Dynamics of soil aggregate-associated organic carbon based on diversity and high biomass-C input under conservation agriculture in a savanna ecosystem in Cambodia

[Display omitted] •Bi-annual rotations restored large macro-aggregates, SOC pools and total N.•Bi-annual rotations reached the highest levels of CMI in the 8–19 mm size aggregate.•Large macroaggregates in the 0–5 cm depth followed: SbCS > RcCS > CsCS.•SbCS due to high biomass-C diversity input reached the highest level of CMI and POXC.•NMR measurement showed higher aliphatic C in NT large macroaggregates than in CT. No-till (NT) cropping systems have the potential to enhance soil aggregation, providing physical protection and soil C sequestration. The existence of discrepancies in the impact of tillage on soil aggregation and soil C sequestration warrants further studies, particularly for different crop rotations. We hypothesized the following: a) NT biannual crop rotations tend to be more effective in restoring large macroaggregation and the concentrations of soil organic C (SOC), total N and permanganate oxidizable C (POXC) associated with macroaggregates than NT systems with a one-year frequency pattern and conventional tillage (CT); b) the continuous biomass-C inputs via crop residues in large macroaggregates under NT tend to increase the proportion of aliphatic C than those under CT. Therefore, the objectives of this study were: (i) to assess changes in the aggregate size distribution and levels of aggregate-associated total SOC, total N and POXC and (ii) to characterize humic acid (HA) using 13C CP-MAS nuclear magnetic resonance (NMR) spectra of 8- to 19-mm soil aggregate size class in a reference vegetation (RV) and in rice-, soybean- and cassava-based cropping systems (RcCS, SbCS and CsCS, respectively) in a clayed Oxisol after tillage and crop rotation management. We evaluated four treatments in each cropping system: 1) CT, and 2) three NT systems in a randomized complete block design with three replicates. Soil aggregate samples were collected at depths of 0–5, 5–10 and 10–20 cm. The conversion of RV to agricultural land influenced the distribution of aggregate size classes, soil aggregation indices and aggregate-associated SOC, total N and POXC in the two surface layers. The formation of large macroaggregates (8–19 mm) dominated the aggregate size distribution with a relatively higher proportion under RV and NT than under CT. Across all soil depths, the proportions of the 8- to 19-mm aggregate size fraction were 59% (NV), 43% and 47% (RcCS), 45% and 53% (SbCS) and 34% and 37% (CsCS) for the CT and NT systems, respectively. Among the three NT sys