31-year contrasting agricultural managements affect the distribution of organic carbon in aggregate-sized fractions of a Mollisol

Evaluation of soil organic carbon (SOC) dynamics is often limited by the complexity of soil matrix. Quantitative information on the distribution of SOC within aggregate hierarchy will help elucidate the carbon flow in soil matrix. However, this knowledge still needs to be documented. Soils were sampled from a surface Mollisol with plots under 100 years of continuous cropping, 31 years of simulated overgrazing to severely degraded bareland, and grassland restoration from cropped soil. A combined density and chemical fractionation procedure within water-stable aggregate was utilized to quantify the distribution of OC after long-term different land use patterns. Results showed that grassland significantly increased total SOC and mean aggregate associated OC compared to initial soil in 1985 with total SOC (g kg(−1) soil) from 46.1 to 31.7 and mean aggregate associated OC (g kg(−1) aggregate) from 31.6 to 44.7. Converting cropland to grassland also enhanced the formation of macroaggregates (>0.25 mm) (from 34.7% to 52.2%) and increased the OC concentrations in density and humic fractions by 48.3%-75.9% within aggregates. But the proportions of OC in density and humic fractions to SOC only increased in macroaggregates in grassland. Alternatively, converting cropland to bareland caused substantial depletion of total SOC, macroaggregates and their associated OC concentrations. The SOC (g kg(−1) soil) and mean aggregate associated OC (g kg(−1) aggregate) significantly decreased from 31.7 to 25.7 and from 31.6 to 26.2, respectively. While the OC concentration of density and humic fractions within aggregates in bareland did not show significant decreases. Principal component analysis demonstrated that the soils were developed by contrasting land use changes, with the grassland soil being more associated with labile OC fractions within macroaggregats and bareland soil more associated with recalcitrant OC fractions within microaggregates and silt-clay units. These findings highlighted the favorable preservation of plant-derived carbon within soil aggregates, particularly in the labile OC fractions within macroaggregates under high plant inputs with 31 years of grassland conversion. For the cropland and bareland soils without organic inputs, more OC was stabilized within fine aggregates via organo-mineral interactions, tending to be more recalcitrant.