A surface seismic approach to liquefaction

The liquefaction potential of soils is traditionally assessed through geotechnical approaches based on the calculation of the cyclical stress ratio (CSR) induced by the expected earthquake and the ‘resistance’ provided by the soil, which is quantified through standard penetration (SPT), cone penetration (CPT), or similar tests. In more recent years, attempts to assess the liquefaction potential have also been made through measurement of shear wave velocity (VS) in boreholes or from the surface. The latter approach has the advantage of being non-invasive and low cost and of surveying lines rather than single points. However, the resolution of seismic surface techniques is lower than that of borehole techniques and it is still debated whether it is sufficient to assess the liquefaction potential. In this paper we focus our attention on surface seismic techniques (specifically the popular passive and active seismic techniques based on the correlation of surface waves such as ReMiTM, MASW, ESAC, SSAP, etc.) and explore their performance in assessing the liquefaction susceptibility of soils. The experimental dataset is provided by the two main seismic events of ML = 5.9 and 5.8 (MW = 6.1, MW = 6.0) that struck the Emilia-Romagna region (Northern Italy) on May 20 and 29, 2012, after which extensive liquefaction phenomena were documented in an area of 1200 km2. The CPT and drillings available in the area allow us to classify the soils into four classes: A) shallow liquefied sandy soils, B) shallow non-liquefied sandy soils, C) deep non-liquefied sandy soils, and D) clayey-silty soils, and to determine that on average class A soils presented a higher sand content at the depth of 5–8 m compared to class B soils, where sand was dominant in the upper 5 m. Surface wave active-passive surveys were performed at 84 sites, and it was found that they were capable of discriminating among only three soil classes, since class A and B soils showed exactly the same VS distribution, and it is possible to show both experimentally and theoretically that they appear not to have sufficient resolution to address the seismic liquefaction issue. As a last step, we applied the state-of-the art CSR-VS method to assess the liquefaction potential of sandy deposits and we found that it failed in the studied area. This might be due to the insufficient resolution of the surface wave methods in assessing the Vs of thin layers and to the fact that Vs scales with the square root of the shear modulus, which implies an intrinsic lower sensitivity of Vs to the shear resistance of the soil compared to parameters traditionally measured with the penetration tests. However, it also emerged that the pure observation of the surface wave dispersion curves at their simplest level (i.e. in the frequency domain, with no inversion) is still potentially informative and can be used to identify the sites where more detailed surveys to assess the liquefaction potential are recommended.