Soil Hydraulic Properties of a Large Undisturbed Macroporous Soil: Subsample Variability and Effective Parameterization.

When deriving the pressure retention curve h(θ) and the hydraulic conductivity curve K(θ) from soil samples, we have the problem of scale-dependency and representativity when using small samples, and the problem of internal heterogeneity for large samples. To investigate these issues, we extracted a cylindrical, 60 cm high undisturbed soil sample of loamy texture with a diameter of 30 cm from a natural hillslope. The column’s hydraulic properties were investigated by the extended multi step outflow (XMSO) method over a period of 60 days. Pressure head and water flow data over time were used to estimate the hydraulic properties by inverse simulation. We then cut the column into two segments of 30 cm height, and repeated the experiment. Finally, we cut the samples again, obtaining four cylindrical samples of 15 cm height. We repeated the XMSO for these four samples and continued, after full resaturation, with evaporation (EVA) experiments, which lasted 10 to 30 days. Results showed that the system was highly complex due to the presence of macropores, which strongly affected the early outflow of water in the XMSO experiments, and due to internal vertical gradients in the bulk density and conductivity along the original column, which makes the parameter estimation by inverse modelling challenging. In the very wet moisture range, where macropores were dominant, the two methods showed different soil hydraulic properties. In the medium moisture range, where the soil texture determines the hydraulic behavior, the retention curves of both methods were similar. In the dryer range, the XMSO contained no information, and the EVA measurements were limited by the range of the tensiometers (approx. -800 cm of pressure head). Evaluation of the EVA experiments by inverse modelling is possible, but requires an appropriate parametric expression for the dry end of the retention and conductivity curve.