The conventional model of aggregate formation suggests a hierarchy where micro-aggregates with lower porosity and therefore reduced soil organic carbon (SOC) mineralization form inside macro-aggregates. This model has however been questioned for highly weathered Oxisols where inconclusive results regarding the presence of aggregate hierarchy have been obtained to date. We hypothesized that in Oxisols (i) an aggregate hierarchy would be present (ii) the porosity of micro-aggregates would be lower than that of macro-aggregates and (iii) pore geometry of aggregates would influence SOC mineralization. We collected topsoils from Oxisols in northern New South Wales, Australia from which macro-aggregates (>250 μm), micro-aggregates (53-250 μm) and <53 μm fractions were isolated from bulk soil by wet sieving. 3D images of macro- and micro-aggregates were produced using X-ray computed tomography (μCT) showing the presence of micro-aggregates inside macro-aggregates, which confirmed the presence of an aggregate hierarchy in the Oxisols studied. Macro-aggregates were more common and SOC in higher concentrations in forest systems compared with agricultural (the cultivation or pasture) land-uses, but aggregate geometry differed little between the land-uses studied. The porosity of macro-aggregates (4%) was significantly lower than micro-aggregates (5.5%). Despite the differences in pore geometry between macro- and micro-aggregates, SOC mineralized (SOC'min') during a 2-month incubation (at 25°C) was similar in macro- (3% of SOC concentration) and micro-aggregates (2.8% of SOC concentration). We conclude that although aggregate hierarchy exists in these soils and that aggregate geometry did differ between aggregate size classes, there was no evidence to support the porosity exclusion principle and the assumption that SOC is preferentially stabilized within micro-aggregates in these soils.

Radiocarbon is a useful tool for studying carbon dynamics in soil aggregates. The objective of the current study was to determine the mean residence time (MRT) of soil organic carbon (SOC) in macroaggregates and microaggregates under contrasting land uses. Contrasting land uses investigated at Alfisol (equivalent to Dermosol in Australian Soil Classification) sites were native pasture (NP), crop-pasture rotation (CP), and Eucalypt woodland (WL), whereas in Oxisol (Ferrosol in Australian Soil Classification) sites, land uses comprised improved pasture (IP), cropping (CR), and forest (FR). Soil aggregates were separated into macroaggregates (250-2000 μm) and microaggregates (53-250 μm) by wet-sieving, and their 14C signatures were determined by accelerator mass spectrometry (AMS). The 14C activity in both macro- and microaggregates was >100 pMC, indicating the presence of post-bomb carbon in the soil. The mean residence time (MRT) of SOC in macro- and microaggregates (MRTagg) was on average 68 yr longer in the Oxisol compared with that in the Alfisol. The MRTagg in microaggregates was 10 yr longer than that of macroaggregates in the Alfisol. However, the MRTagg in microaggregates was 50 yr shorter compared to macroaggregates in the Oxisol. The MRT of macro- and microaggregates can be separated into active, slow, and stable SOC pools. Among the 3 SOC pools, the MRT of the stable pool is of higher significance in terms of SOC stabilization in soil aggregates because of its longer MRT. However, isolation and direct MRT estimation of the stable SOC pool is difficult. The MRT of active and slow SOC pools associated with macro- and microaggregates was measured using a SOC mineralization experiment to estimate the MRT of the stable SOC pool under contrasting land uses by applying a mass balance criterion. The MRT of active (MRTA) and slow (MRTS) SOC pools in macro- and microaggregates varied between 1-50 days and 13-38 yr, respectively. The estimated MRT of the stable pool carbon (MRTP) in microaggregates was 897 yr longer compared to that of macroaggregates in the Alfisol. However, in the Oxisol, MRTP in microaggregates was 568 yr shorter than that of macroaggregates. Among the land uses, WL in Alfisol and CR in Oxisol had longer MRTagg and MRTP compared to other land uses.