Microbial transformation and mineral adsorption control chemical evolution of soil organic matter during semi-arid ecosystem development

The chemical composition and diversity of soil organic matter influence soil organic carbon (SOC) persistence and climate responses, yet their evolution during soil development in drylands remains unclear. We characterized the chemical composition and diversity of bulk SOC and water-extractable organic matter (WEOM) across a 3-million-year-old semi-arid volcanic soil chronosequence, spanning a steep silt and clay concentration gradient. Soils were sampled from beneath pine or juniper tree canopies and inter-canopy spaces covered by grasses and/or shrubs. As soil developed, those beneath pine or juniper canopies became enriched in aromatic C due to a synergistic effect of intensified microbial decomposition of non-aromatic plant material and preferential adsorption of aromatics from WEOM onto minerals in the silt and clay fraction. Consequently, WEOM was depleted in aromatics with soil development. Both bulk SOC and WEOM also showed increasing proportions of microbial carbon, facilitated by higher silt and clay concentrations that mitigated water scarcity and provided suitable pore spaces for microbial proliferation. Inter-canopy soils showed minimal trends with soil development, ascribed to the higher litter quality than tree litter. The WEOM molecular α-diversity remained stable as the influences from microbial transformation and mineral adsorption counteracted each other. However, β-diversity, reflecting compositional dissimilarity of bulk SOC or WEOM across samples at each soil developmental stage, declined as soil developed. This chemical convergence resulted from dominant microbial and mineral interactions overriding vegetation and other influences. Our findings suggest the dual role of silt and clay in controlling SOM chemistry in dryland soils, enhancing accrual of both microbial C and the aromatic portion of the plant-derived C. These new insights can inform process-based models to better describe soil organic C dynamics and persistence in drylands.