LAUSR

Abstract Soil structure is not static but undergoes continuous changes due to a wide range of biotic and abiotic drivers such as bioturbation and the mechanical disturbance by tillage. This continuous alteration of soil structure beyond the pure swelling and shrinking of some stable structure is what we refer to as soil structure dynamics. It has important consequences for carbon turnover in soil as it controls how quickly soil organic matter gets occluded from or exposed to mineralization. So far there are hardly any direct observations of the rate at which soil pores are formed and destroyed. Here we employ are recently introduced labeling approach for soil structure that measures how quickly the locations of small garnet particles get randomized in soil as a measure for soil structure dynamics. We investigate the effect of desiccation crack dynamics on pore space attributes in general and soils structure turnover in particular using X-ray microtomography for repeated wetting-drying cycles. This is explored for three different soils with a range of soil organic matter content, clay content and different clay mineralogy that were sieved to a certain aggregate size fraction (0.63–2 mm) and repacked at two different bulk density levels. The total magnitude of desiccation crack formation mainly depended on the clay content and clay mineralogy. Higher soil organic matter content led to a denser crack pattern with smaller aperture. Wetting-drying cycles did not only effect visible macroporosity (>8 μm), but also unresolved mesoporosity. The changes in macroporosity were higher at lower bulk density. Most importantly, repeated wetting-drying cycles did not lead to a randomization of distances between garnet particles and pores. This demonstrates that former failure zones are reactivated during subsequent drying cycles. Hence, wetting-drying resulted in reversible particle displacement and therefore would not have triggered the exposure of occluded carbon that was not already exposed during the previous drying event.