Soil density, elasticity, and the soil-water characteristic curve inverted from field-based seismic P- and S-wave velocity in shallow nearly saturated layered soils

Abstract

Soil density, porosity, elastic moduli, and the soil-water characteristic curve (SWCC) are important properties for soil characterization. However, geotechnical and laboratory tests for soil properties are costly and limited to point sampling sites. Seismic surveys can provide laterally continuous, seismic soil property values that may complement geotechnical borehole tests with less cost. We have developed a workflow to invert for soil properties and the SWCC from seismic P-and S-wave velocity-versus-depth profiles interpreted from shallow (<25 m depth) unconsolidated sediments under conditions of near-full saturation (>99%). The inversion is performed by using the covariance matrix adaptation evolution strategy to search automatically for optimal input soil property values by minimizing the misfit between field-based velocity profiles and predicted velocity profiles based on the Hertz-Mindlin and Biot-Gassmann theories. The results from seismic soil property inversion are validated by comparison with geotechnical as well as laboratory results conducted independently in the same area as the seismic survey. For each seismically recognizable layer, soil types are interpreted from the inverted soil density and elasticity, aided by the SWCC to help detect thin units that are below the original seismic resolution of the field data. There is flexibility to apply our suggested workflow in future studies. For a known geologic setting, empirical relationships and other velocity prediction models could also be incorporated into the suggested workflow to improve inversion results and extract additional information in soils.

Stochastic Inversion of Geomagnetic Observatory Data Including Rigorous Treatment of the Ocean Induction Effect With Implications for Transition Zone Water Content and Thermal Structure

“…Therefore, our inverse problem formulation is based on a mixed scheme that combines a stochastic optimization technique known as Covariance Matrix Adaptation Evolution Strategy (CMAES) (Hansen & Ostermeier, 2001), with McMC methods (e.g., Mosegaard & Tarantola, 1995). Although the use of CMAES in geophysics is not common, it has been implemented in recent studies as a global minimization method (Alvers et al, 2013;Diouane, 2014;Fonseca et al, 2014;Grayver et al, , 2017Shen et al, 2015) outperforming other techniques such as genetic algorithms and particle Although the use of CMAES in geophysics is not common, it has been implemented in recent studies as a global minimization method (Alvers et al, 2013;Diouane, 2014;Fonseca et al, 2014;Grayver et al, , 2017Shen et al, 2015) outperforming other techniques such as genetic algorithms and particle…”

“…CMAES explores the model space globally and exhibits a remarkable robustness on ill-conditioned problems (Hansen et al, 2011). Although the use of CMAES in geophysics is not common, it has been implemented in recent studies as a global minimization method (Alvers et al, 2013;Diouane, 2014;Fonseca et al, 2014;Grayver et al, , 2017Shen et al, 2015) outperforming other techniques such as genetic algorithms and particle Journal of Geophysical Research: Solid Earth 10.1002/2017JB014691 Swarm optimization (Arsenault et al, 2013;Auger et al, 2009;Elshall et al, 2015). Additionally, showed that the use of CMAES for finding regions of low misfit can improve performance of conventional McMC methods.…”

Stochastic Inversion of P‐to‐S Converted Waves for Mantle Composition and Thermal Structure: Methodology and Application

“…This technique aims to explore the model space globally showing remarkable robustness on ill‐conditioned problems (Hansen et al, ). The use of CMAES in geophysics is not common, but has recently been implemented as a global minimization method (Alvers et al, ; Diouane, ; Grayver et al, ; Munch et al, ; Shen et al, ) outperforming other optimization techniques such as Genetic Algorithms and Particle Swarm Optimization (Arsenault et al, ; Auger et al, ; Elshall et al, ).…”

Seismic velocity inversion for patchy and homogeneous fluid-distribution conditions in shallow, unconsolidated sands

“…The inversion results include SWCC, S w , and the volumetric fraction of the patches ( Figure 5). We minimize the misfit between experimental and predicted V P and V S -versus-depth profiles for each WL, aided by the covariance matrix adaptation evolution strategy optimization (Shen et al, 2015). The best fit for the experimental data relies on the lowest rms misfit to arrive at the preferable inversion result.…”

“…Velocity models that are applied in our inversion (Shen et al, 2015) are based on the commonly accepted Hertz-Mindlin (Hertz, 1882;Mindlin, 1949) and Biot-Gassmann (Gassmann, 1951;Biot, 1962) (HM-BG) theories, but with different averaging methods depending on the patch size. When the patch size is small compared with the diffusion length, an average fluid bulk modulus can be given by the Wood (1941) average, which uses a weighted harmonic mean of the bulk modulus of each pore fluid.…”