Water retention and unsaturated permeability of sandy and silty soils: experimental determination and application to a two-layer capillary barrier system

Abstract

Rainfall-induced landslides in natural- and substandard cut- slopes are one of the most serious worldwide natural disasters. Slope instability in various unsaturated grounds has attracted increasing attention from researchers around the world. The soil water retention curve (SWRC) and permeability function (k(ψ)), where ψ stands for suction, are important soil properties affecting the hydraulic performance of slope covers. However, direct measurement of k(ψ) is often tedious and time-consuming, and in most cases cannot be used to investigate the effects of drying and wetting path. This study developed a soil column apparatus which allows one-dimensional water flow. Using the state-of-the-art sensors pore-water pressure (both positive and negative) and volumetric water content were continuously monitored during a drying-wetting cycle. As a result, SWRC was obtained and k(ψ) was determined using the Instantaneous Profile Method. The developed soil column was used to obtain the water retention curve and unsaturated permeability of a silty soil and a sandy soil during dying and wetting process. The air-entry value of silty soil was found to be 8 kPa, which was higher than that of sandy soil (3.2 kPa). At a given matric suction, the measured volumetric water content along drying SWRC was always greater than that along wetting SWRC, which exhibited the hysteretic behaviour. The sandy soil had a higher saturated permeability than the silty soil and the permeability function of the two soils reduced at different rates with increasing suction while silty soil dropped more slowly. Numerical soil column models were developed using the finite element package VADOSE/W and were validated using the experimental data. Furthermore, two numerical soil column models with silty soil overlying (or underlying) sandy soil were constructed to study water flow behaviour of the two-layer system when subjected to dying and wetting. The simulation results indicated that the two-layer soil column consisting the silty soil as upper layer could reduce the downward flow amount of water. Finally, some representative sets of simulations were further analysed to observe the effects of initial ground saturation, durations of the simulated precipitation and capillary barrier materials on a two-layer capillary barrier system. The particular capillary barrier model subjected to similar precipitation rates performed well for the case of an initially dry ground and became less effective for the case of an initially wet ground. In the case of long precipitation duration, the capillary barrier became less effective as a moisture barrier. When the underlying layer soil became too coarse, the capillary barrier model was less effective in maintaining negative pore-water pressure.