| Summary: | This study investigates a slope failure complex in loess at Allandale, Lyttelton Harbour. Literature relevant to the slope stability and strength of Banks Peninsula loessial soils is reviewed. Laboratory and in situ strength testing shows that both C and P layer loess in a partially saturated state displays a significant reduction in undrained shear strength with increasing degree of saturation. Strength reduction can be attributed to reduced pore water tension due to capillary suction which results from an increased degree of saturation. The moisture controlled strength component in partially saturated loess can be defined by any two of dry density, moisture content and degree of saturation. When comparing loess C and P layer remoulded strengths with peak strengths, the P layer is significantly more sensitive to remoulding than C layer.
Drained direct shear testing of C layer loess produces remoulded and peak strength parameters of c'=O, Ø'=28.4° and c'=6kPa, Ø '=28.4° respectively. Drained direct shear testing of P layer loess produced remoulded and peak shear strength parameters of c'=O, Ø '=28.4° and c'=20kPa, Ø '=28.4° respectively.
The slope failure complex investigated has been formed by an earthflow initiated by a tension crack in C layer loess (which acts as an unconfined leaky aquifer). Subsequent retrogressive upslope and lateral migration of the slope failure complex involves "turfmat slides" in S layer loess which also acts as an unconfined leaky aquifer, and more tension crack initiated earth flows in C layer loess. Back analysis suggests both forms of slope movement may have failed by translational sliding at the base of their respective loess layer, with a piezometric level coincident with the ground surface. Mobilisation of the "turfmat slide", requires drained remoulded shear strengths, whereas mobilisation of the earth flow is more likely to involve drained peak shear strengths.
|
|---|
