Abstract. Prediction and modeling of localized flow processes in macropores is of crucial importance for sustaining both soil and water quality. However, currently there are no reliable means to predict preferential flow due to its inherently large spatial variability. The aim of this study was to investigate the predictive performance of previously developed empirical models for both water and air flow and to explore the potential applicability of X-ray computed tomography (CT)-derived macropore network characteristics. For this purpose, 65 cylindrical soil columns (6 cm diameter and 3.5 cm height) were extracted from the topsoil (5 cm to 8.5 cm depth) in a 15 m × 15 m grid from an agricultural field located in Silstrup, Denmark. All soil columns were scanned with an industrial X-ray CT scanner (129 µm resolution) and later employed for measurement of saturated hydraulic conductivity, air permeability at −30 and −100 cm matric potential, and gas diffusivity at −30 and −100 cm matric potential. Distribution maps for saturated hydraulic conductivity, air permeability, and gas diffusivity reflected no autocorrelation irrespective of soil texture and organic matter content. Existing empirical predictive models for saturated hydraulic conductivity and air permeability showed poor performance, as they were not able to realistically capture macropore flow. The tested empirical model for gas diffusivity predicted measurements at −100 cm matric potential reasonably well, but failed at −30 cm matric potential, particularly for soil columns with biopore-dominated flow. X-ray CT-derived macroporosity matched the measured air-filled porosity at −30 cm matric potential well. Many of the CT-derived macropore network characteristics were strongly interrelated. Most of the macropore network characteristics were also significantly correlated with saturated hydraulic conductivity, air permeability, and gas diffusivity. The predictive Ahuja et al. (1984) model for saturated hydraulic conductivity, air permeability, and gas diffusivity performed reasonably well when parameterized with novel, X-ray CT-derived parameters such as effective percolating macroporosity for biopore-dominated flow and total macroporosity for matrix-dominated flow. The obtained results further indicate that it is crucially important to discern between matrix-dominated and biopore-dominated flow for accurate prediction of macropore flow from X-ray CT-derived macropore network characteristics.
Predicting saturated hydraulic conductivity from air permeability: Application in stochastic water infiltration modeling
