Geomechanical characterisation and dynamic numerical modelling of two anthropogenic fill slopes

Abstract

In the last 22 years, four large (between approximately 1000 and 5000 m3) flowslides in anthropogenic fill materials have occurred in residential areas of the greater Wellington region, New Zealand. Such failures are relatively rare, given that there are over 1600 mapped fill bodies, formed between the early 1900s and present day within the region. Nonetheless, their performance under strong earthquake shaking has yet to be tested, given no strong ground motions (>0.15 g) have shaken the region since their construction. In this study, we used remote sensing techniques, borehole investigations, geotechnical testing and numerical modelling to investigate two fill bodies that are characteristic of those constructed across the Wellington region. This paper presents the results from the dynamic analysis of these two sites under strong earthquake shaking.The maximum anthropogenic fill thicknesses at both sites ranged between 20 m and 30 m. The fills tested varied from sandy gravel with some silt to clayey gravelly silt sand (D50 from 0.02 mm to 10 mm and D90 from 5 mm to 50 mm). Based on the field mapping, site investigation and laboratory results, two-dimensional geotechnical cross-sections of the sites were generated and used as the basis for the numerical simulations. Accelerograms recorded from eleven local and overseas earthquakes were used as inputs for the modelling. These were selected and scaled to simulate the amplitude and frequency content of the different types of earthquake shaking that could affect the sites up to 2.5 g, which represents a peak ground acceleration with an approximate annual exceedance probability of 0.00005. The results indicate simulated permanent ground displacements, due to sliding of the fill in the range of 0.01 m to >10 m, with the amount of displacement increasing with increasing Peak Ground Acceleration (PGA). Although at small displacements it is unlikely the fill slopes would fail catastrophically to form flowslides, the simulated permanent displacements are large enough for an earthquake with a free field PGA of >0.6 g (annual exceedance probability (AEP) of 0.002) to potentially disrupt buried elastic pipe utilities (especially the water pipes). Such deformation could lead to the leakage of water from broken pipes, and an increase in pore-water pressure within the fills, which could in turn lead to the development of post-earthquake flowslides. After a major earthquake, such cascading hazards may pose a major problem for recovery activities.