The strength, deformation, and hydraulic properties of geomaterials, which constitute embankments, vary with fine fraction content. Therefore, numerous research studies have been conducted regarding the effects of fine fraction content on the engineering properties of geomaterials. Howe ver, there have only been a few studies in which the effects of fine fraction content on the soil skeletal structure have been quantitatively evaluated and related to compaction and mechanical properties. In this study, mechanical tests were conducted on geomaterials with various fine fraction contents to evaluate their compaction and mechanical properties focusing on the soil skeletal structure and void distribution. Furthermore, an internal structural analysis of specimens using X-ray computed tomography (CT) images was conducted to interpret the results of mechanical tests. As a result, it was discovered that the uniaxial compressive strength increased with fine fraction content, and the maximum uniaxial compressive strength was observed at a low water content, not at the optimum water content. Additionally, the obtained CT images revealed that large voids, which could ser ve as weak points for maintaining strength, decreased in volume, and small voids were evenly distributed within the specimens, resulting in a more stable soil skeletal structure.

The maximum shear modulus (G(0(ij))) of rooted soils is crucial for assessing the deformation and liquefaction potential of vegetated infrastructures under seismic loading conditions. However, no data or theory is available to account for the anisotropy of G(0(ij)) of rooted soils. This study presents a new model that can predict G(0(ij)) anisotropy of rooted soils by incorporating the projection of the stress tensor on two independent tensors that describe soil fabric and root network. Bender element tests were conducted on bare and vegetated specimens under isotropic and anisotropic loading conditions. The presence of roots in the soil increased G(0(VH)) at all confining pressures (p '), as well as G(0(HH)) and G(0(HV)) at low p '. However, the trend was reversed at higher p ' because the roots reduced the effects of confinement on G(0(ij)) by replacing stronger soil-soil interfaces with weaker soil-root interfaces. Roots made the soil fabric and G(0(ij)) more anisotropic. The proposed model can effectively predict the observed anisotropy of G(0(ij)) under isotropic and anisotropic loading conditions. The new model also offers a new method for determining the fabric anisotropy of sand based on the anisotropy of shear modulus.