Experimental and theoretical studies have demonstrated that behaviour of unsaturated soils is governed by two independent stress-state variables, soil suction and net normal stress. Most documented case histories focused on responses of soil suction but often overlooked possible influences due to net normal stress. The principal objectives of this research are to investigate seasonal performance of a saprolitic hillslope due to changes of soil suction and net normal stress and to identify in situ groundwater flow mechanisms. Moreover, effects of the two variables and drying-wetting cycle on stress-dependent soil-water characteristic curve (SDSWCC) and permeability function, k(ψ) of a saprolite are measured and investigated....[ Read more ]

Experimental and theoretical studies have demonstrated that behaviour of unsaturated soils is governed by two independent stress-state variables, soil suction and net normal stress. Most documented case histories focused on responses of soil suction but often overlooked possible influences due to net normal stress. The principal objectives of this research are to investigate seasonal performance of a saprolitic hillslope due to changes of soil suction and net normal stress and to identify in situ groundwater flow mechanisms. Moreover, effects of the two variables and drying-wetting cycle on stress-dependent soil-water characteristic curve (SDSWCC) and permeability function, k(ψ) of a saprolite are measured and investigated.

Three interconnected research methodologies, namely full-scale field monitoring (Part I), laboratory investigations (Part II) and two- and three-dimensional (2D and 3D) flow analyses (Part III), are adopted. In Part I, a saprolitic hillslope situated in Hong Kong was selected to be heavily instrumented. Two-year seasonal variations of soil suction and net normal stress were closely monitored by heat dissipation matric water potential sensors and earth pressure cells, respectively. Slope movements were also measured. To provide input parameters for flow analyses in Part III, SDSWCC and k(ψ) of the saprolite were measured by two new test apparatuses directly in Part II. A series of 2D and 3D transient flow analyses were performed in Part III to back-analyse measured pore-water pressure (PWP) responses in Part I.

Evident seasonal slope movements are observed throughout the 2-year monitoring period. When a heavy rain fell in wet seasons, measured increases of both positive PWP and equivalent “effective” stress led to a “deep-seated”-type of down-slope ground deformation. An inelastic horizontal displacement of 40mm was resulted and a shear strain of 8% was estimated at 5.5 depth. In dry seasons, owing to suction recovery of 190kPa and substantial reduction of total horizontal stress, “cantilever”- and “translational”-type of up-slope movements were resulted at the top 5m of the ground. After subjecting to 2 cycles of dry-wet seasons, up-slope rebounds in dry seasons appear not to recover down-slope displacements measured in previous wet seasons fully. A general down-slope rachetting is identified.

Under the heavy rainfalls, significant rise of the main groundwater table which was at 11m depth is observed. This is mainly attributed to 3D cross-slope groundwater flow on top of a shallowed decomposed rock stratum. The increase of positive PWP in deeper regions due to the GWT rise and the observed “deep-seated” mode of ground deformation shape imply a reduction of stability for “deep-seated” slope failure.

Drying and wetting full-suction SDSWCCs and stress-dependent k(ψ)s of the saprolite were measured at vertical net normal stresses of 0, 40 and 80kPa. At a given suction, the k(ψ) decreased by up to 1 order of magnitude when average net normal stress increased from 4 to 78kPa. A noticeable hysteresis loop is observed between the measured drying and wetting k(ψ)s. On the contrary, hysteresis between the drying and wetting k(θ)s (where θ is volumetric water content) appear to be not very significant at any stress level.