A kriging method of interpolation used to map liquefaction potential over alluvial ground

Soil liquefaction is a devastating geotechnical phenomenon associated with earthquakes that causes failure of foundations, soil embankments and dams, especially in cities built on young alluvial deposits. Therefore, the ability to predict liquefaction potential is important and valuable for seismic hazard assessment. The conventional approach to evaluating liquefaction potential at any given point is through the use of borehole data. The liquefaction potentials within sedimentary basins vary over short distances with local soil properties. The purpose of this paper is to evaluate liquefaction potentials at unsampled locations where borehole data are not available based on the measured liquefaction potentials at selected borehole locations. The application and the importance of the geostatistical analytical technique (kriging method) for the estimation of liquefaction potential at unsampled locations are shown. The data set consists of liquefaction potential values measured at 86 randomly distributed boreholes in Saitama City, Japan. With the use of calculated liquefaction potentials at sampled locations, the experimental semivariograms were constructed. The experimental variogram characterizes the spatial variability of the calculated liquefaction potential. To interpolate the liquefaction potential over the entire area, the ordinary kriging spherical model was applied. In applying the kriging method, both isotropic and anisotropic models were used; however, the anisotropic model exhibited a better fit for interpolating sedimentary deposits. The interpolated liquefaction potentials are validated with the measured liquefaction potentials of an independently collected second set of 41 boreholes. The interpolated liquefaction potentials show strong concurrence with the measured values.

► Liquefaction potential at unsampled locations was made by using limit sample points.
► Kriging method of interpolation is a suitable and unbiased method for interpolation.
► Direct measured and interpolated values at same point show significance correlation.
► Anisotropic modeling in interpolation gives more significant results.
► Historical liquefied points are plotted in the high liquefaction zone.