Plant- and Microbial-Derived Soil Organic Carbon: Global Patterns and Divergent Climatic Associations

Soil organic carbon (SOC) pools reflect the net balance of plant- and microbial-derived inputs, transformation, and loss, yet a global-scale assessment of distribution patterns, relative contributions, and hierarchical environmental correlates of microbial necromass carbon (MNC) and plant-derived carbon (PdC) remains incomplete. Based on 1167 site-level observations (1235 pool-level records) across four ecosystem types, we applied variance partitioning, random forest, linear mixed-effects models, and structural equation modeling to quantify global patterns and associative pathways of the two carbon pools. Unadjusted pooled comparisons showed higher MNC concentration and SOC contribution than PdC. However, covariate-adjusted linear mixed-effects models revealed significantly higher PdC. Paired observations showed the same directional trend, although statistical significance was not reached due to limited paired sample size (concentration: p = 0.104; contribution: p = 0.140). The apparent MNC dominance constitutes a Simpson's paradox driven by uneven ecosystem sampling, rather than a consistent within-ecosystem pattern. Instead, the relative contribution of MNC and PdC was ecosystem-dependent: MNC was marginally higher in forests, whereas PdC dominated in croplands and wetlands, with no significant difference in grasslands in paired observations. Edaphic variables were the strongest predictors (independent explanatory power: 61.1% for MNC, 47.4% for PdC), with total nitrogen, C:N ratio, pH, and mean annual temperature as key correlates. Our results reveal a hierarchical associative cascade in which geographic factors covary with climate, which in turn correlates with edaphic properties as the strongest direct predictors. Climate shows a weak but significant negative direct association with MNC only. This synthesis provides a hierarchical correlative framework and biome-specific baseline parameters to support explicit representation of plant- and microbial-mediated SOC pathways in Earth system models.