Journal of Geotechnical and Geoenvironmental Engineering

Postshaking settlements observed during centrifuge tests of model levees resting atop soft compressible peat are compared with numerical settlement solutions. Two large-scale ($9\mathrm{m}$) tests and one small-scale ($1\mathrm{m}$) test are analyzed. The models included extensive instrumentation consisting of pore pressure sensors, accelerometers, bender elements, and displacement transducers to measure levee response during and after the application of scaled ground motions at the container base. Postcyclic settlement records suggested an increase in settlement rates within peat on cyclic loading compared with preseismic settlements due to the combined effects of excess pore pressure generation and secondary compression. The observed settlements were compared with the predictions of a one-dimensional nonlinear consolidation code that follows an implicit finite difference formulation. The code includes nonlinear compressibility and permeability properties and models secondary compression strain rate as a function of soil state rather than of time. Secondary compression was found to be the largest contributor to levee settlement. Further, cyclic straining was found to increase the secondary compression rate after earthquake shaking. Incorporating secondary compression reset into settlement predictions resulted in close agreement with measurements, whereas failing to consider secondary compression reset resulted in substantial underprediction of experimental settlement records.

This paper describes a new pseudostatic approach for an efficient seismic design of anchored steel sheet pile (ASSP) walls supported by shallow passive anchorages. As for other retaining structures, energy dissipation during strong earthquakes leading to reduced inertia forces can be achieved by allowing the activation of ductile plastic mechanisms. To this end, a robust method is required to identify all the possible yielding mechanisms and to guarantee the desired strength hierarchy. It is shown that dissipative mechanisms for ASSP walls correspond either to the local attainment of the soil shear strength in the supporting soil and around the anchor, or in the activation of a log-spiral global failure surface. A new limit equilibrium method is proposed to compute the critical acceleration of the system, corresponding to the actual mobilization of its strength, and the maximum internal forces in the structural members. Theoretical findings are validated against both existing dynamic centrifuge data and the results of original pseudostatic and fully dynamic numerical analyses.

This research introduces a new effective approach, including a weak vertical wall (WVW) and a strong inclined wall (SIW) under a shallow foundation, to reduce the rotations caused by a fault rupture. A series of centrifuge tests, followed by numerical modeling and verified by the test results, were conducted to explore the most suitable characteristics of an inclined wall. It is shown that a WVW is, in some cases, ineffective when it is not in the fault rupture path. Furthermore, uncertainties related to determining the exact location of the fault outcrop make it essential to construct a SIW beneath a foundation to protect the foundation that has already been improved by the WVW. The results proved that the application of the proposed approach could reduce the damage potential to the surface and embedded foundations located at various positions relative to reverse faults with various dip angles.

A review of liquefaction case histories shows that sand deposits containing some fines were common in past liquefaction events. While current liquefaction mitigation measures are mostly applicable to clean sands or open sites, development of nondisruptive techniques applicable to sands containing fines is critical. This paper examines the application and performance of microbial-induced partial saturation (MIPS) for liquefaction mitigation of sands with various silt contents. The investigation consisted of a set of undrained strain-controlled cyclic direct simple shear (DSS) tests on untreated and MIPS-treated samples. Experimental results suggested that high excess pore pressure was developed in untreated clean and silty sand specimens, which led to liquefaction depending on the shear strain level. Regardless of fines content and induced shear strain amplitude, MIPS-treated samples with only 4%–5% reduction in degree of saturation did not liquefy. A semi-empirical equation was adopted to predict the excess pore pressure generation in partially saturated conditions. The equation is able to reasonably predict the excess pore pressure generated in both clean and silty sands with variable degrees of saturation.

In this study, an analysis of soil–retaining wall dynamic interaction is conducted using three-dimensional discrete-element method (DEM) simulations. Soil grains are treated as rigid spherical particles that are allowed to overlap one another at contact points. The flexible sheetpile-type retaining wall is simulated using rigid balls glued together by parallel bonds with specific strength and stiffness to mimic the physical properties and stiffness of a real wall. Owing to computational limitations, the high g-level concept and scaling laws for dynamic centrifuge testing are utilized to decrease the domain size and simulation time. In addition, free-field boundaries are employed at the lateral sides of the model to prevent the reflections of the propagating waves back to the assembly and enforce free-field motion. Seismic excitation is introduced to the system through the base wall, which represents the bedrock. The effects of different characteristics of the input seismic wave, such as its frequency and amplitude, on the dynamic response of the soil–sheetpile system are analyzed. Furthermore, data on the lateral thrust and bending moment on the wall and its deflection are collected. It is found that the lateral earth pressure and bending moment increase during seismic excitation and the final residual values are, in most cases, considerably larger than the initial static ones. It is also observed that the maximum amplification of ground acceleration behind the sheetpile, the amount of wall deformation, and the maximum level of internal forces and moments the sheetpile experiences during dynamic loading are strongly affected by the frequency and amplitude of the input motion. The results show that for ground acceleration stronger than a critical limit, the maximum lateral earth pressure stays almost at a constant level. However, the maximum dynamic bending moment on the wall is found to increase even for ground accelerations higher than the critical value.

Initial shear stress is crucial in the three-dimensional (3D) behaviors of soil–structure interfaces but has seldom been examined. This research presents 3D monotonic and cyclic behaviors of a gravel–structure interface, considering initial shear stress from large-scale direct-shear tests. Tangential displacement is distinctly induced by initial shear stress due to the coupling effect, and the shear distance in the $x$- and $y$-directions exhibits a linear formation. The tangential displacement angle shows a close relationship with the initial shear stress and the orientation shear angle, and it can be expressed using a formula. The distinct noncoaxiality of the interface occurs during the yielding phase. The interface deforms toward the shear stress, and the unit tangential displacement increment reaches the peak when the shear strength is mobilized. The critical orientation shear angle and critical initial shear stress are found and determined by test results and formula calculation. Initial shear stress remarkably influences the magnitude of tangential displacement, shear stress, and irreversible and reversible normal displacements, while it slightly affects the shear strength of the interface, which behaves in an isotropic manner.

Study of the Krauklis wave in fractured media reveals the potential for characterizing fracture geometry. In this paper, a finite element scheme is proposed to simulate the Krauklis wave in fractures with different geometries. The influence of fracture geometry on the propagation properties of the Krauklis waves is evaluated. Here, fracture geometry refers to the fracture shape, width, and length. First, a series of numerical models containing a single water-filled fracture is set up, in which the fracture shape can be an ellipse or rectangle, its width varies from 5 to 10 mm, and its length varies from 2 to 3 m. The proposed finite element scheme is applied to simulate Krauklis waves. From these simulation results, the Krauklis wave signals inside the fractures are extracted, and a velocity estimation method is developed and implemented to estimate the average velocity within the finite length fractures. For comparison, the theoretical velocity is solved from a dispersion equation, in which the fracture is assumed to be an infinite thin fluid layer. It is shown that the rectangular fractures generate stronger amplitude and higher velocity waves than the elliptical ones. The amplitude varies along the fracture and can be significantly affected by the fracture geometry. Furthermore, the velocity increases with the fracture width, whereas the fracture length does not affect the velocity for both elliptical and rectangular fractures. These effects indicate that Kraulklis waves are rich in information about fracture geometry, and this is valuable when applied to the quantitative characterization of fractures.

Weathered gap-graded soils are a common geological body in mountainous regions, and they are widely used as construction materials. The shear strength is the controlling parameter for the design of civil projects; however, there is still a controversy on the coarse fraction effect on the shear strength of gap-graded soils. To this end, 22 triaxial shear tests are performed on gap-graded soils, and the factors affecting the coarse fraction effect have been analyzed, including the confining stress, the particle shape of aggregates, and the initial density of sand matrix. Partial contacts and sand bridges between aggregates are responsible for the transmission of loading and thus affect the coarse fraction effect. The results of triaxial tests reveal: (1) The overall shear strength of sand-beads mixtures is rather independent of the confining stress and the coarse fraction, and even the volume of aggregates is as high as 44.5%. (2) The effect of the shape of aggregates is effective only at a high coarse volume fraction (44.5%) for loose-sand-gravel mixtures, where the partial contacts between aggregates play an important role in forming the interaggregate structure. (3) The overall shear strength of gap-graded soils with a denser matrix increases continuously with a rising coarse fraction. The formation of a densified sand bridge is correlated with the initial density of sand matrix, contributing to the loading transmission in the interaggregate structure and, in turn, affecting the overall critical state behavior of gap-graded soils. The insights drawn from this study provide a reference for assessing the deformation behavior of weathered residual soils.

Specimen preparation methods and the resulting fabric can significantly influence the mechanical behavior of sand. To clarify the influence of specimen preparation methods on liquefaction behavior, a series of undrained cyclic triaxial tests were conducted on sand specimens prepared by different methods with a wide range of relative densities. Then, mesoscopic tests, including scanning electronic microscopy and digital image processing methods, were systematically conducted to observe the initial fabrics produced by different preparation methods. The triaxial test results demonstrated that different specimen preparation methods produce distinctly different responses, including the stress–strain relationship, failure mode, liquefaction resistance, and critical state line. Analysis of the mesoscopic testing results indicated that the initial fabrics of the specimens were affected not only by the preparation methods but also by the relative densities. A reasonable explanation for the linkage between the macromechanical behavior of the sand and the fabric formed in the specimen preparation process is given in this study.

The Atterberg limits are empirical indices of the critical water contents defining mechanical states of soil-water mixtures between the semisolid state and plastic state (the plastic limit or $w_P$), and between the plastic state and liquid state (the liquid limit or $w_L$). They reflect semiquantitatively a fine-grained soil’s ability to resist external loading. They are controlled by four fundamental factors: soil mineralogy, particle size distribution, pore fluid chemistry, and pore structure. However, general quantitative relationships among the Atterberg limits and these controlling factors are yet to be established. The authors hypothesize that the total amount of adsorptive water content for a given soil is directly related to the soil’s $w_L$ and $w_P$. A broad suite of 35 soils with measured soil-water retention (SWR) and the Atterberg limits from the literature were synthesized to explore these relationships. Using the measured SWR data and interpreting them through a generalized SWR model, the adsorptive water contents of all 35 soils were quantified. The authors demonstrate that the Atterberg limits, including $w_P$, $w_L$, and plasticity index $I_P(=w_L-w_P)$, were all correlated to mechanisms of soil-water interaction, specifically to a soil’s total adsorptive water content in terms of gravimetric water content, confirming the hypothesis. Further, the correlation between the Atterberg limits and a soil’s volumetric water content was poor, indicating the Atterberg limits’ independence from capillary water retention mechanism. The correlations provide a new pathway to move beyond the Atterberg limits to classify soil directly using more representative soil physical properties like adsorption suction stress, which could be linked to all four fundamental factors of soil mineralogy, particle size distribution, pore fluid chemistry, and pore structure.

This paper presents a comprehensive study of the load transfer mechanism of a soil anchor during pullout test. The anchor was monitored with fiber optic sensors installed along the tendon and embedded in the grout, which enabled continuous measurements of strains with a spatial resolution of about 10 mm. The measured strain profile along the tendon indicates that the load is not transferred from the tendon to the grout at the same rate along the fixed length. Crack development was verified by measured peaks along the longitudinal strain profile. The in situ measurements were compared with results from finite-element simulations contributing to the interpretation of the in situ results. The load–displacement curves and strain distributions obtained with the numerical simulations compare well with measurements. The constitutive model selected for the grout is capable of reproducing the crack development and the numerical simulations suggest that debonding occurred during the anchor pullout test.

The design of large diameter monopiles subjected to cyclic lateral loading is of interest in many applications, such as, for example, in the offshore wind energy industry. Engineering design methods to estimate the deflection of these structures are required for this purpose. In the present work, a $p–y$ model for large diameter monopiles subjected to lateral loading is proposed. The model considers a pile in cohesionless soil subjected to static and long-term cyclic loading. Its formulation features the consideration of nonlinear relations for the ultimate soil resistance $p_u$ and the initial subgrade modulus $E_{py0}$ among the soil depth and a cyclic factor that accounts for the effect of the soil density and loading amplitude. A relation is also proposed to account for a base shear force at the tip of the monopile. The proposed relations were employed and solved under the beam on nonlinear Winkler foundation (BNWF) approach and were adjusted to simulate a number of three-dimensional (3D) finite-element (FE) models accurately, accounting for variations on pile geometry, soil properties, and loading conditions. In the end, the performance of the proposed relations was evaluated through the comparison with a field test and a centrifuge test.

Use of a liquefied strength ratio to estimate strengths for postliquefaction slope stability analyses assumes that the liquefied shear strength of sandy soil can be normalized by the consolidation stress. This implies that the void ratio versus log consolidation stress and void ratio versus log liquefied strength relationships are parallel (or nearly parallel, for engineering purposes). Laboratory isotropically consolidated and monotonically loaded triaxial compression test data for eight normally consolidated sandy soils were used to show that this assumption may be reasonable in certain effective stress ranges—for example, greater than $1\mathrm{kg}/{\mathrm{cm}}^2$, for practical applications. The use of a strength ratio can facilitate postliquefaction stability and deformation analyses by representing the liquefied strength as a function of the preliquefaction effective vertical stress.

Discrete-element method (DEM) simulations of three-dimensional (3D) assemblies of ellipsoid particles were used to evaluate the critical state (CS) for both drained and undrained (constant volume) conditions. A series of conventional triaxial cyclic liquefaction tests with symmetrical cyclic deviatoric stress (${\sigma }_d$) with initial $q=0\mathrm{kPa}$ were simulated to develop a relationship between the cyclic stress ratio ($CSR={\sigma }_d/2{\sigma }_0^{\prime }$) and the number of cycles required for initial liquefaction ($N_L$), where ${\sigma }_0^{\prime }$ is the mean effective normal stress at the end of consolidation. Both cyclic mobility and instability type behaviors were observed depending on the initial void ratio ($e_0$) and ${\sigma }_0^{\prime }$. The micromechanics quantities, i.e., the coordination number ($CN$), von Mises fabric ($F_{vM}$), fabric anisotropy intensity (${\alpha }_c$), and stress-strain behavior, suggested that cyclic mobility and instability may depend on the phase transformation and instability state, respectively. The cyclic resistance ratio ($CR{R}_{15}$), i.e., $CSR$ at $N_L=15$, showed a unique relation with the initial state parameter (${\psi }_0$), irrespective of $e_0$ and ${\sigma }_0^{\prime }$. Two series of postliquefaction monotonic simulations with and without reconsolidation exhibited a unique CS, which perfectly matched with the original CS line. The $F_{vM}$ also reached a unique, narrow range at the CS. The postliquefaction settlement during reconsolidation also showed a linear relation with ${\psi }_0$.

This paper proposes simple numerical models that simulate the influence of tunneling-induced ground movements on masonry buildings founded on shallow strip footings. The focus is on developing an improved understanding of the impact of structural discontinuities due to pre-existing cracks and joints between adjacent buildings. The soil–structure system is idealized with equivalent Timoshenko beams connected to a homogenous elastic half-space with rigid links. The flexibility of the equivalent beam at the location of the crack or building interface is determined using linear-elastic fracture mechanics. The proposed models are first applied to a generic numerical example. Then, they are used to interpret the detailed displacement and strain measurements gathered from a masonry church during nearby tunneling. The results show that pre-existing cracks can have variable influence on the soil–structure interaction depending on their extent and position relative to the building and settlement trough. For the investigated case study, the proposed models illustrate how pre-existing large cracks in the hogging zone localized deformations and how the interaction between adjacent buildings drastically altered response mechanisms.

This paper presents a large-scale backward erosion piping experiment aimed at studying the erosion rate. This temporal aspect of piping complements previous research that focused on the critical head. To study the progression rate in realistic conditions, an experiment was carried out on a 1.8 m high levee with a cohesive blanket on a sandy foundation. The pipe was guided along a row of pore pressure transducers in order to measure its temporal development. Pipe development in space and time was successfully derived from pore pressure changes, showing an average progression rate of $8\mathrm{m}/\mathrm{day}$ during the progressive erosion phase. The results show a relation between upstream gradient and progression rate. Furthermore, analysis of the eroded sand mass shows a relatively large pipe volume compared to existing lab tests, and an approximately linear relation between pipe length and volume. The results and insights from this study can be used to validate and improve transient piping models, leading to more accurate dam and levee safety assessments.

This paper shows and compares the experimental results obtained by using a standard cone ($10{\mathrm{cm}}^2$) and a mini piezocone ($2{\mathrm{cm}}^2$). Tests were carried out at Calendasco (Piacenza, Italy) in a natural soil deposit mainly consisting of clayey-sandy silts. Grain size distribution with depth was also available. Other tests were carried out at Cavezzo (Modena, Italy), where liquefaction-induced phenomena were observed during the May 29, 2012, seismic sequence. The purpose was to investigate capabilities and limitations of mini piezocone and to explore the possibility of obtaining a better prediction of soil stratigraphy in thin layered deposits. Systematic differences in terms of tip resistance and sleeve friction were not observed, even though, generally, $q_c$ standard $<q_c$ mini. The spatial heterogeneity was considered responsible of the higher observed differences. In both test sites thin sandy layers are characterized by $I_c$ mini $\le I_c$ standard, with a ratio of the $I_c$ index as low as 0.88. This allows us to conclude that the minicone could be a valid alternative to the standard in identifying thin sandy layers.

Databases used in developing cone penetration test (CPT)-based design methods for driven piles have been dominated by data for the preformed variant, with the driven cast-in-situ (DCIS) pile category poorly represented. A database of 30 DCIS piles was used to appraise the ability of seven established CPT-based methods for driven piles to predict the total capacity of DCIS piles in granular soil. A subset of instrumented piles within the database was used to appraise the ability of these methods to predict shaft and base resistances separately. In general, the UWA-05 method provided the most reliable predictions for total and base capacity, whereas the LCPC-82 method is recommended for prediction of shaft resistance using shaft coefficients for standard driven piles. The paper collated findings from the recent instrumented pile tests conducted by National University of Ireland (NUI) Galway examining the mechanisms of shaft resistance in DCIS piles. In spite of uncertainties regarding mechanisms affecting the shaft resistance of DCIS piles during installation, the data from the instrumented tests presented herein provide strong evidence that the shaft resistance of a DCIS pile in sand is comparable to that of a preformed displacement pile. The NUI Galway pile tests also showed that the normalized base resistances for preformed and DCIS piles are equivalent.

The seismic performance of Balboa Boulevard during the 1994 $M_{\mathrm{W}}$ 6.7 Northridge earthquake was examined through nonlinear deformation analyses (NDAs) using advanced tools to (1) investigate the failure mechanism leading to ground deformations at this site; (2) evaluate the accuracy of the adopted analysis methods, engineering procedures, and state-of-the-art tools to reasonably estimate horizontal ground displacements; and (3) identify key factors and parameters contributing to earthquake-induced ground deformations at this site. One-dimensional (1D) liquefaction vulnerability indexes (LVIs) and permanent displacements using Newmark sliding block analyses were also estimated and compared against ground deformations observed after the earthquake. The geotechnical characterization of Balboa Boulevard was assessed based on field and laboratory data obtained from two investigation campaigns. Transitional probability geostatistics were used to develop stratigraphic models that capture the heterogeneity and the spatial variability patterns of sand-like and clay-like soils present at this site. The stratigraphic models were implemented in the finite difference software FLAC and the behavior of sand-like and clay-like soils simulated using the PM4Sand and PM4Silt constitutive models, respectively. Sensitivity analyses were performed to address uncertainties associated with the spatial variability of soils, input ground motions, the proportion of sand-like and clay-like soils within the soil deposit, and the strength properties of these materials. Results from NDAs suggest that a compounded effect of both liquefaction of sand-like soils and cyclic softening of clay-like soils led to the excessive ground deformations at Balboa Boulevard. This study sheds light on the importance of using appropriate engineering procedures and numerical modeling protocols in the prediction of deformation patterns, the selection of key input parameters, as well as the applicability of LVIs in complex sites.

The C Line of the Rome underground underpasses the historical city center, facing significant problems for the presence of archaeological artefacts and the necessity to prevent damage to the historical and monumental heritage. This paper describes the field performance of a 30-m-deep excavation that has been constructed at a short distance from the Aurelian Walls at Porta Asinaria (third century). An extensive instrumentation program was conducted during construction, as well as field observations including deflection of the diaphragm walls, ground movements, pore water pressures, settlements, and rotation of the Aurelian Walls. Observed ground settlements and diaphragm wall displacements were smaller than those monitored in other case histories worldwide, inducing no damage to the Aurelian Walls. The lessons learned from this case study showed that a stiff retaining system and a strict control of the construction sequence were the key to minimize the effects of a deep excavation with millennia of history at risk.

Undissolved gas bubbles, which are widely formed in marine sediments, could modify the behavior of the soil. To date, the cyclic behavior of fine-grained gassy soil has been rarely studied, limiting any reliable analysis of cyclically loaded offshore foundations in fine-grained gassy seabed. This study presents a series of undrained cyclic triaxial tests on normally consolidated (NC) fine-grained gassy soil. Each sample was charged with the same amount of ${\mathrm{N}}_2$ gas before consolidation. Then, samples were consolidated to an identical mean operative stress (total stress minus pore water pressure) of 200 kPa, but at different initial pore water pressures ($u_{\mathrm{w}0}=0–600\mathrm{kPa}$) before cycling. Compared with a saturated sample, the inclusion of a small amount of gas bubbles can either significantly decrease the level of cyclic stiffness degradation when the initial pore water pressure $u_{\mathrm{w}0}$ is lower than a threshold value (150 kPa in this study), or vice versa. This behavior is associated with two competing mechanisms during the cyclic shearing: (1) collapse of gas-filled pores, which increases cyclic pore water pressure (CPWP), and therefore stiffness degradation; and (2) partial water drainage into gas cavities, which reduces CPWP and stiffness degradation. The former and the latter are likely to dominate at relatively high and low $u_{\mathrm{w}0}$, respectively. This implies that water depth would significantly influence the cyclic deformation and stiffness degradation of offshore foundations in fine-grained gassy sediments.

Despite the potentially significant influence of tension cracks on earth pressures, almost all of the existing methods either do not account for cracks or treat them in a simplistic manner when dealing with unsaturated soils. The main objective of this study is to quantify the influence of tension crack formation on active earth pressures (AEP) in unsaturated backfills. A limit equilibrium-based framework for the determination of the tension crack depth (TCD) and AEP in unsaturated backfills is proposed. An effective stress approach is employed along with logarithmic spiral failure geometry to assess the influence of the interdependence between normal stress distributions, suction stress, and soil tensile strength on the formation of tension cracks under steady flow conditions. The maximum TCD is explicitly determined by evaluating a normal stress distribution and associated tensile stresses mobilized along the critical failure surface in retained backfill, with or without tensile strength. A parametric study is performed on a hypothetical marginal backfill to investigate the impact of a tension crack on AEP with and without considering the soil tensile strength under different flow rates. The results indicate that the formation of tension cracks will significantly increase AEP, while considering a nonzero tensile strength as a function of the matric suction will decrease the TCD and AEP. Upon comparison, it is shown that classic earth pressure theories may significantly underestimate or overestimate TCD in unsaturated backfills. The proposed approach presents a useful tool for forensic studies of failed structures and determining the depth of cracks in existing retaining earth structures that are to be rehabilitated.

This technical note presents simplified parametric solutions for the axial response of surface-laid submarine pipelines subjected to axial drag from debris flows. In assessing the response of pipelines impacted by debris flow emanating from a submarine landslide, both normal and axial responses must be considered. Previous work has indicated that these can be decoupled, at least as a first-stage analysis. The most critical aspect of axial drag is the potential for the pipeline to buckle. However, to make preliminary estimates of displacements and forces along the pipeline prior to buckling, simple assumptions of elastic pipeline response with elastic perfectly plastic interaction with the seabed are justified. These allow the development of parametric solutions that contain only three nondimensional quantities. This technical note documents the solutions and illustrates their application for some typical input conditions.

Sediments experience shear and volumetric strains during freeze–thaw cycles. Measurements during globally drained and undrained cycles under constant deviatoric stresses show that the asymptotic shear and volumetric response vary with sediment type and drainage conditions. In particular, the sediment response is intimately related to the ice pore habit that results from effective stress and the ice capillary pressure ${\sigma }_z^{\prime }/\mathrm{\Delta }{u}_{iw}$. Pore-invasive ice formation in coarse-grained soils may trigger some contraction during the first freeze–thaw cycle, even in sands denser than the critical state. Grain-displacive ice growth in fine-grained soils causes cryogenic consolidation of the surrounding sediment; subsequent melting of the segregated ice lenses yields a high increase in pore water pressure during undrained thawing, a pronounced volume contraction under drained conditions, and preferential shear deformation along melting ice lenses in either case. Both dilative sand and normally consolidated (NC) clay specimens subjected to deviatoric loading exhibit unceasing vertical strain accumulation (i.e., ratcheting) during freeze–thaw cycles; the void ratio evolves toward asymptotic values in all cases. The freezing rate relative to the pressure diffusion rate $\Pi =D_T/C_v$ regulates drainage conditions during freeze–thaw cycles; globally drained freezing and thawing are anticipated in coarse-grained sediments.

Heat flow controls the design and operation of a wide range of engineered geosystems. This study uses transient thermal probe measurements to determine the evolution of the thermal conductivity of air-dry and water-saturated sand–silt mixtures as a function of effective stress. Results confirm that the thermal conductivity of soils varies with state of stress, dry mass density, mineralogy, and pore fluid properties and highlight the effect of thermal contact resistance on the thermal conductivity of granular materials. Thermal conductivity follows a linear relationship with the logarithm of effective stress as a consequence of fabric compaction, increased coordination number, contact deformation, and reduced thermal contact resistance. The bulk thermal conductivity of water-saturated soils is more than seven times that of air-dry soils for the same fines content (FC) and effective stress. Pore-filling fines contribute conduction paths and interparticle coordination; the peak in thermal conductivity takes place at $\mathrm{FC}\approx 0.4$; this mixture range corresponds to the transition from fines-controlled to coarse-controlled mechanical response (i.e., both fines and coarse grains are load bearing), in agreement with the revised soil classification system.