Morphological and topological dynamics of soil pore structure during wetting–drying

Wetting–drying cycles significantly influence soil structure and are a major driver of its dynamics. Previous studies, based on X-ray tomography imaging, have primarily used classical scalar parameters – such as porosity, density, and pore-size distribution – to assess this dynamics. However, the information content of such metrics has recently been questioned. In this paper, we propose using high information content vector metrics: correlation functions (CFs), pore-network models (PNMs), and persistence homology. While the first two have proven effective for analyzing soil structure dynamics, the third – persistence homology – is not yet widely applied in soil science. Its application can capture topological changes in the structure that may be difficult to detect with CFs and PNMs alone. Using these three metrics, we quantified and compared samples from three different horizons of Greyzemic Chernozem (Loamic, Pachic) across various saturation states during a steady wetting–drying cycle. In total, we analyzed 150 3D images obtained using X-ray microtomography (with five replicates per horizon to account for soil variability). Our results enable a numerical analysis of the morphological and topological changes in the samples during steady wetting–drying. We also observed structural hysteresis and its degree of return to the original state after the cycle, which varied in intensity across horizons. We discuss the major advantage of high information content metrics – their capacity for a complete description of soil structure – while acknowledging that their comprehensibility is more complex than that of classical soil metrics, as they incorporate numerous structural properties simultaneously. The methodology proposed in this work paves the way for robustly describing and comparing soil structures across different scales.