No abstract

To advance the understanding of the rainfall-induced landslide triggering mechanism, a comprehensive field monitoring programme was implemented in a saprolitic hillslope in Hong Kong. The instrumentation covered the measurements of the two stress-state variables (i.e., net normal stress and matric suction) and their effects. The monitoring results, including pore-water pressure (PWP), volumetric water content, subsurface total horizontal stress, horizontal displacement, and rainfall intensity, are reported. Most instruments recorded reliable and good quality data, which have strong correlation among each other. The site-specific infiltration and deformation characteristics of the hillslope subjected to heavy rainstorms are investigated. Shallow transient perched groundwater tables were believed to be developed at colluvial deposits on the top 3 m, where PWPs up to 20 kPa were typically measured. The main groundwater table probably rose by 6 m when the hillslope was subjected to rainfall intensity of 133.5 mm/h. It is possible that cross-slope groundwater flowed along a shallow, dipped decomposed rock stratum at the central portion of the landslide body. Besides, two distinct types of slope movements were generally observed, namely the “cantilever” and the “deep-seated” mode. Rupture surfaces have possibly been developed at 5 m below ground or deeper, resulting in a multiple translational- and rotational-slide type of failure.

It is well recognised that field measurements of stressdependent soil-water characteristic curves (SDSWCCs) and the permeability function (coefficient of permeability with respect to the water phase) are vital for assessing transient seepage, pore water pressure changes and the stability of unsaturated soil slopes. In this study, SDSWCCs and the permeability function were measured by using the instantaneous profile method on a saprolitic hillside in Tung Chung, Hong Kong. The ground profile comprises colluvium and completely decomposed tuff (CDT). A 3 m diameter circular plot was instrumented and subjected to two wetting-drying cycles. Throughout a 48-day monitoring period, variations of volumetric water content and matrix suction in the uppermost 3 m of soil were measured continuously using time-domain reflectometry moisture probes and jet-fill tensiometers, respectively. The field observations reveal that there is a clear trend of decreasing sizes of hysteretic loops of SDSWCCs with depth. In a comparison of the field-and laboratory-measured SDSWCCs, general agreement can be seen in the first wetting-drying cycle but not in the second cycle. The field-observed unsaturated permeability function is highly dependent on suction history (wettingdrying cycles). The field-measured permeability ranges from 4 3 10 27 m/s to 3 3 10 24 m/s in the colluvium, whereas the corresponding measured values vary from 4 3 10 27 m/s to 1 3 10 24 m/s in the CDT. Predictions of SWCC and permeability function using existing empirical approaches based on particle size distributions (PSDs) should be treated with caution. There is a clear qualitative and quantitative inconsistency between predictions of permeability function from PSD and field measurements.

No abstract