatic environments.