UC Riverside

Mountainous soils usually contain a large number of rock fragments, particles with a diameter larger than 2 mm, which can influence soil hydraulic properties that are required to quantitatively describe soil water movement in stony soils. The objective of this study was to numerically estimate both the saturated hydraulic conductivity of a stony soil and its dependence on a relative content of rock fragments (stoniness), and the shape, position and distribution of rock fragments in a soil matrix. The assessment method was based on a numerical version of Darcy's classic experiment that involved steady-state flow through a porous material under a unit hydraulic gradient. Our experiments, involving hypothetical stony soils in this particular case, were simulated using mainly the two-dimensional (2D) numerical model, HYDRUS-2D. A limited number of simulations were carried out using a three-dimensional HYDRUS model. Three different shapes of hypothetical rock fragments were used in the study: a sphere, an ellipsoid with two different positions, and a pyramid, all represented by their 2D cross-sections (i.e., a circle, an ellipse, and a triangle, respectively). The mean relative effective saturated hydraulic conductivity (Krs) for the same stoniness was almost the same for all simulated scenarios and fine soil textures. A stoniness between 0.07 and 0.5 cm3 cm− 3 can cause a decrease of Krs in the range of 0.17–0.70. Numerical experiments were divided into 3 scenarios. The largest and the smallest values of Krs were different for different shapes of RFs (scenario A), different orientations of the slab-sided elliptical RFs (scenario B), and regular or irregular distributions of spherical RFs (scenario C). The largest difference between Krs values (0.26) was found in scenario B when the slab-sided elliptical RFs were oriented either horizontally or vertically for stoniness of 0.24 or 0.31 cm3cm− 3. Simulated Krs values were underestimated in all scenarios as compared to the Ravina and Magier (1984) function. The smallest differences (− 1.1%–2.5%) between numerically simulated and calculated (the Corring and Churchill (1961) method for a cylindrical shape of RFs) Krs values were found for scenario A with its 2D representation of spherical rock fragments. Calculated (the Corring and Churchill (1961) method for a spherical shape of RFs) Krs values corresponded well with those simulated using a 3D representation of spherical rock fragments. Numerical models provide a unique opportunity to evaluate the effects of different factors on the saturated hydraulic conductivity of stony soils that may be nearly impossible to measure in practice.