A geotechnical characterisation of volcanic soils in relation to coastal landsliding on the Maungatapu Peninsula, Tauranga, New Zealand.

Maungatapu Peninsula is a northeast trending peninsula located within the Tauranga Basin covering an area of 1.6km². Maungatapu is underlain by a sequence of volcanic tephras, ashes and fluvial deposits derived both locally and from the Taupo Volcanic Zone. In late May 1995 three landslides occurred at 83, 85 and 89 Te Hono Street, and again in late December 1995 at 330 Maungatapu Road. The purpose of this study was to carry out a geotechnical investigation of these landslides, and to establish the mechanisms that produce cliff failure on the Peninsula. Landslides were identified from aerial photographic interpretation and engineering geological mapping at a scale of 1:5000, and were classified as, 1) probable large scale block failures, 2) piping-triggered block failures, 3) wave erosion triggered block failures, and 4) colluvium/topsoil failures. Geotechnical core logging at a scale of 1:50 identified a number of stratigraphic units including the Post-Rotoehu Ash Tephras, Rotoehu Ash, Palaeosol, Hamilton Ash, Pahoia Tephras, Cross-bedded sequence, Upper Bounding Aquitard, Aquifer, and Lower Bounding Aquitard. The total thickness of the sequences are approximately 15m, and failures in 1995 were associated with a piping failure within the aquifer and lower section of the Crossbedded sequence triggering a block landslide. Geotechnical testing involved both field and laboratory testing to characterise the various stratigraphic units present within the logged cliff faces. In-situ shear strength testing indicated variable strength through out the profile, with the Palaeosol demonstrating the highest shear strength, and the Aquifer the lowest. This relationship was also confirmed by unconsolidated undrained triaxial laboratory testing. Clay mineralogy analysis indicated that the main constituent clays present were mixed layer 7 & 10 Å Halloysite and Allophanes. Atterberg Limit testing demonstrated a range of plasticities from low to very high. Direct shear testing indicated low cohesions and high friction angles for the Cross-bedded sequence and Aquifer, and a moderate cohesion and friction angle for the Lower Bounding Aquitard. Dispersion and Erodibility testing showed the Post-Rotoehu Ash Tephras, Rotoehu Ash, and Palaeosol to be non-dispersive and non-erodible, whilst the Cross-bedded sequence was dispersive and highly erodible. Both in-situ and laboratory permeability testing indicated low permeabilities associated with the stratigraphic units of the Peninsula. From field and laboratory investigations a hydrogeological model was developed to explain the fast lag times delineated by plots of piezometric water level response to rainfall. The hydrogeological model combined components of a "defect controlled permeability model" and a "hydraulic head response model". The "defect controlled permeability model" indicates that these fast lag times can be produced by soakage water permeating through high permeability flow pathways such as exfoliation defects, fractures, and heavy bioturbation structures. The "hydraulic head response model" involves the rapid transferral of a pressure wave along the Aquifer and lower section of the Cross-bedded sequence in response to changes in the hydraulic head of the Peninsula due to recharge within a much larger catchment of approximately 5km² Stability analysis using a non-circular failure mode was conducted for an increasing phreatic surface and landslide block size. The phreatic surface was related to piezometric water levels and showed that with an increase in the phreatic surface there was a decreased in the factor of safety by 0.1 from 1.0 to 0.9. Increasing the landslide block size was undertaken to determine whether larger blocks were likely to fail. From calculations it was concluded that failure of blocks greater than 10m back from the cliff edge were unlikely for the piping triggered model. Two principal conclusions can be drawn from this study. Firstly a 2H:1V slope line projected back up to the Peninsula's surface from the base of the cliff delineating a geotechnical assessment zone is not a correct representation of the failure types threatening cliff top properties. Therefore, this assessment criteria should be reassessed, and a policy adopted where by any future development on a cliff top property should require a geotechnical report if deemed necessary by the Consents Officer from evidence of slope failures in adjoining properties or other evidence of instability on site. The second conclusion is that it takes approximately two months of double the average rainfall to produce adverse pore water conditions at the cliff edges where a rainfall event can trigger a piping-triggered block slide such