The Engineering Geological Assessment of an Urban Landslide at Wallis Road, Gisborne, New Zealand

Abstract

Instability at the southern end of Wallis Road (85 m asl) landslide, Gisborne city, has been recognised for several years, resulting in the house at No. 1 Wallis Road being demolished in 2015. Significant slope failure has occurred since then, particularly during winter of 2016 and 2017, culminating with runout from a channelized earthflow onto Kaiti beach below. The upslope extent of the landslide is characterized by rotational slumping bounded by a broad headscarp. Lateral scarps, back-tilted blocks and ponded surface water are also present. Below the upper area, a transition zone occurs, and below this, is the channelized earthflow, bounded by levees, and smoothed internal surfaces caused by scouring from rapid sediment discharge. At present, incipient slumping is most evident at the southern boundary of No. 8 Wallis Road, and active slumping of the headscarp has occurred at No. 3. Unmanned Aerial Vehicle (UAV) imagery taken in August of 2018 shows significant re-vegetation of the landslide, compared with late 2017, indicating less reactivation has occurred since the previous year. Historic imagery has identified the currently activated landslide as the central component of a larger landslide complex, seen in previously failed portions below Titirangi Drive and to the south in Gisborne District Council land. Dynamic cone penetrometer, shear vane testing and boreholes indicate the slip base at c. 6 m depth in the upper section, marking the boundary between the moderately to highly sensitive surficial early Pleistocene silts, clays and paleosols of the Mangatuna Fm and underlying impermeable mudstone of the Tunanui Fm. Permeability testing and electromagnetic conductivity surveys indicate that soils in the base of the gulley are reasonably saturated while the upper slopes are relatively dry, further suggesting stabilisation has occurred during 2018. Nevertheless, monitoring during the 2018 winter with a timelapse camera, a network of survey pegs and a piezometer, all indicate that deformation occurs in response to wet antecedent conditions, heavy rainfall and elevated pore pressures with the onset of winter. X-Ray Diffraction (XRD) and Scanning Electron Microscopy (SEM) were used to classify soils of the Mangatuna Fm, and an array of laboratory tests was used to identify the soil behaviour. Finally, limit equilibrium modelling provided insight in to pre- and post-failure activation scenarios in relation to seismicity and rainfall triggers. It is inferred that a single large, or multiple small seismic events triggered the initial (Holocene?) failure and subsequent reactivations have occurred due to elevated pore pressures along the boundary of the impermeable Tunanui Fm following periods of heavy rainfall. Furthermore, highly sensitive Mangatuna Fm. soils remould at residual shear strengths to culminate in the earthflow.