Shallow Shear-Wave Velocities from Surface Wave Dispersion, with Applications to Earthquake Hazard Assessment

Abstract

Amplified ground shaking and soil liquefaction can cause major damage in the event of a large earthquake. Estimates of the shear-wave velocities of rock in the shallow subsurface can be obtained quickly and efficiently. These make an important contribution to determining the seismic hazard at a site. In this thesis a relatively new technique to determine shallow shear-wave velocities is investigated, and is used in a case study of ground shaking and liquefaction hazard at three sites in the Wairarapa. The technique involves determination of velocities using the dispersion of Rayleigh surface waves analysed using the ReMi(TM)(refraction microtremor) method. Firstly, the ReMi method which as yet has had limited use in New Zealand is explored. Surface waves were recorded on refraction style geophone arrays and analysed through the ReMi slant-stack method to produce a one-dimensional shear-wave velocity model of the sub-surface at each site. Guidelines have been produced for experiment design and particular advantages and disadvantages of this method have been identified. Different sources of surface waves are compared for their effectiveness. Accurate dispersion results were obtained from both controlled sources (sledgehammer, dynamite and a hydraulic thumper) and noise sources (such a vehicle running along the length of the array). Case studies at two previously well-studied Wellington sites (Miramar Polo Ground and Boyd Wilson Field) test the accuracy of the method. In general, good agreement down to at least 30m is achieved between ReMi-derived shear-wave velocity profiles and structure determined through existing bore logs, SCPT and refraction data. The technique was able to successfully identify a dip in shallow basement, but was not able to reveal a thin layer of low velocity compared to the surrounding material. In general, the method was found to produce consistent, well-determined average velocities, with a larger uncertainty in the velocities and thicknesses of individual layers. The case study at three Wairarapa sites illustrates the application of the method to hazard assessment. Shear-wave velocity is derived at each site to depths of at least 15m and up to 60m at the Lake Wairarapa Barrage site. Shear-wave velocity was lowest at the Barrage site with a V30 (the average velocity in the top 30m) of 135 +/- 10m/s. This gives it the greatest ground shaking hazard of the three sites. V30 was 220 +/- 40m/s at Kahutara bridge also indicating a significant ground shaking hazard and was 340m/s +/- 80m/s at the Ruamahanga Bridge site. The higher velocities observed at the latter site, indicate a lower ground shaking hazard and also a lower liquefaction potential. The Barrage and Kahutara Bridge sites have a high liquefaction potential during a large earthquake, whereas at Ruamahanga Bridge liquefaction is possible, but would be confined to thin layers of liquefiable material such as sand, if any exist. Finally, this thesis also investigates the application of the ReMi method to larger crustal scale shear-wave velocity determination. Earthquakes sourced in line with the CNIPSE (Central North Island Passive Seismic Experiment) provide a basis for dispersion analysis. We determine a broad average shear-wave velocity model for a profile across the boundary of the Central Volcanic Region of the North Island. Although the uncertainty in our average profile is large and we are unable to determine a detailed velocity structure, our model is consistent with shear-wave velocity derived previously in the region through receiver function analysis.