A method is presented for estimating the diffusion coefficients for chloride and sodium in unsaturated coarse sand and fine gravel based on parameters obtained from saturated diffusion tests conducted for similar material. The method is tested by comparing the observed and predicted diffusion profiles through unsaturated soil. The method is shown to work well for predicting the advective-diffusive migration of chloride and sodium through a two-layer soil system consisting of a compacted clayey silt underlain by an unsaturated fine gravel. Over the range of conditions examined, it is concluded that existing solute transport theory along with the proposed procedure for estimating the unsaturated diffusion coefficients can adequately predict chloride and sodium diffusion through both unsaturated coarse sand and fine gravel as well as predict advective-diffusive transport through a compacted clayey layer and underlying unsaturated fine gravel.

The mechanical response of rock slopes and excavations depends largely on the geometry of discontinuities and associated rock blocks. Commonly used rigid block models for determining stability, such as the sliding wedge model or the keyblock analysis, assume that initial displacements are pure translations. This paper examines kinematic and kinetic constraints on block rotations. The kinematic constraints include conditions that are independent of the free surface and therefore belong uniquely to the joint pyramid, as well as conditions that depend on the orientation of the free surface. The relevant criteria, expressed in terms of vector inequalities, are represented on the stereographic projection, allowing graphical solution of the kinematical inequalities. Additionally, rotational equilibrium of tetrahedral blocks is considered, the failure modes of block theory are generalized to include rotational modes and procedures for evaluating rotational stability are discussed. The results show that blocks found to be stable with respect to the translational modes of sliding and falling may yet fail by rotation.

The process of soil formation in the tropics is quite different than that of more temperate regions. The weathering processes tend to be intensified by the warm and humid climates that are prevalent in tropical regions. Chemical weathering processes that dominate in these regions, such as decomposition and dissolution, are fairly well understood phenomena. However, the effect this weathering has on the physical, mechanical, and dynamic properties of the rock/soil is not fully understood. This paper presents some laboratory test results and correlations that indicate behavioral trends to be caused by the effects that weathering has on a residual soil from the western coast of Puerto Rico. The study focuses on the dynamic shear modulus G and material damping ratio D of both natural (undisturbed) and remolded samples, for lowto mid-level amplitudes of vibration. The understanding of these effects will allow for a better prediction of phenomena such as soil amplification that may trigger landslides in such soil formations. This study was accomplished with the aid of a computerized resonant column device. Correlations and empirical formulations are presented that demonstrate the effects that the degree of weathering has on the dynamic properties of these residual soils for undisturbed and remolded samples.

Sinkhole development in a profile with a weakly cemented sand overlain by an uncemented sand and underlain by a karst limestone was studied using 51 reduced-scale stress-correct centrifuge models. The interaction between the thickness ( H _c ), cohesive strength ( c ), and unit weight (γ_c) of the cemented sand layer; the thickness ( H _s ) and unit weight (γ_s) of the uncemented sand overburden; and the diameter ( D ) of the cavity in the underlying karst formation was examined. Collapse of the cemented sand layer into the underlying opening was observed to occur either as a breakthrough plug with steeply inclined sides at an average angle of 61° to the horizontal or as a dome fallout that left behind a stable arch. The type of failure depended on the model geometry, described by H _c /D. A dimensionless design chart was developed to predict failure, relating N γ_cH _c/c to H _c /D for configurations without uncemented sand overburden. The overburden pressure over the cavity at collapse was less than the geostatic overburden pressure. The fraction of full geostatic stress decreased with increasing overburden depth relative to the size of the cavity opening and approached an asymptotic value of about 5%.

Improvements to existing methods in the literature have been made for predicting cumulative plastic deformation for fine-grained subgrade soils. The soil deviator stress, number of stress applications, soil physical state, and soil type are considered. The improved method incorporates multilevels of deviator stresses and multisoil physical states that result from load-level variations, as well as seasonal and weather changes throughout traffic. Measurements of plastic deformation for a railroad-track subgrade are presented and show a significant influence of soil physical state, soil type, traffic tonnage, and wheel loads on the accumulation of plastic deformation. Comparisons between predicted and experimental results show good applicability of the improved method.

Two-dimensional (2D) finite-element analyses were carried out to study undrained soil deformation around piles displaced laterally through soil. The load-transfer p -δ curves produced were found to be applicable for design during passive lateral loading but not for active lateral loading of pile groups. The p -δ curves characterize the local soil-shear deformation around the pile, whereas p - y curves used in the subgrade-reaction method of active lateral-pile-loading design also include the effects of global soil displacement. Hence, determination of p - y curves from basic soil parameters requires consideration of both the local and global soil behavior and the pile-group geometry and loading. Because p -δ curves stiffen with reducing pile spacing, whereas active load-transfer curves soften, there is a significant stiffness disparity between the two forms of load-transfer curves for closely spaced pile groups. Care should be taken in choosing the appropriate form of reaction curve for design.