Unsaturated Soils 2006

The paper presents an evaluation of the current state of development of unsaturated soil mechanics by directing the attention to the capabilities of existing methods of analysis to face real problems. A review of recent contributions in several fields of application: fills, embankments and dams; pavements, foundations; landfills; slopes; nuclear waste disposal and infiltration is presented. The review shows the increasing sophistication of available tools although well documented case records remain scarce.

Geotechnical and Environmental Engineers face the challenge of managing ever-increasing quantities of mine and municipal solid waste, including disposal, rehabilitation of depositories and long term maintenance and monitoring. Waste management is a long-term process that may be required to continue for centuries into the future, as there are no walk-away solutions to the maintenance of waste depositories. Climates that are water-deficient, to various degrees, cover 65 percent of the inhabited Earth. In some ways, waste management is simplified by a water-deficient climate, in others it is made more difficult. After describing the characteristics and arising of mine waste (MW) and municipal solid waste (MSW), the paper proceeds to describe selected aspects of waste management in water-deficient climates, including: (1) water retentivity, (2) shear strength and slope stability, (3) erosion of waste depository surfaces, and (4) settlement of MW backfill and MSW landfill.

Application of unsaturated soil mechanics to practice requires measurement of soil suction. This paper examines several devices and techniques that are used for suction measurements in the laboratory and the field. Salient features of each suction measurement device are highlighted.

This paper presents a description of a new volume-mass constitutive model for unsaturated soils. The model requires relatively simple laboratory tests to obtain the calibration information. Volume and water content changes can be independently computed using the model. The model is also capable of taking into account: i) the hysteretic nature of the soil-water characteristic curve; and ii) both elastic and plastic deformations in the soil. The predicted results for several artificial soils (i.e., sand, silt and clay) are consistent with observed behavior. A presentation of the model predictions for the volume-mass constitutive surfaces is also presented.

Calibration of soil-atmosphere models to in situ measurements of suction, moisture content, and lysimeter outflow in response to atmospheric conditions can be time consuming. Calibration usually involves iterative and often arbitrary adjustment to estimated or laboratory-determined material properties until an acceptable match is achieved between predicted and measured responses to atmospheric conditions. Stepped irrigation and draindown tests on instrumented lysimeter test plots, however, can provide an improved basis for estimating bulk material properties for use in model development beyond that available from laboratory measurements or calibration to measurements during normal atmospheric responses. Irrigation tests provide, in part, information regarding the change in the volumetric moisture content and unsaturated hydraulic conductivity. The relationship can be directly measured during the test. This provides valuable information for use in calibrating a numerical model, specifically for capturing scale effects and constraining the relationship where standard models based on a proposed soil water characteristic curve are not valid. This paper describes a stepped-irrigation test conducted on an instrumented lysimeter in a mine waste rock pile. The irrigation portion of the test was conducted for seven days at two application rates. The draindown portion was monitored for 14 days following the test. The soil-atmosphere model based on the parameters from the irrigation test achieved a closer match to the observed conditions than the model calibrated to in situ measurements alone.

An instantaneous profile (IP) test was conducted at Sandia National Laboratories, Albuquerque, New Mexico. The IP method was used to determine the field-saturated and unsaturated hydraulic conductivities of a natural soil profile. The soil profile was instrumented with two neutron probe access tubes and three clusters of paired tensiometers and time domain reflectometry (TDR) probes. The soil profile was flooded for a period of 48 hours. Flooding was discontinued when the soil was field-saturated. Neutron probe and tensiometer data were acquired on a continual basis for 890 days and 200 days, respectively. As the soil profile was allowed to drain, soil moisture and suction data were recorded throughout the soil profile. These data from the IP test were used to calculate the unsaturated hydraulic conductivity as a function of the volumetric moisture content of the soil profile. The curves representing hydraulic head, H, change with depth, z, exhibit a clear change in slope between 90 cm and 120 cm due to a caliche layer. Values of hydraulic conductivity were calculated from the slope of the ∂H/∂z curves corresponding to a given depth and time. The hydraulic conductivity of the soil profile is represented by a family of curves that represent the variations in soil characteristics within this soil profile.

A model of dynamic soil moisture distribution in unsaturated soil is proposed as a simple solution to Klute's equation. This model can represent relationships among soil moisture, depth and elapsed time using a sigmoid function. By substituting this model into Klute's equation (the governing equation of one-dimensional vertical seepage flow in the unsteady state), Klute's equation can be solved easily and the hydraulic conductivity of unsaturated soil can be obtained without suction data.

To study seepage behavior caused by rainfall, field tests were carried out on model embankments that consisted respectively of Hachinohe Shirasu and Tsukidate soil. They are volcanic pyroclastic sediments with pumice that exist in Aomori and Miyagi Prefectures in Japan. Due to rainfall the maximum value of water content in earthfills which consist of Hachinohe Shirasu and Tsukidate soil were 40–42 percent and about 42 percent respectively. Their peak had tend to occur after each 1day and 2.5day of rainfall peak. Furthermore from the field test results, storage coefficient, which is normalized by porosity, at wetting process was 0.01–0.25 in Hachinohe Shirasu and Tsukidate soil. Subsequently, using hydraulic properties determined from laboratory tests, finite element analyses of saturated-unsaturated seepage in those embankments were performed. Comparing the calculated behavior with measured one for about three months data, after 20 days we could conclude that the calculated seepage behavior agreed with observed one.

Infiltration characteristics of a single-layer soil and a two-layer soil were investigated in this study by conducting small-scale field infiltration tests and numerical analyses. The field infiltration tests were conducted using double-ring infiltrometers, where the water pressure head was maintained throughout the tests. Changes in pore-water pressure during the tests were measured using tensiometers installed at various depths in the soil. Simulations of the field infiltration tests were performed using a finite element program. The hydraulic properties of the soils used in the simulation were measured independently in the laboratory. The results showed that the difference in the hydraulic properties of the two-layer soil caused a delay in the downward water flow into the lower soil layer. The numerical simulation results and the field infiltration test results were found to be in reasonably good agreement.

The monitoring of saturation and desaturation processes is a critical issue for the assessment of potential landslides. In this study, results from the first year of a monitoring experiment on a 15 by 13 metre square test site in the vicinity of previous landslides are presented. A top layer of clayey sand overlies weathered sandstone just above the rock surface and this layering system allows rapid infiltration. Preceding geophysical investigations helped to optimize the configuration of the test site. Meteorological data (e.g. rainfall, temperature and air pressure), infiltration and outflow, suction and volumetric water content at different depths up to 1.5 m have been measured continuously using among others tensiometers, Time Domain Reflectometers and MoisturePoint devices. Interaction has been observed between suction and volumetric water content and between rainfall intensity and the reaction at different soil depths. Seasonal hydrological changes can also be presented in terms of the response characteristics of the soil to rainfall events.

When new development occurs in arid regions, the boundary conditions and in particular the amount of water applied to the soil is usually changed. This change in the boundary condition at the surface can induce a change in the soil suction profile. In the case where the soils underlying structures exhibit volume change in response to a change in soil suction, a change in the soil suction can in turn induce volume change in the soil. As a consequence, the degree (amount by which the suction changes) and extent (depth over which a change occurs) of soil suction change are of interest to designers developing recommendations for many aspects of new construction, including grading, floor and foundation support, and landscaping. A series of suction measurements are reported for samples recovered from over 30 sites in the Denver metropolitan area where single-family residential structures have been in-place between 7 and 62 years. Samples were obtained to depths of approximately 12 meters. A companion set of results are presented for sites at which no development has yet occurred. In general, comparison of these results suggests that residential development affects the underlying soil suction profile in complex ways. In several cases, no effect can be observed. In most, suction decreased in the upper 3 to 7 meters, and in a few cases deeper wetting was suggested. The depth of suction decrease did not correlate to the age of the structure. However, operating characteristics, in particular landscaping choices and watering policies, have a significant influence over the degree and extent of suction changes.

With the advent of improved foundation design and analysis methodologies, more accurate strength and stiffness values are required to yield reliable structure performance. In particular, lateral modulus of deformation values (synonymous with pressuremeter modulus in this paper) are needed in the design of drilled pier foundations for electric power transmission single pole structures. Parameter values can be developed from either in-place field tests or laboratory tests. More commonly, these parameters are estimated from published empirical correlations with Standard Penetration Resistance Test (SPT) N values or soil consistency. A number of researchers provide empirical pressuremeter relationships in both cohesive fine grained and granular deposits, but few look solely at unsaturated soils. The Salt River Project (SRP) has performed field pressuremeter testing for use in design of a number of its high voltage electric transmission structure foundations over the past 20 years. This paper presents the data and discusses the development of empirical correlations between field pressuremeter tests and standard penetration resistance blow counts obtained by the SRP in desert southwest unsaturated alluvial soils. Data from four projects within Maricopa County, Arizona are examined to develop the relationships. Data is supplemented with work performed by the Electric Power Research Institute at an SRP substation in the Phoenix area. The purpose of the work is to provide more reliable estimates of lateral deformation modulus values for projects in the SRP service territory where it is not economical to perform site-specific pressuremeter testing. The correlations from this work appear to be very similar to those determined by previous researchers. The data from this work indicates somewhat less scatter than previously observed by others with the ratio of measured versus estimated deformation modulus values ranging from 0.3 to 2.5 for the linear regression average lines and 0.4 to 3.3 for the low bound 90 percent confidence interval curves.

Existing methods for the detection and measurement of volatile organic compounds (VOCs) in the vadose zone are often hampered by issues associated with ease of use, accuracy, and cost. As a result, there remains a need for inexpensive, minimally invasive, real-time instrumentation and sensor systems that can be used for characterizing or long-term monitoring of contaminated sites. A new sensor system to meet this need is under development at Boise State University, in collaboration with Washington State University. A miniature Ion Mobility Spectrometer (IMS) has been fabricated from a machinable ceramic material and has been shown to have a resolving power comparable to a commercial IMS. The IMS system will be packaged in a probe housing for deployment using direct push methods and will be ultimately equipped with water content, temperature, and pressure sensors. The proposed system is designed for use in multi-probe arrays and features wireless transmission of data directly to the user. To our knowledge, an in-situ IMS for detection of subsurface gaseous VOCs has not been previously developed.

This study was undertaken to begin evaluating the influence of soil suction on shear strength parameters determined from the Iowa Borehole Shear Test (BST) in unsaturated soil, namely the friction angle and cohesion intercept. A new technique, the Flooded Borehole Shear Test (FBST), was employed in this study. The BST, before and after flooding the borehole, was conducted for comparison of results. Suction was determined using the filter paper technique on undisturbed specimens in accordance with the ASTM D 5298 test method. Results showed that the friction angles determined from the FBST were much lower than those determined from the BST, as expected, apparently due to the reduction in suction due to flooding. The reduction in friction angle due to flooding generally followed a predictable pattern that was a function of the suction prior to flooding. However, the effect of suction on the cohesion intercept was much less pronounced. The cohesion intercepts from the FBST generally decreased slightly in comparison to the BST results, although cohesion intercepts were small in both cases. This paper discusses the practical implications associated with BST testing in unsaturated soil.

Seasonal Monitoring Program (SMP) data available in the Long Term Pavement Performance (LTPP) database DataPave was analyzed to investigate the variations of volumetric water content. The SMP data includes volumetric water contents from time domain reflectometry (TDR) probes in pavement sections located in the United States and Canada. Water content, or the associated degree of saturation, is used to compute resilient modulus for unsaturated unbound base and subgrade soils in the Mechanistic-Empirical Pavement Design Guide (M-EPDG) that was developed for the Federal Highway Administration. The purpose of this paper is to discuss results of analysis of volumetric water content data from the most recent release of DataPave (Release 19) and the resulting variations in resilient modulus that would occur. Results from analysis of the data indicate that there are variations of volumetric water content that occur over time. For a few of the pavement sections, the variations of volumetric water content were seasonal. For most of the sections, it was not possible to determine consistent trends in the moisture variations on a temporal scale or when comparing the different climate zones, soil types (coarse or fine-grained), pavement types or depth to the water table. The volumetric water content variations typically were greater than 3 percent and less than 9 percent. These findings indicate that subgrade soils undergo varying degrees of saturation. The resilient modulus, computed using the water content variations and an empirical equation developed for the M-EPDG, can vary by as much as a factor of 2. The resilient modulus variations are generally higher in wet climates than in dry climates.

Several models exist to predict the interaction between unsaturated soils and slab foundations or pavements. Validation of these models requires real-world data about the distribution and movement of water within the soil. Electrical resistivity tomography (ERT) offers an economical, non-destructive, and relatively non-invasive method of creating 2- and 3-dimensional views of moisture distribution. This is a case study of an asphalt bicycling/walking path constructed on an expansive clay subgrade. A test section was underlain by a horizontal moisture barrier, in an attempt to minimize seasonal shrink/swell behavior. A string of electrodes was installed across the pavement section, beneath the barrier. By measuring the subgrade resistivity at monthly intervals, it was believed possible to estimate the depth of the active zone, the edge moisture variation distance, the long-term equilibrium moisture beneath the covered area, and seasonal fluctuations of moisture at the edge.

This paper presents the results of a field experiment in applying horizontal membrane barriers to control longitudinal cracking in a bike trail pavement. A detailed site investigation of the test section was made to characterize the subsoil conditions which include classification, gravimetric and volumetric moisture content, engineering properties, matric suction versus depth, and soil water characteristic curve. The field experiment involves evaluating the effect that three widths of the horizontal membrane barrier beyond the edge of the pavement have on reducing longitudinal cracking. The premise controlling the behavior of this pavement is that changes in matric suction result in a horizontal shrinkage component that causes the longitudinal cracking. The evaluation of the effectiveness of the horizontal membranes involves the influence of the edge moisture variation distance and the active depth. In summary the horizontal membrane barrier was successful in eliminating the longitudinal cracking to date.

A detailed investigation is reported of the effects of two deciduous trees on water contents in a shrinkable soil stratum supporting a small house on a shallow foundation. The trees involved are a willow (Salix tortuosa) presently 8.5 m high and growing 7.5 m from the west wall of the house, and an American ash (Fraxinus americana) presently 9 m high, growing 6.5 m from the south wall. While it appeared that the growth of the Salix was not yet affecting the house, desiccation by roots from the Fraxinus were clearly causing the south wall to subside and crack. As a trial remedy, a root barrier was installed between the house and the Fraxinus. Frequent (approximately two-monthly) measurements of soil moisture profiles between the trees and the house have been made over a period of a year and a half. Changes in soil moisture have been linked to the input and output variables for the soil water balance of rainfall, surface infiltration, evapotranspiration, soil water storage and deep infiltration, to both demonstrate the phenomena involved and aid in their understanding.

A companion paper submitted to this conference ("Interaction between trees and buildings on shrinkable soils — a detailed field study") investigates changes in the water content of a soil stratum caused by the activities of trees. The main activity is the abstraction of water from the soil by the trees' root systems. This must take place to provide for the evapotranspiration of water from the trees' branches and leaf canopies, that is necessary for growth. Evapotranspiration from a tree can be measured by considering the tree'ss solar energy balance, because it is the energy absorbed by the tree branches and leaf canopy and the soil penetrated by the tree root system that drives the evapotranspiration and growth (photosynthesis) processes. The paper describes the techniques developed to assess the solar energy balance for a tree and hence enable assessment of evapotranspiration from the tree branches and canopy as well as the soil shaded by the tree. The accuracy is checked by means of the soil-water balance for the grass-tree-soil system.

This study describes geotechnical practice for expansive soils in Texas based on a controlled research survey of Texas geotechnical engineers conducted in April–May 2002. Expansive soils are Texas' defining geotechnical issue; however, the findings of this research show a remarkable amount of uncertainty in the geotechnical community as regards expansive soil movement prediction, analysis and design. The findings address the demographics of Texas geotechnical engineering practice, education in the expansive soil area, the frequency of use for several methods of expansive soil identification and characterization, the validity and procedures for expansive soil laboratory tests, the validity and frequency of use for several expansive soil movement prediction methods, and common practices for expansive soil analysis, reporting, and design. Despite the focus on Texas, the expansive soil practice issues described herein go beyond state boundaries and are of national and international interest with respect to the public health, safety and welfare.

A wide variation in geologic conditions within the North Texas region allows for numerous methods to reduce the potential movement of a ground-supported floor slab constructed over expansive soils. Methods commonly used include isolation, removal and replacement of the expansive soils with inert fill, and preswelling via either mechanical excavation or high-pressure water injection. The preswelled soils are then capped to limit post-construction drying and to increase the subgrade modulus. A brief description of each of the processes used to deal with expansive soils is provided. Specific emphasis is given to the use of the high-pressure water injection process to preswell soils. Typical moisture and suction profiles for varying geologic conditions are provided for background information. An alternative procedure, designated the "picture frame method", consists of construction of an earthen, preswelled "dam" around the building perimeter to isolate the interior portion of the structure from moisture change. Analysis using antidotal information from observation of building performance was utilized to evaluate the width of the required dam. An 18-meter width was used for the case history. The depth varies with the geologic setting. A case history for a 59,130-square meter (sq.m.) office/warehouse is provided.

A manufacturing facility was completed in Loveland, Colorado in early 1993 on the clay shale Pierre Formation. As expansive subsurface materials were identified prior to construction, the facility was founded on drilled shaft foundations with a structurally supported basement underlain by a crawl space. The site was landscaped with bluegrass turf, and there were initially unlined ponds on the up-gradient side of the building. Significant heave of almost 9 inches occurred under portions of the building within three years of occupancy. Early responses included replacement of damaged concrete masonry unit (CMU) partition walls and other cosmetic repairs. In addition, the ponds were lined, the lawn removed adjacent to portions of the building, and deep dewatering wells installed on the up-gradient side of the building. Heave and distress continued in spite of these efforts, progressing slowly across the facility, and currently (2005) impact over three-quarters of the building. Maximum heave is almost 12 inches, and over half the building has heaved more than 6 inches. The locus of the maximum rate-of-heave is moving across the building at about 7 feet per year. Repairs have been performed over time as needed. Lateral heave has sheared a number of the 36-inch and 42-inch drilled shafts, which have been repaired by external banding with post-tensioned cables or carbon fiber wrap. Heave appears to be associated with depths of wetting greater than typically documented in the area. Both construction details and geology appear to have contributed to the heave. Manufacturing has continued uninterrupted with only moderate inconvenience to the building occupants. Interestingly, single-family residences that abut the site have little reported heave damage.

The application of conventional non-swelling soil hydrology to predict wetting depths of submerged swelling soil suggests that equilibrium will only be achieved with moisture profiles having constant saturated moisture content throughout the entire depth. Considerable disparity has been observed between the wetting depths predicted by the conventional approach and equilibrium moisture conditions that have been empirically observed in swelling soil profiles many years after construction is complete. This paper uses well established water retention and volume change properties to introduce the mechanisms governing equilibrium moisture profiles in swelling soils as compared to non-swelling soils. The effort extends the work of previous researchers which has shown that the effect of overburden pressure on the volume change of swelling soils produces equilibrium moisture profiles entirely different from those predicted for non-swelling soil. It is shown that soil can be at equilibrium in both saturated and unsaturated states to great depths. Examples of equilibrium moisture profiles and ground heave in swelling soils are worked out using existing data to highlight the consequences of using non-swelling soil theory to analyse swelling soil hydrology.

The most common foundation for light structures founded on shrink-swell soils is the stiffened slab on grade. The beam depth is an intrinsic design parameter for this type of foundation, it usually ranges between 0.6 and 1.2 m with common beam spacing of 3 to 6 m placed in both directions. The following methods are among the most common methods used to calculate the beam depth, spacing, and reinforcement needed for a safe and serviceable foundation: 1) BRAB Method (1968); 2) WRI Method (1980, 1996); 3) PTI Method (1996); 4) AS 2870 (1996); 5) PTI Method (2004). WRI and PTI Method (2004) are the most recently developed method in the USA while the AS 2870 is the most recent version of the Australian Standards. For the same input design data, applying these three design methods may result in different beam depths. In order to compare these methods, 27 cases are designed. They include three simple rectangular stiffened slabs subjected to three different weather patterns, built on three different shrink-swell soils, and loaded with a uniform pressure and perimeter line load. The resulting beam depths using the WRI, PTI 2004 and the AS 2870 methods in different soil and weather conditions are also presented and discussed.

Several changes have been introduced in going from the 2^nd Edition to the 3^rd Edition of the PTI procedure for the design of slabs-on-grade on expansive soils. The changes in the design procedure, when applied to desert regions, tend to result in doubling of the edge moisture variation distance and a change from center lift to edge lift design governing conditions. The new design parameters along with the change of critical mode of slab deformation lead, in general, to thicker slabs. It is the objective of this paper to examine the changes in the PTI design method and compare the results obtained with both the 2^nd and the 3^rd Edition design procedures for typical Phoenix, Arizona soils and climatic conditions.

This paper presents the analysis of a slab-on-ground foundation influenced by a water line leak and the subsequent soil moisture condition and soil volume changes (shrink/ heave) in the supporting expansive soil. Several possible situations are considered: depth of leak and (1) edge swelling distortion mode resulting from a lengthy wet climate situation following a dry suction profile, (2) edge shrinkage distortion mode (center lift distortion resulting from a lengthy dry climate situation following a wet suction profile), and (3) edge shrinkage distortion mode resulting from a dry climate situation following an equilibrium suction profile. The analyses compares differential movements occurring from the no leak situation to that from a leak located at two different depths beneath the center of the slab foundation. The analysis also compares the resulting differential movements and the induced bending moments in the foundation structure. Some conclusions about the potential for water leaks beneath a slab foundation to cause structural distress to the supported superstructure are offered.

A new 400 MW coal-fired power generating unit was planned to be constructed next to two existing units in the desert of central-eastern Arizona. An expansive clay layer was encountered at depth within the site of the two existing units. The clay layer was stiff to hard, and groundwater was located several hundred meters below the ground surface at the site. However, previous foundation heave experience obtained from another power plant built to the north of the site in the late 1970s/early 1980s led the designers to use drilled piers installed to a depth of about 24 m to support the two existing units. An extensive subsurface investigation program performed at the site of the new unit revealed the presence of alluvial sands and gravelly sands to a depth of about 8 m, where the top of the clay layer was encountered. Based on the results of field and laboratory tests and geotechnical analyses, it was decided to support the new unit using footings and structural mats placed directly on the alluvial soils. The use of these shallow foundations would result in significant schedule and cost savings for the project. Detailed foundation settlement monitoring for more than 1 year during plant construction indicated excellent foundation settlement performance and the adequacy of the selected foundation system.

The evaporative and sedimentary environment that has prevailed over the southern shores of the Arabian Gulf region (eastern Saudi Arabia, Qatar, and United Arab Emirates) has produced salt-encrusted flat areas known in Arabic as "sabkha". Building activities on sabkha areas have posed problems and challenges to the construction industry, triggered primarily by the excessive salts present in both the sediments and their shallow ground waters. In general, sabkha sediments are cemented and uncemented layers of sand /silt material, interbedded with pockets of clay and mud, where calcium carbonate and more recent digenetic minerals (gypsum, anhydrite) serve as the principal cementing agent. The geotechnical aspects of sabkha are addressed with particular reference to building foundations. Simplified soil profiles from selected sabkha sites with Standard Penetration Test Results are shown. The effectiveness of some soil densification methods, as a means of improving engineering properties of sabkha sediments, is touched upon. The paper concludes by offering guidelines for proper field investigation in sabkha terrain, with a geochemical component as an essential part of the geotechnical investigation.

Block anchor is an interface element used to restrain horizontal movement of structures. This paper investigates the effect of moisture conditions (or degree of saturation) on the pullout capacity of block anchor embedded in sand at three different moisture conditions. The approach taken consists of experimental work, and analytical calculation. The experimental work is pullout tests, made in the laboratory, on 0.15 x 0.15 x 0.15m concrete block anchors embedded in sand, at a depth of 0.15m. The sand is deposited in a 1.2 x 0.6 x 0.8m box using a pluviation method to ensure a uniform and reproducible density. Materials used were characterized to find their properties, and the equipments used were calibrated before usage. The load and the corresponding horizontal and vertical displacements were recorded. In addition, visual observations were made on the failed soil body. The experimental results are compared with the analytical calculations. The 3-D effect was also considered. The block anchor was found to have higher pullout capacity than a plate anchor. The results show that the moisture condition significantly affects the pullout capacity of the block anchor. The pullout capacity of the block anchor embedded in unsaturated (wet) sand is about double that for the block embedded in dry sand, while that for the block embedded in saturated sand is only about one half of that for the block embedded in dry sand. These findings have very significant implications in the analysis and design of the block anchor embedded in unsaturated sand. Also, these have contributions to the hazard risk assessment of block anchors embedded in sand subjected to variations in degree of saturation.

A model was developed to predict pavement roughness due to both expansive soils and traffic in terms of serviceability index (SI) and international roughness index (IRI) by correlating the roughness analysis to the vertical movement estimated from a vertical movement model. The total vertical movement, including both swelling and shrinkage, at the edge of pavement sections, the geometry of the pavement, and site conditions were used as model parameters. Total movements calculated at the edge were based on exponential suction envelopes, volume change coefficients, pavement treatments and roadside conditions. Pavement treatments include vertical and horizontal barriers, inert soil and lime- or cement-stabilized layers. The movements in wheel paths at a distance from the edge of pavement are estimated based upon both field observations and the computed results of a transient finite element analysis. A relationship between IRI and SI was developed. The design equations that were developed for both flexible and rigid pavements include the effects of traffic and expansive soil and permit the selection of the desired level of reliability.

Vertical movements calculated by the new design method for each of the six case study cross sections, in three different climatic zones in Texas, were compared with the swelling movement predicted by the PVR method which has been used by Texas Department of Transportation (TxDOT) in the past. The vertical movements included both swelling and shrinking. Pavement treatments were selected to provide an acceptable predicted performance at high levels of reliability, and the vertical movements were calculated both at the edges of the pavements and beneath the outer wheel paths. The case studies were based upon a variety of traffic levels, site conditions, and sample testing. The treatments include lime- and cement-stabilized layers, removal and replacement with inert layers, and vertical and horizontal moisture barriers. The results of the case studies show that the PVR overpredicts the swelling movement that can be expected using the new method, and does not predict shrinkage movement. The resulting evaluation is that the PVR criterion is conservative.

Treatment of expansive soil subgrades is a worldwide technical problem, which needs to be studied and solved urgently due to the rapid development of large-scale highway construction in China. In order to solve the problem, a research group have carried out a field survey in Ningming area. There the construction of Nanning to Youyi Guan Expressway (NYGE) has undergone serious geological disaster caused by Ningming expansive soil. On the basis of worldwide successful and unsuccessful experiences of treating expansive soil projects, the research group analyzed the causes of landslides and conducted a lot of laboratory and field tests on properties of expansive soil. Aiming at the three typical properties of expansive soil and the characteristics of subgrade failure, the group adopted enveloping techniques, and executed some creative schemes to treat slopes of embankments and cuts. All measures taken by research group were successful and have effectively solved the technical problem in the subgrade construction of NYGE. The research outcomes are also useful for railway construction, hydraulic works, and municipal works.

Compressed air tunneling is one of the rare cases in applied unsaturated soil mechanics, where air has to be considered as an active fluid phase. An excess air pressure inside the tunnel is applied to prevent groundwater inflow. As soil and also the shotcrete lining of the tunnel are permeable materials, a transient air flow from the tunnel into the surrounding ground takes place. In order to allow an accurate prediction of the amount of air losses in the design process and to provide pore pressure distributions for coupled flow-deformation analyses, a 3-D numerical tool considering unsaturated multiphase flow has been developed. The results of conducted case studies show a very good agreement with measured air losses for tunnels driven in sandy soils, whereas a big discrepancy for tunnels driven in silty soils was experienced. Recently performed laboratory tests are indicating that permeability increasing macro flow paths are likely to develop in cohesive soils. Therefore, a continuum mechanic based numerical approach to predict the air losses in cohesive soils requires the use of adapted relative permeability functions.

Seasonal change of water level in reservoirs is an important factor causing a slope failure of banks. Unsaturated-saturated seepage analyses were carried out to simulate the changes of phreatic line and pore-water pressure in a bank slope subject to a reservoir level change. On the base of the seepage analysis results and Fredlund's shear strength equation for unsaturated soils, limit equilibrium analyses were performed to evaluate the change in factor of safety (FS) with reservoir level. The analysis results demonstrate that the value of FS in the silty soil slope increases with the rising reservoir level, which is inconsistent with the result found in a literature. The increase of FS is attributed to an increase in the water thrust on the slope surface. During the drawdown process, there exists a critical reservoir level at which the value of FS is the lowest. The critical reservoir level is at a level equal to 1/3 of the slope height.

A large number of slides occurred in the South-Eastern part of Norway during the autumn of year 2000. The slides were triggered by extreme duration of rainfall, leading to saturation of slopes that are partly unsaturated under normal conditions. On this background, a study aiming at establishing a methodology for assessment of stability of slopes prone to rain-induced landslides was started. The suggested methodology for calculation of the effect of rainfall on slope stability is described and applied on a typical example (model slope). Meteorological data together with assumed soil properties are used to define a yearly normal infiltration function and similar functions for extreme events. The normal infiltration function is used to calculate soil water content variation through a normal year using the computer code SEEP/W (2004). Extreme events are analysed by starting calculations from normal conditions. The resulting pore pressures are used to calculate safety factors for slopes during normal and extreme events. Stability of the model slope is calculated using effective stresses based on changes in pore pressures and water content in the undrained and drained zones. Calculations have been performed for the normal year, for the 100-year situation, and for the extreme situation represented by the autumn of year 2000 for both shallow translatory slides and deep circular surfaces. The calculations have demonstrated a significant reduction in safety factors for both deep seated and shallow potential failure surfaces. A tentative conclusion from these calculations is that the safety level (margin) for climatically influenced slopes should be minimum in the order of 0.2 when considering the normal year situation (i.e. FS in a normal year > 1.2). This will ensure that stability of deep-seated shear surfaces do not drop to a critical level during a period of extreme rainfall, like the rain experienced in the South-Eastern part of Norway during the autumn of year 2000. For shallow slides, the safety margin must be evaluated based on top layer material properties.

The paper presents results of studies performed at PUC-Rio to evaluate the mechanism of failure of a natural slope in Nova Friburgo, Rio de Janeiro, Brazil. The slide provoked the death of two people and important material losses. It happened in a profile of unsaturated gneissic residual soil. Failure occurred in a period of low precipitation, a few days after the rupture of a water pipe buried in its top portion. In an attempt to understand the mechanisms that caused failure, field and laboratory investigations were carried out and stability and flow analyses were performed. Flow analyses were executed using a 2D finite element saturated-unsaturated flow program. Resulting pore pressures were used in 2D stability analyses under different scenarios. Also analyses were run to have an assessment of 3D effects on the slope failure. It was verified that failure was caused by the rupture of the water pipe.

Although soils are generally assumed fully saturated below the groundwater table, they may be semi saturated near the state of full saturation under certain conditions. The situation of partial saturation may be caused by several factors, such as variation of water table level due to natural or manmade processes. It has been known in geotechnical engineering that the partial saturation of soil usually gives higher strength and lower compressibility. Additionally, it was shown in the recent studies that unsaturation of soil can also cause much greater amplification in ground motion then a fully saturated model and more resistance to liquefaction of sandy soils. Moreover, the resistance of sand to onset of liquefaction tends to increase with a reduction in the saturation ratio of soil specimens that was expressed in terms of the Skempton's pore pressure coefficient, B. Since the B-value is defined as the ratio of the induced pore water pressure to the applied effective confining stress, it can be simply obtained in the laboratory. The Skempton B-value method has been widely used to determine the state of saturation of laboratory soil specimens. In this study, in order to examine the stress-strain behavior and the strain softening response of partially saturated cohesive soil, a cyclic torsional shear apparatus were used, in which it was possible to perform cyclic shear tests on the specimens under undrained conditions. A series of the shear tests were conducted on the partially saturated-undisturbed clay samples. Therefore, the stress-strain properties were investigated to identify the saturation degree of near-saturated samples. Undisturbed clayey specimens were subjected to multi-staged cyclic torsional shear stress, and deformation characteristics of soils, such as the variations of the shear modulus and damping ratios were evaluated in terms of the saturation ratio. For the cyclic phase of torsional tests, the degradation curves were shown to move the right side with the decreasing Skempton B-value. It was also observed that initial values of the shear modulus were considerably affected the Skempton B-value.

A special laboratory experimental setup is described that permits preparation, and testing of large size partially saturated sand specimens. A unique liquefaction box was designed and built that has two of its walls hinged at the bottom of the box thus allowing rotation and inducing simple shear in the sample. A shaking table is utilized with the box to induce controlled shear strains in the sand specimen. Electrolysis technique was used to generate oxygen and hydrogen gases uniformly within the specimen without causing any change in the density of the loose sand. Experimental test results demonstrate that partially saturated sands have larger strength against liquefaction than fully saturated sands. Induced partial saturation can be a liquefaction mitigation measure. Further research is required to explore the technical and practical feasibility of this potential.

A micro-mechanical study of the sedimentation process and dynamic response of unsaturated granular soils is presented in this paper. The discrete element method was used to idealize the granular soil skeleton. Suction forces were used to model the effects of interparticle water bridges in pendular state. Periodic boundaries and a high gravitational field were used to achieve a realistic simulation of a soil deposit with a computationally manageable number of particles. Numerical simulations were performed to investigate the impact of moisture content on the sedimentation and dynamic response of level ground deposits of unsaturated soils. The outcome of conducted simulations provided a valuable insight into the response mechanisms of these soils.

This paper presents the results of an investigation on the variation of the stiffness at very small strains of an unsaturated clayed soil subjected to drying processes. The soil stiffness was measured by using the wave velocity technique (ultrasonic pulse transmission method and bender elements testing). The test were conduced on samples of kaolinite clay derived from volcanic ash. Those samples were prepared by using a static compaction method known as the flouting mould. Compacted samples of kaolin were allowed to dry gradually up to air-dry condition. At intermediate equilibrium stages, the Shear and Young's modulus and the matrix suction were measured. The effects of initial void ratio, compaction water content and matrix suction on the dynamic response of the soil were investigated. The experimental results suggest that suction significantly increases the stiffness at very small strains. An S-shaped relationship between the initial stiffness and the degree of saturation was observed. The modulus increased significantly at early stages of the drying process, with the changes becoming more gradual as drying continued. However, at final stages, near to air-dry condition, the effect of suction on the stiffness was important again. The initial void ratio appears to have more influence on the stiffness than compaction water content. Contours of equal suction in the plot of the modulus versus degree of saturation, during drying processes, were analyzed. Additionally, changes in the soil fabric during drying and wetting were analyzed considering changes in void ratio.

Construction specifications generally require that subgrade soils be compacted in the field at or near optimum moisture content (OMC). As such, subgrade soils should be treated as unsaturated soils. In addition, moisture content in the pavement components changes over time as a result of environmental and traffic factors. Numerous studies indicated that the moisture content of the subgrade, after construction, would attain an equilibrium with the environment and this is referred to as the equilibrium moisture content (EMC). However, the influence of water content or soil suction on the unsaturated cohesive subgrade soils has not been systematically considered in the conventional subgrade soil design. This paper aims to evaluate the variations of resilient modulus and plastic strain with the post-construction moisture content and soil suction for cohesive subgrade soils. This study carried out the resilient modulus and plastic strain tests via the MTS cyclic triaxial system. To simulate subgrade soils at in-service conditions, soil specimens were compacted at the optimum moisture content, and then placed in the damp room to increase the moisture content to equilibrium condition for resilient modulus and plastic strain test. Based on the above test results, the characteristic of resilient modulus and plastic strain of two unsaturated cohesive subgrade soils were investigated. The experimental results demonstrated that the stress state, moisture content, and soil suction influenced the resilient modulus and the plastic strain. Unsaturated subgrade resilient modulus reduces with increasing deviator stress and decreasing matric suction. The subgrade plastic strain increases as the deviator stress increases and the matric suction decreases. Furthermore, the high subgrade moisture content or low soil suction would result in sharp decrease in subgrade resilient modulus, and should be avoided for pavement design. Finally, based on the concept of effective stress of unsaturated soils, a prediction model incorporating cyclic deviator stress and matric suction for unsaturated cohesive subgrade soils was established. The matric suction of soil proved to be a good prediction variable for resilient modulus. Also, the effects of seasonal variation of moisture content on the resilient modulus of subgrade soils are reflected in the deviator stress-matric suction model for the prediction of resilient modulus.

Time Domain Reflectometry (TDR) is widely applied for measuring soil water content and electrical conductivity. The method works by analyzing the signal responses under the excitations by a small magnitude electrical pulse, which are used to estimate the properties of the soil sensed by the TDR measurement probe. The commonly used method of TDR signal analysis empirically determines the characteristic reflection points. The soil properties obtained with this method are the average values over length of the TDR probe. The variation of material properties with distance along the probe are generally ignored, which otherwise could provide supplementary information of practical importance. The study of vadose zone is an example where the soil moisture varies with depth. This paper presents an approach for obtaining soil property variation within the probe from TDR measurements. It makes use of a forward model that consists of a frequency domain model for a non-uniform TDR system and a simplified model for soil dielectric permittivity. Inversion analysis was used to determine the geometry and soil properties for each layer within the probe. Both the forward model and the inversion method are first illustrated by numerical examples. They are then applied to analyze data from experimental measurements on layered soils. The preliminary results indicate that the framework provides accurate determination of soil properties, including the thickness of each layer as well as the corresponding soil properties. The capability of obtaining material profiles could further enhance the current TDR-based field monitoring system. It could also be incorporated into the existing cone penetrometers to obtain more information on soil properties.

Geophysical down-hole surveys can be used to measure the small strain dynamic properties of soils by the effects these properties have on wave propagation. The relevant effects include amplitude decay (corrected for beam divergence) and velocity dispersion. In this paper, down-hole data collected during the GeoInstitute's Denver 2000 field day are presented and analyzed as a Kelvin-Voigt solid. Findings for these unsaturated soils include viscous damping and stiffness which differ significantly for shear and compressional waves. A strong viscous damping is observed in compression, but weak damping is presented in shear. Lumped parameter constitutive models are discussed which mathematically represent the soil dynamics. It appears that, in the case of unsaturated soils, the relatively low level of viscous damping in shear may be explained by the low mass of the air in the pores. That is, it is difficult for inertial decoupling to occur between the soil frame and the pore fluid when the pore fluid (air) is of such low density. Thus, a pore fluid in coupled motion with the frame can not produce significant viscous drag. On the other hand, large viscous damping is observed for compressional waves. This larger damping may be due to the larger relative motions between air and frame which can be forced by compression of the frame matrix. These observations may be relevant in areas such as the design of driven piles and the estimation of potential for damages from vibrations due to construction.

The application of surface and downhole geophysical methods for site characterization can provide data, information and insight into the subsurface that often cannot be achieved by traditional drilling and sampling techniques. Geophysical measurements applied over a broad area can complement geotechnical sampling methods applied at discrete locations. From a practical perspective, access to areas that cannot be drilled provides subsurface coverage that would otherwise be missing or could only be achieved at extreme cost. This paper presents the results of a comprehensive investigation and characterization program that included the use of state-of-the-art surface seismic microtremor surveys and downhole geophysical methods to complement rotosonic drilling and sampling methods and an innovative method of pressuremeter testing. Refraction microtremor surveys provided continuous simplified vertical shear wave profiles and associated low strain modulus profiles in 36-m (120-foot) increments along an 1,800-m (6,000-foot) alignment profile. Downhole geophysical methods included resistivity logging using an induction electrical sonde to provide vertical subsurface profiles to high resolution (about 0.5 m) that were then correlated back to estimates of clay content in the coarse sand-gravel-cobble (SGC) formation. Geophysical measurements were correlated to downhole pressuremeter test results to verify the relationships between the surface and downhole geophysical methods and in-situ high-strain borehole modulus results. Results from these methods were used to estimate the depth of the Holocene-Pleistocene contact, estimate strength and modulus parameters, and identify the extent and frequency of clean lenses and boulders within the geologic profile. The results of the investigation provided a detailed geologic characterization that was used to optimize tunneling depth, perform engineering analyses for support of excavation and underpinning, and evaluate standup time and tunneling methods.

Unsaturated soil profiles with cementation varying from non-existent to rock-like in nature are typical in the arid and semi-arid regions of the American Southwest. Geotechnical sampling of these materials can be difficult, especially as the materials become more rock-like. Furthermore, excavation of the harder cemented materials may require rock excavation methods, so that effective material characterization is needed for selection of appropriate excavation methods. Geological characterization of cementation strength, excavatability and erodability can be accomplished through various means. Surface seismic methods such as seismic refraction for p-wave velocities and refraction microtremor for s-wave velocities can provide subsurface profiles to effectively characterize geomaterial strengths as well as subsurface profiles. Without a water table to interfere with p-wave velocities, seismic refraction is very effective for obtaining depths to materials with increasing strengths. Velocity reversal conditions can be interpreted from the refraction microtremor s-wave results. Visual observations of stage cementation in test pits excavated using various sized backhoes and trackhoes and other geotechnical test data for dams, highways and other infrastructure are correlated with shallow seismic p- and s-wave velocity results. These correlations provide empirical testing for excavatability as a function of seismic velocity and stage cementation that is consistent with and correlated to the Kirsten excavatability criteria. The Kirsten excavation criteria have been expanded by others to address erodability and scour.

Seismic refraction methods have been used for more than a decade to locate earth fissures and other significant discontinuities in unsaturated, cemented soils that present potential risks to various facilities. Recently, these methods have been adapted to locating areas of potential significant cracking within earthen structures, including large flood control embankments. Earth fissures and connected cracks in normally dry embankments have potential under flood induced erosion to cause failures in such structures or at linear infrastructure such as roads and canals. The essence of these methods is to identify anomalous large signal attenuations and (or) first arrival time delays in standard multi-channel seismic refraction data. Locations of anomalous attenuation or time delays are interpreted in the field and marked on the ground. Using a series of profiles along an earth fissure or other discontinuity trace, actual positions of the features can be determined. Once the locations of such features have been determined, further geologic investigation, including test pits, are typically used to verify and assess the features and potential hazard. These methods have been used to locate and trace earth fissures in Arizona and Nevada, including in urbanized areas of Las Vegas and at industrial ponds and flood control dams.

Covers with capillary barrier effects (CCBEs) can be used as an alternative to more traditional covers that rely on materials with a low saturated hydraulic conductivity. A typical CCBE includes a fine-grained soil layer placed on a coarser material. Other layers can also be added to help the cover play its role(s) efficiently. In a cover built on a flat area, the capillary barrier effect at the interface between the fine and coarse materials allows the finer soil layer to store incoming water, which can later be released by evaporation. Such Store-and-Release covers can be quite convenient in arid and semi-arid conditions. In sloping areas, a CCBE also acts as a lateral water diversion system. Part of this diversion occurs along the sloping interface between the two superimposed soils. It can contribute significantly to the reduction of water percolation deeper into the underlying wastes. However, inclined CCBEs are more complex, as they are influenced by many factors that are not yet fully understood. In such layered covers, moisture is not evenly distributed along the length of the slope. Under some conditions, the moisture-retaining layer can reach a critical degree of saturation at a certain down dip location, which increases infiltration of water into the coarse material and reduces the cover efficiency. This paper presents some of the main results of an ongoing investigation of inclined CCBEs that includes testing on a physical model, field work on large scale covers, and numerical analyses of various cover scenarios. Emphasis is placed here on numerical simulations to assess some of the influence factors that affect the diversion capacity and moisture distribution. The numerical results show how the diversion capacity is affected by the saturated hydraulic conductivity of the fine-grained soil, by the precipitation rate and duration, and by the layer thickness. The findings presented here can be useful for the design of inclined CCBEs.

The infiltration component of the water budget in a soil cover layer is difficult to measure under field conditions. Often, zero-tension (or pan) lysimeters are used for that purpose. Although they are relatively cheap and easy to build, as well as easy to install, their design may be more complex than one thinks. Indeed, the design depends on the unsaturated properties of the soil in which it is installed, in particular the hydraulic conductivity function. The latter is often derived from the water retention curve and the limitations inherent in this process may lead to errors, resulting in the lysimeter being dysfunctional. The walls of a zero-tension lysimeter must be high enough to permit the development of a pressure greater than atmospheric at the base of the container, from where the infiltrated water drains. Using theoretical and numerical approaches, and considering construction constraints (in terms of layer thicknesses), this paper presents a procedure to determine the height of zero-tension lysimeters that were installed in a 300 m² experimental cover (a double capillary barrier constituted of a layer of deinking by-products, sand and gravel) placed above a municipal solid waste facility. Numerical simulations were used to evaluate lysimeter performance.

Assessment of liquid (soil water) flux under unsaturated flow conditions is an important factor in permitting and design of evapotranspirative final cover systems (ET covers) for landfills. Unsaturated flow flux assessments are employed to demonstrate regulatory compliance for permitting of ET covers, to develop construction specifications for approved ET covers, and to establish performance criteria for monitoring of constructed ET covers. Ideally, unsaturated flow flux assessments for ET covers should be based upon site-specific data and should be made using boundary conditions and soil profiles as close as possible to prototype conditions. In practice unsaturated flow flux assessments for ET cover performance have typically been based upon either direct flux measurements from lysimeter test sections or unsaturated flow modeling. The unsaturated flow model may often be calibrated using either percolation and run-off data from lysimeter test pads or internal soil moisture content data from the as-built cover. Lysimeter test pad data are adversely impacted by unrealistic lysimeter boundary conditions, difficulties in accurately measuring small values of percolation, and the inability of unsaturated flow models to accurately predict runoff, while flux values predicted using models calibrated solely upon internal soil moisture content data are of questionable reliability. Therefore, choice of the appropriate approach with which to demonstrate regulatory compliance is often a question of "the lesser of two evils." However, consideration of the limitations of these two approaches suggests that for municipal solid waste landfills in arid and semi-arid climates the use of internal soil moisture measurements from a prototype cover in situ may be preferable for both regulatory compliance evaluation and design of ET covers due to a more accurate simulation of prototype conditions as well as other practical engineering considerations.

An "artificial soil" created from industrial and municipal by-products is evaluated as an evapotranspirative (ET) cover via finite-difference based seepage modeling. Wetting front propagation is simulated under a series of long-term (sustained) and short-term (intense) precipitation events and compared with actual infiltration data obtained from an instrumented test section over a natural 24-day precipitation/evaporation sequence in Northeast Missouri. The predicted depth of the wetting front is 250 cm, which compares reasonably well with moisture data obtained from the field site and is less than the total thickness of the cover (425 cm). Complete breakthrough of the wetting front is predicted not to occur if cumulative precipitation remains less than 16 cm, which is equivalent to an event slightly larger than the 25-year 24-hour storm for Northeast Missouri (14.6 cm). Results indicate that the artificial soil acts as an effective ET cover system by restricting wetting front propagation and causing water to either evaporate or flow laterally. Design charts are developed that may be used to assist in making decisions regarding proposed modifications to the design of the cover system.

The computation of soil-atmosphere water fluxes such as infiltration, evapotranspiration, and runoff is required for the analysis of numerous problems in geotechnical, geoenvironmental engineering and hydrogeology. The soil-atmosphere interaction processes can be represented by a series of partial differential equations. This paper presents a PDE formulation that was developed for soil-atmosphere analysis and presents three cases demonstrating the application of the formulation developed to laboratory and fields conditions. Comparisons against experimental data show that evaporative fluxes can be successfully reproduced by theoretical models. The PDE solutions were used for the simulation of the fluxes through two soil cover configurations to exemplify the application of theoretical models to design. The results indicate that the manner how runoff is computed strongly affects the results. The numerical solutions appear robust and can be applied to the design of soil structures such as soil-cover systems, geo-hazard hazard quantification, and other unsaturated soil problems.

The importance of climate and earth-atmosphere interactions for the design of soil covers is now well-established. Modeling tools have been developed to predict the performance of such covers, based on local climate data, and the flow of subsurface water and heat in either one or two dimensions. These tools are commonly used in the design of soil covers. The application of these design tools carries the assumption that climate will be relatively constant throughout the site under consideration. However, it has been well-established in the field of microclimatology that local climates can be significantly impacted by the topography at a site, particularly with respect to the direction faced by sloped ground surfaces, and the steepness of the ground slope. Presented in this paper is a proposed approach for factoring the impact of slope and exposure into the design of a soil cover, incorporating established models for predicting actual evaporation with a recent model for net radiation on sloped surfaces.

This presentation describes a comprehensive effort to address landfill cover design using innovative numerical estimation methods with site-specific data. The sensitivity of predicted cover performance was evaluated while incrementally changing engineering variables including cover thickness, soil type, and percent plant cover. For each of the 72 different combinations of these variables, a Monte Carlo technique was used to vary soil hydraulic properties for 1000 simulations, with a statistical range determined by laboratory analyses of soil samples collected from the site. The multiple simulations represent a stochastic technique that allows for the variability of measurement results to be included in numerical predictions. Modeling results demonstrated that a minimum thickness for covers of all soil types and the presence of at least a minimum plant cover were necessary to prevent drainage of more than 0.5 cm/yr. Variations in both cover thickness and percent plant cover showed the presence of threshold values for each soil type tested, beyond which incremental additions added little to cover performance. The methods developed demonstrate innovative uses for numerical simulations, which can provide for cost effective cover designs and contribute to improved decision making by improving the technical basis for negotiating final cover systems with regulatory authorities.

A decision framework was developed to identify the optimal cover thickness, soil type, and soil density for an evapotranspirative landfill cover. A simplified model was used to consider the hydraulic performance of several cover alternatives during a design precipitation event, neglecting effects of interflow, surface water runoff, and moisture removal due to evapotranspiration. Monte Carlo simulation was used to estimate the probability of failure of each alternative given a selected failure criterion. Expected costs for each alternative were calculated using the probability of failure and the costs associated with implementation, compaction effort, and failure. The optimal alternative was identified as having the lowest expected cost.

Physical and chemical soil properties including soil particle size distribution, texture, soil water retention characteristics, hydraulic conductivity, and soil productivity influence potential evapotranspiration (ET) cover performance. Borrow source characterization is necessary to define the spatial variability of soil properties, the location(s) of suitable material, and a method to select material during excavation. Understanding the soil environment of a borrow source provides insight for site characterization including soil sampling design and data interpretation. The methods for soil characterization at three sites in the United States are discussed. Soil near Colorado Springs, Colorado developed predominantly from aeolian (loess) deposits and had distinct soil horizons with relatively uniform soil properties. The site soil sampling plan and data interpretation considered soil morphology and was based on collection of discrete soil samples from each soil horizon. Soil near Golden, Colorado developed from fluvial deposits. Horizons from soil forming factors were not present due to prior soil removal (quarry operation) and soils were spatially variable throughout the profile with interbedded fine and coarse-grained horizontal layers. A statistical approach was used to collect soil samples from this fluvial deposit. Composite soil samples were obtained from the vertical face of excavated pits that were randomly selected. Soil characterization was based on analysis of duplicate samples, composite samples that included fine-grained clay clods, and composite samples from which clay clods were removed by sieving. This evaluation was conducted to assess the potential variability of soil properties that impact ET cover performance. Soils for a site near Arlington, Oregon were located in facility stockpiles. Characterization of stockpiles is problematic, particularly if materials are not segregated by soil type. Soil characterization considered the potential variability of site soil types that were excavated for stockpiles.

Monolayer evapotranspiration (ET) covers are the baseline method for closure of disposal sites for low-level radioactive waste (LLW), mixed LLW, and transuranic (TRU) waste at the Nevada Test Site (NTS). The regulatory timeline is typically 1,000 years for LLW and 10,000 years for TRU waste. Covers for such waste have different technical considerations than those with shorter timelines because they are subject to environmental change for longer periods of time, and because the environmental processes are often coupled. To evaluate these changes, four analog sites (approximately 30, 1,000 to 2,000, 7,000 to 12,500, and 125,000 years in age) on the NTS were analyzed to address the early post-institutional control period (the youngest site), the 1,000-year compliance period for disposal of LLW, and the 10,000-year period for TRU waste. Tests included soil texture, structure, and morphology; surface soil infiltration and hydraulic conductivity; vegetation and faunal surveys; and literature reviews. Separate measurements were made in plant undercanopy and intercanopy areas. The results showed a progressive increase in silt and clay content of surface soils with age. Changes in soil texture and structure led to a fivefold decline in saturated hydraulic conductivity in intercanopy areas, but no change in undercanopies, which were subject to bioturbation. These changes may have been responsible for the reduction in total plant cover, most dramatically in intercanopy areas, primarily because more precipitation either runs off the site or is held nearer to the surface where plant roots are less common. The results suggest that covers may evolve over longer timeframes to stable landforms that minimize the need for active maintenance.

Evapotranspiration (ET) landfill cover designs are increasingly being considered for use in the arid western United States, in place of standard regulatory designs that require impermeable layers or synthetic materials. Prior to regulatory approval, ET covers are often required to pass a field test designed to demonstrate adequate performance. Daniel B. Stephens & Associates, Inc. (DBS&A) has conducted field demonstrations of ET covers at several locations using a method that estimates flux in ET covers based on daily soil matric potential values measured in-situ. The method uses heat dissipation sensors (HDSs), soil matric potential/soil moisture content relationships for site-specific soils, and Darcian flux calculations to estimate the rate, volume, and direction of soil moisture movement. Results from two separate multiple-year flux monitoring studies—one where the method worked well and one where it did not—suggest that this method is valid for monitoring flux in covers where relatively low moisture conditions exist, but that it can be highly inaccurate under conditions approaching saturation or where extremely large gradients exist (across wetting fronts or near a capillary break or other impending layer). The uncertainty associated with these measurements is strongly dependent upon site-specific factors including soil properties, magnitude of the hydraulic gradient, and the rate of change of moisture content. These factors produce inaccurate flux calculations that arise from limitations and assumptions of the Darcian flux calculation, rather than improper functioning of the HDS. These studies suggest that the Darcian flux method, although limited in its application, can provide accurate results when applied under the appropriate conditions, at a substantially lower cost than other methods.

This paper reviews the relative merits of two approaches for performance monitoring of alternative earthen final covers (AEFCs): (1) direct measurement of percolation via large-scale pan lysimetry and (2) indirect assessment using data from water-content sensors. Large-scale lysimeters account for spatial variability and preferential flow and can resolve percolation rates on the order of 0.1 mm/yr, which is at least ten times lower than typical percolation criteria applied in practice (1–3 mm/yr). However, factors such as the capillary break effect, thermal water fluxes, and leaks can affect the accuracy of percolation rates measured with lysimeters. Lysimeters are also expensive and difficult to install. In contrast, water-content sensors are relatively inexpensive and are relatively simple to install, permitting multiple monitoring points within a cover at relatively low cost. However, demonstrating compliance using water content data is confounded by several factors, particularly spatial variability, scale effects, pedogenesis, and preferential flow. An analysis conducted using Monte Carlo simulation shows that estimates of percolation rate from water content data can vary by as much as six orders of magnitude due to spatial variability alone. For sites where a percolation criterion has been established, the best approach is to combine lysimetry and water-content sensors. Percolation rates measured with the lysimeter can be used for compliance monitoring and water-content data can be used for interpreting the lysimeter data and to assess reliability and representativeness.

Concern at the long period required for landfilled municipal solid waste to decompose and stabilize in water-deficient climates has led to an extension of the concept of the evapotranspirative or ET cover. The ET cover is designed to store all rain infiltration and re-evapotranspire it during dry weather, but prevent infiltration from entering the waste. With the ISE (Infiltrate-Stabilize-Evapotranspire) cover, rain infiltration during wet weather is stored in the cover and any excess is permitted to enter the underlying waste. This acts to accelerate the decomposition and stabilization process. Excess infiltration is then removed from both waste and cover by evaporation during dry weather. The construction and performance of a set of experimental ISE covers of various thicknesses in a summer rainfall water-deficient climate is described. Field observations of rainfall, soil evaporation and water stored in the covers has allowed water balances to be constructed for the covers of various thicknesses (so far, for a period of 18 months). These observations show that the ISE concept is viable. In the limit, when there is insufficient rainfall to exit the cover and infiltrate the waste, an ISE cover operates as an ET cover.

The reclamation of potentially acid forming waste rock is site specific, being a function, among other factors, of the rock types, the dumping and storage method employed and, importantly, the climatic setting. The "store/release" cover developed to manage acid rock drainage from mineralised waste rock dumps at Kidston Gold Mines' open pit operations in the semi-arid, seasonal, sub-tropical climate of North Queensland, Australia, has been monitored for up to 9 years. The paper describes the philosophy behind the store/release cover design and its adaptation over time to suit Kidston's conditions. The results of monitoring of the store/release cover system over 9 years are presented, and estimates are made of the water balance of the store/release cover for a range of annual rainfall totals and of the overall water balance of the waste rock dumps. The Kidston story is a valuable case study of a successful approach to remediating an identified source of acid rock drainage in a semi-arid climate, which has actively engaged all Stakeholders.

As part of research investigating cover systems on hypersaline tailings deposits in the arid climate of the Eastern Goldfields region in Western Australia, different cover trials have been monitored for salinity up-take into the cover. Among the parameters monitored were moisture content, suction, salinity and pH in the covers and underlying hypersaline tailings, and the success of revegetation. The performance of a large (60 m by 60 m) conventional "continuous" cover has been compared with the performance of small (nominally 10 m in dimension) "discontinuous" cover mounds. The trial covers comprised a 400-mm-thick capillary break layer above the hypersaline tailings, overlain by a minimum 200 mm thickness of growth medium ("topsoil"). Four different capillary break materials were investigated, including coarse quarry rock, hard pit rock, caprock and oxide waste rock. Since the construction of the cover trials, there has been a gradual increase in salinity in each of the cover systems, due to the evaporative (matric suction) and osmotic suction-induced upward migration of salt from the underlying hypersaline tailings. The hard pit rock performed best as a capillary break layer, remaining drier than the other materials used for this purpose, but salt still breached this layer and migrated into the overlying growth medium. Perhaps counter-intuitively, the salinity levels were found to be much higher in the small cover mounds than in the large cover, due to the lateral migration of salts into the small cover mounds from beneath the surface of the surrounding tailings, via the growth medium overlapping the sides of the mounds. Salt uptake into the cover system potentially limits revegetation, as is the case for a natural saltpan. The low regional rainfall further challenges revegetation. The greater salt uptake into the small cover mounds makes their revegetation even more of a challenge.

In May 2004, lysimeters were constructed at Great River Energy's Coal Creek Station to allow a comparative field-scale evaluation of the environmental performance of three earthen final cover designs. Two of the cover profiles use conventional compacted clay hydraulic barriers to limit percolation, while the third profile is designed to store and release water via evapotranspiration. Although a considerable amount of progress has been made in modeling and understanding both types of cover systems in recent years, less attention has been devoted to the challenges introduced when attempting to obtain regulatory acceptance of an alternative final cover system. In this paper, the authors present a case history of the project, with an emphasis on permitting challenges associated with alternative final cover equivalency demonstrations. Cover system field performance to date is also discussed.

The U.S. Department of Energy (DOE) and the U.S. Environmental Protection Agency (EPA) collaborated on the design of an alternative cover for a uranium mill tailings disposal cell at Monticello, Utah. The Monticello design relies on the water storage capacity of a thick, fine-textured soil layer overlying a sand-and-gravel capillary barrier to store precipitation while plants are dormant, and on evapotranspiration (ET) to remove stored water during the growing season. A series of field lysimeter studies were conducted to help design and then monitor the performance of the Monticello cover. Initially, small weighing lysimeters containing mature native vegetation growing in intact monoliths of undisturbed soil provided evidence that an ET cover was a reasonable option even at Monticello, which has an annual average precipitation of 390 mm and a relatively short growing season. Favorable results led to installation of 15 additional small weighing lysimeters to compare water-storage capacities and drainage from different borrow soil types and layer thicknesses. Loam and clay loam soils had the highest water storage capacities. Drainage was less than 3 mm/yr for all treatments during an 8-yr period. Two caisson lysimeters (3-m diameter) were installed adjacent to the construction site to mimic and monitor the range of conditions as built in the actual cover. A loam topsoil compacted to 1.45 g/cm, installed in one lysimeter, had a 40 percent greater water storage capacity than a gravely clay loam subsoil compacted to 1.65 g/cm² installed in the other lysimeter. Only 0.1 mm of drainage occurred in both lysimeters, an amount well below the EPA target of <3.0 mm/yr.

The performance of a prototype reclamation cover constructed over saline-sodic shale overburden was tracked over a six-year period. The test cover, constructed on a 5H:1V slope, was comprised of a thin layer (approximately 20 cm) of a peat-mineral soil mixture overlying a thicker layer (approximately 80 cm) of `secondary' (glacial lacustrine or till). The primary objective of the cover design was to provide sufficient moisture storage for vegetation while mitigating the upward diffusion of salts from the underlying pyritic shale. In situ measurements of saturated hydraulic conductivity were conducted using a Guelph permeameter. Hydrologic measurements included soil moisture, matric suction, soil temperature, surface runoff, and interflow. The major ion chemistry of both surface runoff and interflow waters were measured. In situ measurements of hydraulic conductivity over the six-year period demonstrated that the cover developed secondary structure within four years of placement, likely due to repeated freeze/thaw and wet/dry cycles. The hydrologic response data and interflow chemistry suggests that the mechanisms responsible for the rapid delivery of precipitation to the base of the cover during spring melt are controlled by interactions between soil conditions and climate. Preferential flow occurs when a threshold wetting is achieved and when the ground is either frozen or the matric suctions are low. It also appears that the relative contribution of `event' and `pre-event' water evolves as interflow proceeds. Lateral interflow appears to be an important mechanism for mitigating the upward diffusion of salts into the cover.

Estimation of volume changes in compacted soils is an important aspect for the analysis, design and performance of geotechnical works. It has been observed that several stability problems, involving those materials, are due to water content changes and therefore to matric suction changes that occur periodically in nature. This paper shows the results of an experimental programme carried out to study the volume change behaviour due to wetting processes on compacted kaolin samples. The influence of the dry density, compaction water content and the stress conditions prior to wetting, on the collapsible and expansive behaviour, is discussed paying special attention to the influence of suction. The experimental programme used conventional oedometer and suction controlled oedometer testing. Different wetting stress paths were followed verifying the applicability of an elastoplastic constitutive model for the volume change behaviour of compacted soil.

This paper discusses the volume change characteristics of compacted soil based on oedometer test results. After making specimens with different water contents and under different compaction loads, compression tests are conducted under both saturated and unsaturated conditions. The effects of the compaction conditions on the compression curves are discussed. It is confirmed that the yield stress and the water content are important factors when evaluating a compression curve, while dry density and suction are important factors when evaluating the yield stress.

This paper summarizes the preliminary results of a study on the effects of suction on stiffness and compressibility of ashy layers which cover the Pizzo d'Alvano relief (Sarno-Southern Italy). These soils, similarly to other pyroclastic soils, show some peculiar characteristics, as a remarkable secondary settlement, a metastable structure and a limited range of suction variation. At first, the influence of secondary settlement on the long-term deformability is presented. The influence of suction on the apparent preconsolidation stress and, in general, on the stress-strain behaviour of such soils is also shown by means of suction controlled tests. Finally, the metastable structure of these soil is investigated by means of a suction controlled oedometer apparatus. Results show that a reduced time of load permanence, in oedometer tests, is suitable to determine both the entity of primary settlements and the law of variation of secondary settlements of solid phase. More time is needed for the complete measure of the water content variation. The effect of suction is evident on the apparent preconsolidation stress value and on both the primary compressive index (C_t) and the creep coefficient (C_α). The influence of suction on the deformability parameters of the solid skeleton seems to vanish at high stress levels. Isotropic triaxial tests show a general anisotropic behaviour of the soil within the apparent preconsolidation range although a not so clear influence of suction on this behaviour is observed. Comparing the collapse tests on undisturbed and remoulded specimens, it is possible to identify a metastable structure in the first ones and, on the contrary, a lacking of collapse in the others.

Soil microstructure is a key parameter of the swelling behaviour of fine-grained soils, but the detailed mechanisms of its role are still insufficiently described and understood. The paper proposes a novel microstructural model which permits to quantitatively describe the evolution of aggregate and inter-aggregate void ratios of plastic compacted clays upon wetting. The interpretation of wetting tests on the basis of this microstructural model shows that the macroscopic swelling due to wetting is resulted from an increase of aggregate void ratio and a decrease of inter-aggregate void ratio. Moreover, the inter-aggregate void ratio decrease seems to be proportional to the inter-aggregate void ratio before wetting. Mercury intrusion porosimetry tests performed on samples before and after wetting confirm the relevance of the microstructural model and the wetting tests interpretation.

A good understanding of the water retention properties and volume change behaviour of swelling clays is important for the analysis of hazards related to clays swelling upon wetting or clays shrinking upon drying. In this paper, the results of Romainville clay from the basin of Paris are presented and discussed. The retention curve was determined together with the corresponding volume change behaviour. Two suction control methods were used: the osmotic method for low suction range (1 – 6.1 MPa) and the vapor equilibrium technique for high suction range (6.1 to 216 MPa). Both wetting and drying paths were followed. At each suction, when equilibrium was reached (which is characterized by weight stabilization), soil samples are weighed and their volume determined by using a mercury immersion system. The obtained results showed a good repeatability of the tests, attesting a satisfactory homogeneity of the studied soil. The air entry value 13.1 MPa which is quite high. Significant hysterisis was observed either on water content changes or on void ratio changes. But it clearly appears that this hysteresis only occurs at suctions lower than 131 MPa, beyond this value the soil behavior has been found to be recoverable. This observation confirms that at high suction only recoverable physico-chemical phenomena are involved.

Mixtures of mine waste rock and tailings, tailings alone, and waste rock alone were examined for soil-water characteristic curves (SWCC's) along the drying path. Laboratory data were compared with matric suction measurements and observations from a two year meso-scale column study of self-weight consolidation under free drainage. Waste rock alone had a near zero air-entry value and reached residual water content at low matric suctions. The mixture remained saturated relative to waste rock alone, a finding that was attributed to the presence of a fine-grained tailings matrix. The tailings had a high air-entry value, but also had significant volume change with increasing matric suction. The mixture had less total volume change than tailings for a given change in matric suction. The lower total volume change of mixtures was attributed to the presence of an internal "waste rock skeleton." Both the mixture, and tailings alone had SWCC's with a change in slope that was dominated by changes in volume, rather than a distinct air entry value (AEV), and a volume change correction was required to accurately determine the SWCC. Under prolonged drainage and/or high matric suctions, the mixture developed cracks within the tailings matrix. In general, the results indicate that mixtures can remain tension saturated for long periods of time relative to waste rock alone. Mixing for disposal therefore has the potential to limit the rate of acid rock drainage, particularly for wet climates.

A test on the effect of vertical pressure on the deformation and strength characteristics of undisturbed expansive soil obtained from Ningming (a city in southern Guangxi, China) was conducted during wetting and drying cycles. In addition, the expansion and shrinkage were measured. The testing data indicated that the swelling and shrinking of expansive soil were not completely reversible, and upon repeated wetting and drying, the soil showed signs of fatigue after every cycle resulting in decreased swelling ability. Furthermore, it was noted that the first cycle causes the most reduction in swelling-shrinking potential. As the number of cycles increased, additional reduction was observed until a state of equilibrium was reached. For the Ningming expansive soil, the equilibrium cycle ranged between 3 and 4. Moreover, the vertical pressure in certain range had effects to restrain the swelling-shrinking deformation and the attenuation of shear strength. The higher the vertical pressure, the more obvious the effect was. Therefore, in order to acquire the parameters in accord with the engineering practice, the vertical pressure must be taken into account in the wet-dry cycling test of expansive soils.

Expansive clays in eastern Saudi Arabia are generally fissured and contain high quantities of anhydrous calcium sulfate. Similar to clay minerals, this secondary mineral causes swelling when hydrated and forms gypsum that, in turn, dehydrates and transfers back to anhydrite during compression. The main objective of this paper was to understand the volume change behavior of a local expansive clay containing 50 percent anhydrous calcium sulfate. Swelling and consolidation tests were conducted on undisturbed field samples according to the constant volume method. A conventional oedometer sample was used to determine the maximum possible volume change. Likewise, a large-scale oedometer sample was used to capture the influence of fissuring on volume change. The large-scale sample was thoroughly instrumented to determine both the vertical and the lateral swelling pressure. Results indicated a two-fold increase in swelling pressure of the desiccated clay when corrections were applied to account for sample disturbance. The corrected vertical swelling pressure of the clay was 320 kPa for the conventional sample and 245 kPa for the large-scale sample. The lateral swelling pressure was 40 percent of the vertical swelling pressure and the swelling index was one-third of the compression index for the investigated soil.

Compacted mixtures of sand-bentonite (commonly named Bentonite Enhanced Sand, BES) are widely used to control the movement or migration of hazardous substances from the disposal site to the surrounding environment. Barriers built with those materials exhibit high strength and low compressibility with a very low permeability. Those characteristics can be reached by using mixtures with enough sand (to ensure good stability of the compacted mixture) and bentonite (to fill the voids between the sand particles). This paper shows the results of an experimental study of the volumetric behavior of compacted mixtures of sand-bentonite following wetting under constant vertical stress. The influence of initial compaction conditions, the applied vertical stress and the bentonite content were studied. The experimental program included a series of odometer tests performed with samples compacted at different densities and water contents. The results indicate that the development of swelling deformations with time is highly dependent on the compaction conditions of the BES and the confining stress.

Since the use of bentonite based engineered barriers has been considered for nuclear waste disposal facilities, swelling pressure measurement techniques have been improved. However, some experimental evidence shows that vertical strain control plays an important role on such measurements. In this paper, some results are presented to show that strain control is a key aspect to be considered when measuring swelling pressure.

Results of two long-term infiltration tests conducted in a triaxial cell are presented. Results of the testing are examined to assess mechanical and hydraulic behavior of an unsaturated swelling clay. In one test, radial flow is enforced with water pressure applied at the perimeter and drainage is allowed at the center of the specimen. In the second long-term test, vertical flow is applied with water pressure at the base and drainage is allowed out the top of the specimen. In both cases water flow into specimens and air flow at the downstream side is measured along with volume changes. Suction at the center of the specimen is measured in the vertical flow test using an embedded sensor. Water flow, volume change and suction responses are discussed to assess how flow and boundary conditions affect behavior. Differences in behavior during infiltration are noted due to anisotropy in the compacted specimens. Final distributions of water content, bulk density, dry density and degree of saturation are presented and discussed. Results show that these infiltration tests are important for understanding the physical movement of water through unsaturated swelling clay materials.

This paper describes a methodology to determine the constitutive parameters needed for prediction of heave using oedometer test data. Heave prediction using an effective stress based approach utilizes data from the oedometer test. An important constitutive parameter used in that method is the heave index, C_H. The parameter C_H is the ratio of the percent swell observed in the oedometer test to the vertical stress applied to the sample when it was inundated, i.e. the inundation pressure. This paper outlines methodology for determination of C_H using oedometer test results obtained using reconstituted Pierre shale samples from both consolidation-swell (CS) tests, performed according to the ASTM D 4546 Method A procedure, and constant volume (CV) tests, performed according to the ASTM D 4546 Method C procedure. Most commonly, however, test result data from only one type of test is available, usually the CS test. Therefore, it is useful to establish a relationship between the parameters determined in the CS test and the CV test so that data from only one type of test is necessary. This paper outlines a proposed methodology for correlating test results from CS and CV tests and presents results of a laboratory investigation to assess values of parameters needed in the relationship between swelling pressure from consolidation swell tests and constant volume oedometer tests.

The paper presents the results of a systematic experimental investigation under oedometer conditions aimed at understanding the chemical influence of the water chemistry on the swelling capacity of a statically compacted Ca-bentonite. Specimens were soaked under constant stress (0.02, 0.5 and 2.0 MPa) using distilled water, and NaCl and CaCl₂ aqueous solutions with concentrations of 0.5, 2.0 and 5.5 M. test results showed important mechanical changes induced by the effect of the hydration fluids. The swelling strains of compacted samples on soaking were systematically reduced with increasing solute concentrations. This observation had an important consequence on the development of swelling pressures, which are a key design parameter of the engineered barrier system.

Artificially prepared bentonite-based granular mixtures at very low water content have been proposed as possible candidate materials for the construction of engineered barriers for radioactive waste disposal. This material, which is capable of being emplaced by auger technique, improves the emplacement operations and minimizes air gaps of engineered barriers in horizontal drifts installed in competent clay formations. An extensive laboratory test program was performed to characterize this partially saturated material at nearly emplacement scale and at a low emplacement dry density (around 1.4 Mg/m³) using a small-scale hydration equipment. A constant volume hydration column 70-mm in diameter and 200 mm high was used to simulate different hydration conditions. One of the aims of this paper is to describe the small-scale equipment by detailing its capabilities and presenting selected test results. Sample preparation techniques and main properties of the mixture are also presented.

The Parameter Evaluation Method algorithm is applied to determine parameters of the Barcelona Basic Model and the Blatz and Graham Model for a constant volume infiltration test on an unsaturated swelling clay soil. The limitations of both constitutive models to simulate the stress path for this example are indicated. The recommendation to incorporate different swelling induced pressure for drying and wetting stress paths may be made by introducing parameter κ_rat = κ/κ_s, which represents the degree of the swelling induced pressure. The numerical modeling of this test is also accomplished using finite difference analysis incorporating fluid-mechanical interaction. It shows that the hydraulic constitutive models (soil water characteristic curve (SWCC) and permeability laws) play a significant role for this type of analysis. The future application of the PEM algorithm combined with triaxial tests with controlled and measured suction can reduce the necessity of independent permeability testing to establish hydraulic parameters. The PEM can also be used to create a `parameter generator' to characterize the behavior of unsaturated high plastic clay. Consequently, the number of parameters used in the constitutive model is independent of the difficulty associated with calibrating the parameters using physical measurements.

The state behavior of unsaturated soils is controlled in part by the soil-water characteristic curve (SWCC) and unsaturated permeability. Both these soil properties are dependent on pore geometry. The SWCC exhibits hysteresis, so the wetting and drying curves of the SWCC are different. In this paper, SWCCs of remolded Guangxi and Hefei expansive soils of varying dry densities have been measured and studied. The research shows that the shape of SWCC depends on the dry density of soil samples. The slope of soil-water characteristic curves between the air entry value point and residual water content point will be steeper when the dry density increases. The slope of the soil-water characteristic curve fluctuates for low dry density soil samples and it is stable for higher dry density soil samples. The air entry value of the soil-water characteristic curve increases with the dry density, but the variation of the residual water content is small over the dry density range considered. The SWCC characteristics are correlated to the pore structure. The authors present the volumetric water content (which is defined as the ratio of the volume of water in sample to the total volume of sample) in the sample as a parameter, normalizing a single variable for plotting the soil-water characteristic curves of expansive soil. The volumetric water content can embody the effect of water content, degree of saturation and void ratio of sample.

Soils are naturally subjected to cyclic and strong drying and wetting paths due to atmospheric conditions. Clayey soils subjected to these processes develop swelling and shrinkage deformations, which give rise to the accumulation of expansion or compression strains during suction cycles. In order to study the main features of these processes, a systematic experimental programme was performed on an artificially prepared bentonite-sand mixture using an oedometer cell. Several controlled-suction wetting-drying cycles with suction ranging between 130 MPa and 4 MPa were applied using vapour equilibrium technique. Tests were performed at different vertical net stresses. Along the successive cycles, the tested samples showed cumulative shrinkage strains at increasing vertical net stresses. Test results were interpreted and simulated using the suction-dependent elastoplastic model.

Collapsible soils have porous structure, high void ratios, and relatively low dry unit weights in dry conditions. These soils consist mainly of wind-blown silt. Collapsible soils are moisture sensitive and susceptible to large reductions in volume primarily, with increase in moisture content. In this investigation, a commonly found collapsible soil, known as loess was tested on the laboratory compacted specimens to study its volume change behavior. The collapse tests were conducted using conventional oedometer. The collapse potential of loess was found to be dependent on the initial water content, initial dry unit weight and flooding stress (vertical stress at inundation).

Volume decrease associated to an increase of water content of a soil, at almost constant total stress has been termed collapse. Some evidences from laboratory tests and field observation indicate that soil collapse can arise from moisture variation or, more appropriately, from suction variation conditions that are far from saturation or far from zero suction. This paper deals with the influence of soil-water characteristic curve of an undisturbed sandy soil on collapse deformation. The results analyzed come from suction controlled oedometer tests and soil-water retention curves. The results show some features of the gradual collapse behavior of the tested soil and its relationship with the form and parameters of soil-water retention curve and presents an empirical equation able to calculate the collapse potential development under matric suction variations using the parameters of the soil water retention curve.

Collapse and direct shear box tests on an unsaturated compacted soil with suction and water content measurements due to soaking are first presented. The compressibility, strength and dilatancy characteristics are discussed on the basis of the experimental results, paying attention to the initial water contents and the normal stresses. Then, a simplified relationship for rationally predicting the collapse settlement under one-dimensional compression condition is presented based on the experimental considerations. The applicability of the empirical relationship presented is finally discussed by comparison with the experimental data obtained by one-dimensional compression tests.

Laboratory infiltration tests have been performed to predict the extent of wetting zone within a collapsible soil. During these tests, the rainfall water was simulated in laboratory to investigate the infiltration of rain water around and beneath a footing model resting on collapsible soil. Distribution of soil water content beneath and around the footing model was assessed. Consequently, accurate details of the progression of wetting zone front were also measured. The effect of the relative density of the formed soil, the initial water content of soil, and the intensity of induced water are studied. A relationship between depth of wetting zone and time was established, and a proposed mathematical formula for predicting the optimum depth of wetting zone is presented.

This paper presents an investigation of the potential of collapsing soil or loess from Thailand for possible use as a pavement material or subbase. Earlier studies indicated high CBR values for well-compacted material. The collapsing soil is classified as silty sand, with a slight to moderate degree of collapse index. Thai collapsible soil samples were tested generally in the compacted condition (Modified compaction) at optimum moisture content and maximum dry density. Consolidated Isotropic Undrained (CIU) triaxial testing with pore water pressure monitoring was performed on back-saturated samples to quantify the permissible stress states for resilient modulus testing. The resilient modulus of the soil was found to be stress dependent and trends agreed reasonably well with the current AASHTO empirical equation for the resilient modulus of sand.

Basic phenomena involved in crack generation during drying, and the ways to control or avoid it, are still elusive. Mechanisms of shrinkage and cracking are addressed, and especially the critical role in crack onset during drying of boundary constraints and induced tensile stresses. Two kinds of experimental tests were carried out on samples of fine-grained soil under controlled conditions: free desiccation tests and constrained desiccation tests. Measurements of strains, gravimetric water content and suction allow the drawing of some conclusions about the influence and the evolution of these tensile stresses during drying.

Samples of two clays and two marls consolidated or reconstituted to various initial void ratios were left to dry to constant volume and the limiting lower value of the void ratio at that state, e_r, was measured. Drying of the initially fully saturated samples of each soil was achieved by leaving the samples under atmospheric conditions. The data presented suggest, at first glance, that limiting lower void ratio after drying, e_r, for each soil increases as the initial void ratio of the soil increases. Normalizing e_r by the initial void ratio results in the ratio e_r/e_o increasing as the initial void ratio decreases. Plotting the ratio e_r/e_o for all soils against initial void ratio normalized by the void ratio at the liquid limit e_L (= w_L⋅G_s) shows that the results from all soils lie very close to a unique line, which passes through point [e_o/e_L=0, e_r/e_o=1] and has a very high correlation coefficient. This linear relation appears valid for values of e_o/e_L between 0.6 and 1.7. Using the relation between e_r/e_o and e_o/e_L the limiting lower void ratio after drying can be predicted when the initial void ratio of the soil e_o (expressing the initial state of the soil) and its liquid limit w_L and specific gravity G_s (expressing the nature of the soil) are known. Also, according to this relation, e_r exhibits a maximum value as e_o increases and then starts to decrease again as e_o increases further. The empirical relation derived may be used for the prediction of a mechanical parameter of use to modern constitutive modeling of unsaturated soils such as the limiting lower void ratio after drying. It may also be used as a means to study unsaturated soil behavior as far as the evolution of e_r with initial void ratio and physical properties of soils is concerned.

In many parts of the world the shrink-swell characteristics of fine-grained material is of considerable importance and of potential economic significance. Expansive soils cause significant damage to structures and roadways by cyclically shrinking and swelling within the active zone, which is defined as the depth in a soil to which periodic changes of moisture occurs. In the U.S., expansive soils cover large parts of Texas, Oklahoma and the upper Missouri Valley, and are typically montmorillonitic in nature. Traditionally there are two ways to identify the shrink-swell potential of the clay deposit: measure the shrinkage characteristics or measure the swelling characteristics. This paper discusses the shrinkage characteristics of fine-grained soils. Four natural and pure clays were tested to determine shrinkage characteristics and the respective shrinkage curves are presented and discussed. The direct measurement of the limit of shrinkage from the shrinkage curves of the Linear Shrinkage and Shrinkage Limit tests are discussed and compared to the calculated Shrinkage Limit from the ASTM D-427 Shrinkage Limit test. It is shown that the Linear Shrinkage and the Shrinkage Limit tests produce similar shrinkage curves and a consistent Limit of Shrinkage. The calculated Shrinkage Limit from the Shrinkage Limit test underpredicts the limit of shrinkage for the soils tested and is extremely operator dependent.

A series of unconfined undrained triaxial tests were performed, on compacted and reconstituted samples of clayey soils, at different degrees of saturation. Pore water pressure changes were recorded. Stress-strain-suction relationships were studied and analyzed. The soil water retention curve (SWRC) or any scanning curve exhibit an S-inverted shape. At low suctions the curve has a small gradient which increases and tends to be constant at intermediate suctions. Two different stress-strain-suction behaviour patterns were identified depending on the initial position (before shearing) of the soil sample on the SWRC. Samples initially located in the zone of small gradient, of the SWRC, exhibited a stress-strain-suction behaviour which is similar to that of normally consolidated samples (i.e. strain hardening), whereas samples initially in the higher gradient zone, of the SWRC, exhibited a similar behaviour to that of over-consolidated samples (i.e. dilatancy occurred). On shearing, the behaviour of the soil at peak conditions was not dependant of the initial fabric. There was no influence of the suction history of the sample before shearing.

This paper presents an experimental investigation of the small-strain stiffness behavior of an unsaturated compacted subgrade soil. The small-strain stiffness tests were performed under two testing schemes, namely as-compacted state (at the initial compaction condition during construction) and post-compacted state (service life after construction when subject to moisture regime changes). A non-destructive elastic wave propagation technique using bender elements was utilized to monitor and to evaluate the shear wave velocity and corresponding small-strain shear modulus (G_o) of the compacted soil specimens. A new test apparatus was designed to control two stress state variables, the net confining pressure and the matric suction, during the G_o measurements. Finally, a model was proposed for the small-strain stiffness behavior of unsaturated compacted subgrade soil.

Mudflow type failure of a gentle fill slope occurred at Dateshita in Tsukidate town during the 2003 Sanriku-Minami earthquake in Japan. The flowed fill material was pyroclastic sediment that was classified as a volcanic sandy soil with pumice. This study reports the results of cyclic shear triaxial tests of unsaturated volcanic sandy soil considering suction. Tests were conducted for specimens with different initial suctions and the same dry density under undrained conditions for both air and water. Volcanic sandy soil specimens lost their effective stress under cyclic shear loading even if the degree of saturation is about 70 percent. If the initial suction is lower than the air entry value (AEV) of the soil-water characteristic curve, the effective stress of the specimens became zero or an extremely small value, and stiffness decreased. On the other hand, stiffness did not decrease in cases where the initial suction was higher than the AEV of the soil-water characteristic curve. It is considered that the variation of air distribution pattern in specimens below under the AEV and above the AEV causes the apparent difference in cyclic behaviors.

This paper is focused on the results of an experimental study of the effects of mean net stress (p-u_a) and suction (u_a-u_w) history on the initial shear stiffness G_o of a compacted silt. Isotropic tests were performed using two suction controlled devices, a triaxial cell and a Resonant Column Torsional Shear (RCTS) cell, so as to investigate the volumetric behavior of this material. In addition, by using the RCTS cell, initial shear stiffness was measured almost continuously along different isotropic stress paths in the plane (p-u_a):(u_a-u_w), including constant suction- compressions and single stages or cycles of drying and wetting. The obtained results show a strong dependency of G_o on mean net stress and suction. Cycles of suction, in particular increasing suction beyond the past maximum value, induce significant accumulation of irreversible strains and increase of stiffness, confirming that G_o is not univocally linked to the stress state expressed in terms of (p-u_a) and (u_a-u_w). The influence that mean net stress — suction history has on initial shear stiffness is analyzed and ascribed to the accumulation of irreversible volumetric strains. A model able to predict the observed behavior is proposed.

Matric suction and cementation together have significant influence on the stress-strain behavior of many lightly cemented soils. The combined effects of suction and cementation during soil formation may stabilize the fabric of the soil at a very high void ratio. This is the case of wind-blown loess of the Pampas formation in Argentina. The scope of this work is to present some fundamental aspects related to the unsaturated stress-strain behavior of lightly cemented silty clay. Triaxial test were performed in laboratory and strains were monitored by means of local displacement transducers. Undisturbed structured specimens as well as remolded unstructured specimens were tested both in saturated and unsaturated condition. It was observed that cementation and unsaturation cause an increase in yielding stress and shear strength, decrease in volume contraction during loading, and a higher propensity to stress localization. These effects are mainly controlled by water content and confining pressure. At low confinements, the unsaturated and cemented specimens behave more brittle as compared to the uncemented specimens due to the much lower deformation threshold required to break cementing bonds than menisci. A bilinear model is proposed to model the fragile behavior of cemented specimens.

Compacted soils are often used in engineering structures such as pavement, backfills and embankments. The stiffness of the compacted soils is affected by dry density, degree of saturation and applied strain. This paper describes the stiffness-strain relationships of a compacted residual soil from undrained triaxial compression tests. The triaxial apparatus is equipped with load transducer, bender elements, local displacement transducers, linear variable differential transducer and proximity transducers. Therefore, stiffness from very small strain to large strain can be measured. The small strain stiffness is important in pavements where deformations in the subgrade can be small whereas the larger strain stiffness is applicable to retaining walls and foundations. The effects of initial matric suction and confining pressure on the stiffness-strain relationships are examined in the paper. Both initial matric suction and confining pressure affect the very small strain stiffness, G_max. However, only initial matric suction affects the entire stiffness-strain relationship of the compacted residual soil.

The shear strength of two colluvium soils, situated at the Serra Geral Formation scarp in Southern Brazil, was determined with conventional and suction-controlled direct shear tests. One investigated colluvium soil was originated from the weathering of a volcanic rock (Sample RO) and the other, possibly from the degradation of intertrapp sandstone layers (Sample AV). A study of matric suction influence on the shear strength of undisturbed samples of both soils is presented in this paper. Suction-controlled direct shear tests showed that shear strength envelopes are linear for matric suctions between 0 and 150 kPa. Further direct shear tests, carried out with air-dried samples, showed that those shear strength envelopes tend to become bilinear for higher suctions. In addition, the experimental data indicated that for both soils the internal friction angle seems to increase with matric suction and values are higher than corresponding values.

A conventional triaxial equipment, originally used for testing soil specimens in the saturated condition, was modified in order to allow the testing of specimens in the unsaturated drained condition. In this device, the axis-translation technique is used to avoid cavitation in the pressure measuring system. The volume change in the specimen is determined from the measurement of changes in the volume of the triaxial cell confining fluid. Those tests were carried out using specimens of a colluvium soil of sandstone both in the saturated condition (drained and undrained tests) and in the unsaturated condition. The observed behavior was (i) an increase of shear strength with an increment of matric suction and (ii) the non-linearity of the shear strength envelope after the application of a higher value of matric suction.

The shear strength of unsaturated soils is a subject which has received a lot of interest recently. Experimental results have been used to define mathematical formulations, but the data available on tropical soils is quite scarce. This paper presents the results of triaxial tests on unsaturated lateritic and saprolitic soils with the same rock mother. The variation of shear strength parameters with suction was analyzed. Air dried specimens tested with constant water content were used to define the maximum shear strength.

Shear strength of unsaturated soil is commonly obtained from Consolidated Drained (CD) or Constant Water content (CW) triaxial tests. A series of CD and CW tests was carried out on statically compacted silt specimens. The results indicate that the effective angles of internal friction, φ', and the effective cohesions, c', of the compacted silt as obtained from both the CD and CW tests are identical. The φ^b angles from the CD and CW triaxial tests are essentially identical at matric suctions lower than the air-entry value. This could be attributed to the fact that the soil specimens remain saturated at matric suctions lower than the air-entry value in both the CD and CW tests. Similarly the φ^b angles from both tests are the same at matric suctions higher than the residual matric suction when the soil is at a relatively dry condition. However, the φ^b angles from the CD and CW tests are different at matric suctions between the air-entry value and the residual matric suction value. The difference is due to the lower failure envelope with respect to matric suction from the CW tests as compared to the failure envelope from the CD tests. This difference can be attributed to the hysteretic behavior of soil-water characteristic curve of the soil which will be explained in detail in the paper.

Aiming at the examination whether two factors of the net normal stress and the suction can express the shear strength of unsaturated soil under different conditions of drainage, triaxial tests were conducted on unsaturated specimens of a clayey silt under drained and undrained conditions. In drained tests with the pore-air pressure u_a and the suction s held constant (D-tests), the volume contraction occurred as well as the draining of the pore water. Two kinds of undrained tests were conducted with u_a regulated. In undrained tests with u_a constant (Ua-tests), s decreased during shear and the shear strengths became lower those in D-tests. Another kind of undrained tests in which u_a was regulated in order to keep s constant (Us-tests) resulted in simultaneous increases in u_a and u_w, and the state of stresses reached one corresponding to the unconfined compression. Shear strengths from Us-tests were much lower than those from D- and Ua-tests. Measured shear strengths were analyzed using an expression for the shear strength as a linear function of s and the net mean normal stress; results showed that, although the expression would be satisfactorily accepted, the effects of the difference of the testing conditions were detected.

This paper is concerned with the shear strength of undisturbed natural samples of volcanic pumiceous soil. The shear strength varies in a complicated manner with the normal stress lower than the consolidation yield stress p_c. Considering the suction as a factor affecting the shear strength characteristics, direct shear box tests were conducted after the consolidation during which the suction was measured and regulated. When consolidation pressures higher than p_c were applied, the suction decreased, eventually, to null with almost full degree of saturation; the shear strength with such high normal stresses obeyed the Mohr Coulomb criterion. To some specimens the suction was regulated with a normal stress lower than p_c; they mobilized shear strengths exhibiting a good correlation with the suction. Conclusions emphasize that the suction should be measured for unsaturated specimens even if they have so high degree of saturation as 80 percent or more, and that the suction will be effective for interpreting the variation of the shear strength, especially determined with relatively low normal stresses, of undisturbed natural samples.

Shear strength properties of unsaturated soils are important in evaluating the stability of geotechnical structures such as natural slopes, embankments, retaining walls, excavations, and footings. This importance is more in countries, where intense rainfalls cause the instability of natural and man-made slopes. An unsaturated soil behaves differently at wetting and drying and this difference in behavior is referred to as hysteresis. Hysteresis is often exhibited in soil-water characteristic curves of unsaturated soils. The hysteresis in the shear strength behavior of unsaturated soil under wetting and drying has not been fully investigated. Shear strength tests on unsaturated soils have usually been conducted following the drying process. For example, in a multi-stage shear test, the soil is subjected to increasing matric suctions at subsequent stages. Very few studies have been conducted on the unsaturated shear strength parameters on wetting process by which slopes become more unstable. Re-compacted soil specimens of silt-sand were tested in a modified triaxial apparatus to examine the effects of wetting and drying on shear strength. Identical soil specimens were tested at wetting and drying under controlled air and water pressures. The study was conducted for low suction range, (i.e. 0–50 kPa). The soil at the drying had lower shear strength as compared to the soil at the wetting under the same matric suction. Furthermore, the study was extended to investigate the effects of suction and the hysteresis of SWCC on shear strength parameters such as effective friction angle and the apparent cohesion, c. Results exhibited no significant effects of suction and the hysteresis of SWCC on the effective friction angle. It was observed that the apparent cohesion at the wetting was higher than that was obtained at the drying under the same suction. The c increases as the suction increase with the decreasing rate at wetting.

Considerable attempts have been made in recent years to develop a better understanding of the mechanical behavior of unsaturated soils in terms of elasto-plastic critical state constitutive models. An important feature of the models is the suggestion of the existence of a yield surface in mean net stress: deviator stress: suction space. This paper presents the results of experimental tests for investigating the shape of yield surface for compacted silty soils. A comprehensive set of laboratory experiments has been undertaken in a double-walled triaxial cell. The method of axis translation was used to create the desired suction in the samples. The results show that constant suction cross sections of the yield surface were approximately elliptical in shape, and inclined to the axes; the size of the elliptical yield curves increased with increasing suction and the flow rule was non-associated.

Shear strength characteristics of unsaturated soils under low confining pressure are important in order to analyze the mechanism of slope failures induced by rainfall. In this paper, we studied the results of unconfined compression test and constant water content compression test on silt specimen under low confining pressure performed in the past and examined the residual strength and deformation characteristics. The residual stress states of specimens having different degrees of saturation and matric suctions at failure show a unique relationship for the confining pressure when the suction stress is considered. This means that "suction stress" affects the residual strength considerably. It is also found that when the specimens have nearly the same initial void ratio and expanded in shear process, critical state lines parallel to the original one is expected. From these results, it is inferred that in some cases, the extended Cam clay models proposed need to be modified when they are applied for low confining stress state.

In this paper, tests conducted on a low plasticity clay in a Bromhead ring shear apparatus adapted to allow for suction control during the shearing phase are presented. Conclusions are drawn concerning the effect of suction on the residual strength of the material. Apparatus and test procedure are firstly detailed. Afterwards, a series of single and multi-stage shear tests performed on samples saturated, dried in one step, dried in several steps or wetted in several steps are described. It was concluded that suction does not introduce any cohesion-like component of strength but rather produces an increase in friction angle, possibly due to particle aggregation during drying.

Laboratory tests performed on unsaturated kaolinite clay show that the shear strength is function of the water content. Water contents used for the tests were obtained by drying the clay. Three different intervals of water contents in the clay were identified for which the shear strength varies in different way. When the clay has high water content, close to saturation, the shear strength increases as the water content is reduced. If the clay has a water content in an intermediate interval the shear strength is almost constant. For the interval of low water contents the clay presents lower shear strength as the water content is reduced, which is the opposite tendency for the interval of water contents that is close to saturation. In this article, the three water content intervals are related to three water distribution states of the water in the pores namely: the saturated-funicular state, the complete pendular state and the partial pendular state. On the other hand, the shear strength in the soil is represented as a function of the equivalent effective stress and the friction angle in the clay. The equivalent effective stress is determined through direct and indirect tensile tests. The friction angle that results from the above relationship is very similar to the friction angle of the clay when it is saturated.

It is generally recognized that unsaturated soil behaviour and properties are governed by two stress state variables: soil suction and net stress. Although a soil-water characteristic curve (SWCC) can provide a measurement of the water retention ability of an unsaturated soil, very often only a single stress state variable, i.e. soil suction, is considered in experimental programs. In addition, the volume changes of soil specimens during the determination of SWCC have often been ignored. Some recent experimental studies have demonstrated that the other stress state variable, i.e. net stress, is also important and needs to be considered. Ignoring the effects of net stress and volume changes on measuring any soil property of an unsaturated soil may impose severe constraints in the development and verification of advanced constitutive soil models and hence may affect the accuracy of numerical predictions significantly. In this study, a modified volumetric pressure plate extractor was used to determine the state-dependent soil-water characteristic curve (SDSWCC) of a completely decomposed granite (CDG) with due consideration of volume changes. Intact block samples of CDG were obtained from a fill slope in Hong Kong. In the laboratory, both intact and recompacted specimens were tested and subjected to different stress levels, together with one cycle of drying and wetting paths. The measured results from the CDG specimens are compared with the previous test results of a completely decomposed volcanic tuff (CDV) in Hong Kong. Differences in the observed soil-water characteristics are examined and discussed.

This study focused on the effects of temperature on the suction characteristic curve of bentonite-sand mixtures. Mixtures of bentonite (Calcigel) and quartz sand were used. Three different suction measurement methods (i.e., chilled-mirror hygrometer, non warmed-capacitance sensor, and warmed-head capacitance sensors) were used for measuring total suction. A special device was developed to maintain the specimen conditions during test at a temperature up to 80 degrees C. The procedures adopted in the study are described in this paper. The total suction of the specimens measured by the different methods both at 20 degrees C and 80 degrees C are compared. Factors affecting the alteration of suction due to temperature change are discussed in the paper.

Soil water retention is important for characterizing the storage within and the rate at which water moves through a granular material under both saturated and unsaturated conditions. This study characterized the drying curves of 18 compacted aggregate base and granular subbase materials used in pavement construction in Minnesota. In general, drying curves at matric suction greater than 30 cm were within a narrow range of water contents for most materials. This is expected considering that particle size distribution that contributes to water retention (<2 mm) was nearly similar for most samples. Suctions less than 30 cm represented the air entry values of many materials. Pedo-transfer functions showed that water retention properties of roadbed materials could be explained by differences in sand content and dry bulk density of the samples. Correlation coefficients between the water retention function parameters of van Genuchten, Brook and Corey, or Fredlund and Xing and the particle size distribution were good. However, the correlation coefficients of above function parameters versus percent particles passing #200, D10, D60, or the Minnesota's grading number were poor suggesting that complete water retention curve can not be described by a single parameter of the particle size distribution. Existing empirical and physico-empirical models also did not predict well the water retention properties of the roadbed materials because these materials were highly compacted (up to 2.20 Mg m^–3) and were low in clay content. Since aggregate base and subbase materials contain large aggregates that do not contribute to water retention but strongly affect saturated hydraulic conductivities, we point out the difficulties of using water retention characteristics along with saturated hydraulic conductivity to predict unsaturated hydraulic conductivities of these materials.

This paper presents a series of experimental results obtained from wetting and/or drying cyclic tests and constant suction triaxial tests on unsaturated compacted clay with different initial densities using two suction-controllable triaxial apparatus. The primary objective of this research is to investigate the influence of the soil density on the soil-water-retention behaviour. The initial density and initial degree of saturation were controlled by changing the compaction energy using clay powders with almost the same water content. The test results contain the water-retention curves and deformation behavior at different initial densities using different specimens and one specimen before and after collapse. The test data indicate that the measured water-retention curves vary with the current specimen density. A denser specimen results in a higher degree of saturation at the same imposed suction, which implies that the water-retention curve shifts to the right in the ln s-S_r plane.

This paper summarizes a database of sixteen measured soil-water retention curves (SWRC) of compacted soils and their conventional properties from the published literature. In addition, six SWRC were determined for a compacted clay from Ottawa region using different initial compaction water contents and compaction efforts. Using the above summarized information, relationships were developed between the parameters of the SWRC and the conventional soil properties. This study shows that liquid limit coupled with clay fraction can be a useful tool to correlate with the rate of desorption for compacted soils. In addition, a new technique is also proposed for the estimation of the entire SWRC of compacted fine-grained soils using one point measurement of matric suction versus water content and the derived relationship between the liquid limit and the clay fraction. The proposed method provides good comparison between the measured and estimated SWRCs for the compacted fine-grained soils with a relatively rigid pore structure.

Compaction during construction will reduce the overall pore volume of soil and change the size and distribution of the pores within the soil. The resulting pore space changes are expected to change the moisture characteristic of the soil. The soil moisture characteristic has been examined using experiments where a single soil type was tested for a range of porosity values. Moisture characteristic data recorded in published papers and measured at the University of New Mexico were evaluated. A new term, the Moisture-Density Variation Ratio, was used to quantify the relationship between the dry bulk density and the unsaturated soil moisture. The test data was used to predict changes in the soil moisture characteristic with changes in the porosity. A method is presented to modify the parameters in the van Genuchten soil moisture characteristic equation with changes in porosity or dry bulk density.

The water retention properties of a chalk from an abandoned underground mine in Northern France were studied and related to changes in relative humidity (RH) in the mine in an attempt to better predict the long term stability of the mine. Chalk microstructure was investigated by scanning electron microscopy (SEM) and mercury intrusion porosimetry (MIP). It has been shown that some experimental techniques commonly employed for controlling and measuring suction in unsaturated soils can be easily used to determine the retention properties of chalk. The comparisons between the results obtained using these techniques allowed to have an insight into the effect of mineralogy and pore morphology on the retention properties of the chalk.

Experimental research was conducted to evaluate the effects of salt content and osmotic suction on suction measurements, moisture retention, and shrinkage behavior of a natural compacted clay soil. Results are presented for: (1) soil water characteristic curves (SWCCs) in terms of matric, total, and osmotic suction, and (2) unconfined shrinkage tests on natural and salt-amended compacted soil specimens. These experimental results show that the measurement of total suction is dominated by osmotic suction, even when only moderate amounts of salt are present. Although osmotic suction dominates the total suction measurement, its influence on soil behavior is relatively small compared to the effects of changes in matric suction. It is common practice in engineering to use total suction as a substitute for matric suction, because total suction can be readily measured over the entire range of soil moisture up to 10⁶ kPa using psychrometric techniques. Matric suction is measurable only up to approximately 1500 kPa maximum. Theory and experiments described in this paper are intended to illustrate how the use of total suction as a proxy for matric suction may be problematic when osmotic suction is large due to the presence of salts in the soil.

The soil water retention curve of a soil is usually not unique. It depends on the soil water flow process. Fully primary curves for draining and wetting processes are necessary in soil water flow modeling. Usually, the primary draining curve can be produced from an experiment with a hanging column apparatus together with a pressure plate apparatus. However, the pressure plate cannot supply the necessary data for the wetting process; therefore only the scanning wetting curve can be derived from the hanging column. By using the primary draining curve and the wetting scanning curve as guides, the primary wetting curve can be constructed. Six saline soil samples from around Khon Kaen City in northeastern Thailand were used for testing. The hystereses of the soil water retention curves were found for all soil samples with varying degrees of hysteresis. The degree of hysteresis was found to vary with the percentage of silt portion of the soil sample.

The objective of the paper is to compare drying, wetting, and suction characteristic curves of a bentonite-sand mixture. A mixture, of 50 percent bentonite and 50 percent sand was prepared and used in this investigation. The main drying curve of the mixture was determined for a slurried specimen of the mixture using a chilled-mirror hygrometer. Several drying and wetting tests were also performed on the compacted mixture with an initial water content of about 9 percent and initial dry density of about 2 Mg/m³ incorporating different suction paths using a desiccator and a pressure plate apparatus. The suction characteristic curve for the mixture was obtained at loose as well as compacted states by utilizing the chilled-mirror hygrometer. The experimental results obtained are presented and compared. The drying, wetting, and suction characteristic curves of the material are different. Factors causing the differences including suction paths, hysteresis, and fabric effects are also discussed in this paper.

This paper presents data on the water retention characteristic properties of some typical eastern Australian bituminous coals. "Soil water" characteristic curves, comprising wetting and drying paths, derived using filter-paper methods are presented and compared. The water retention properties of processed coal differ considerably from those of most soils, due primarily to the significant number of hydrophobic sites on coal particle surfaces. The properties of different coals are also found to vary widely, according to the rank or carbon content of the coal. In general, matric suctions in coal are very low for a large range in water content below saturation, and become significant only at water contents below about 50–70 percent of the saturated value. The difference between total and matric suctions is significant at all water contents, and it increases with reducing moisture content.

The measurement of the Soil-Water Characteristic Curve (SWCC) is an important initial step in applications that require knowledge of the partially saturated soil behavior. The relationship between the pore-water suction and the soil water content is known as the SWCC. This relationship is fundamental to the description of the behavior of unsaturated soils. SWCCs are hysteretic, with the drying limb plotting above the wetting limb. Hysteresis is generally occurring due to entrapped air inside the pores. In this study, only the drying limb of the SWCC was measured. The standard method in determining the SWCC relies on the pressure plate (PP) apparatus (D3152-72, 1994). The PP apparatus is generally used to determine the equilibrium water content for various pore-water suctions allowing the construction of SWCC. The PP methodology is time consuming, expensive and elaborate. Thermocouple psychrometer (TCP) provides an alternative device to the PP for the SWCC measurement. TCPs are a simple and reliable method in the measurement of the soil total suction in the field and in the laboratory. Previous studies showed that the use of TCPs in the measurement of total suction in order to construct the SWCC of compacted fine-grained soils is acceptable. In the current investigation, an evaporation scheme that utilizes TCPs was used in the determination of SWCC. This scheme is referred to as the "Evaporation Technique." This study will attempt to use the Evaporation Technique in the SWCC determination. In order to construct the SWCC using the Evaporation Technique, samples were prepared using three soil types (Plasticity index (PI) 7-60) compacted at or near wet of optimum water content. The findings of this investigation are in good agreement with the data reported in the literature using the PP. Samples were arranged in the order of their PI. Higher PI samples at the same suction resulted in less water content. The Evaporation Technique created as part of this study is proven to be efficient, and reliable.

Vapour equilibrium techniques are usually used in determining the water retention curve of soils for high values of suction. These high values may regularly occur in highly plastic soils such as compacted bentonites, which are often proposed for the engineered barrier in high-level nuclear waste disposal. In the present work, the use of saturated saline solutions for vapour equilibrium techniques at different temperatures was studied. First, a hygrometer was used to measure the evolution of the relative humidity generated by salt solutions at different temperatures. Then, several vapour equilibrium techniques were used for suction control in the determination of water retention curves. The results on relative humidity measurement were in good agreement with that found in the literature. It was observed that the equilibrium time during water retention tests depended on several parameters as the exposition surface of the solution, the dimension of soil samples, and the air circulation, etc. Finally, the water retention curves obtained evidenced the effects of temperature, dry density, hydric history, and negligible influence of the technique used.

Soil water retention curves were determined using four different techniques for suction measurement. Transistor psychrometer and non-contact filter paper techniques were used for total suction determinations; pressure plates and in-contact filter paper techniques were used for matric suction measurements. The soil water retention curves (WRC) were determined for bentonite, speswhite kaolin and a mixture of kaolin and sand (70 percent:30 percent by mass). Specimens were prepared at their liquid limits and the WRC were determined for drying from this initial state. At least four measurements were made for each point on the WRC to ensure reliable results. The transistor psychrometer was found to be suitable for testing suctions between 200 kPa and 18,000 kPa. However, results were found to be very sensitive to temperature change. The pressure plate gave reliable results but was limited to 1500 kPa and tests took long time periods. Osmotic suction was also determined for kaolin using a saturation extract technique followed by measurement of electrical conductivity. The results for total suction determinations showed good agreement with the sum of the matric and osmotic suction measurements. The filter paper was used successfully for both total and matric suction determinations. The results suggest that there is not a need for separate calibration curves for matric and total suction, but that a single curve can be applied to both, if the equilibration time for total suction measurements is sufficient. It was found that reasonable agreement between techniques was obtained using published calibration curves for filter paper measurements.

The structure of a soil, generally derived from the particle arrangements, significantly influences the way it responds to loading. In unsaturated soils, grouping of clay particles forms bi-modal pore size distributions in which the aggregates (a group of fine particles) are generally separated by larger air voids. This paper examines the bi-modal pore size distribution of unsaturated compacted kaolin and the effect of confining pressure on the soil structure. Triaxial tests were performed on samples of kaolin prepared at 25 percent water content which were initially saturated and subsequently taken to various consolidation pressures: 200kPa, 400kPa, 800kPa, 1600kPa and 2000kPa. At the end of the tests a small specimen from each sample was subjected to a Mercury Intrusion Test in order to evaluate the bi-modal distribution of pore sizes. Further tests were performed on samples prepared from kaolin slurry for comparative purposes. The results have shown that the presence of bi-modal distribution of particles cannot be destroyed by subjecting the soil to high pressure. At pressures reaching 2000kPa both the samples prepared from kaolin slurry and those prepared from kaolin aggregate exhibited a similar pressure-volume relationship.

This paper presents a method for estimating the soil-water retention curve (SWRC) for plastic soils, in terms of gravimetric water content versus soil suction, according to the soil type, and also the stress history of the soil. The method is based on a study where 49 SWRCs from 18 soils were examined. The method requires the liquid limit of the soil, one set of data of water content/suction measurement and knowledge of the stress history of the soil. In order to facilitate the use of the method a theoretical approach using the over consolidation ratio is presented. The objective of the method is not to substitute the necessary laboratory tests, but top give the user some guidance for the general trend for the experimental result.

During the last decade several researchers and practitioners have stressed the importance and the need of adopting unsaturated soil mechanics in the analysis of problems associated with swelling clays and other problematic unsaturated soils. To accomplish this goal, the determination of the soil-water characteristic curve (SWCC) becomes essential. However, practitioners perceive the SWCC determination as a time consuming and expensive laboratory process. This fact has lead to a slow implementation of unsaturated soil mechanics into geotechnical engineering practice. In an effort to encourage practitioners to employ unsaturated soil mechanics, a study was performed to determine the SWCC based on soil index properties and one measurement of matric suction. In this study, the matric suction of approximately 82 field soil samples was determined using either suction plate or suction membrane devices. It was determined that the proposed procedure provides acceptably reliable approximation of the SWCC.

A number of geoengineering problems, involving unsaturated soils, require an understanding of the hydraulic relations between degree of saturation and matric suction head. For example, knowledge of hydraulic relations is required for solution of coupled flow and stress-strain behavior of unsaturated soils. Similarly, in modeling the flow of contaminants in unsaturated soils information about hydraulic relations is essential. In this paper, major features of hydraulic relations under dynamic flow condition are highlighted and discussed using experimental results from tests involving static equilibrium and dynamic flow of water. The experimental apparatuses used included Buchner funnel and fully instrumented 2-D tank. Several scanning loops were simulated to investigate the nature of hydraulic relations.

A method for direct measurement of hydraulic conductivity of unsaturated soil is presented. The method is based on a system of capillary flow from a controlled water table up into a soil column in initially air-dry condition. By measurement of change in soil water content at a known vertical position from the water table, together with a prepared soil water retention curve, the unsaturated hydraulic conductivity K(θ) can be calculated. The calculation is based on Darcy's Law and continuity equation. The proposed method was tested with 6 soil samples and compared with the indirect methods of van Genuchtent and Gardner. The results of K-θ relationships show linear trends in semilog scale in a form of K = ce^aθ where a and c are parameters.

The hydraulic conductivity — volumetric water content relationship (k-function) of an unsaturated soil is often estimated using predictive models based on the soil-water retention curve (SWRC). The accuracy of such models in predicting the actual soil k-function is seldom verified by direct measurements of hydraulic conductivities, mainly due to the considerable large timeframe necessary to evaluate precisely several magnitudes of hydraulic conductivities at different volumetric water contents. In order to contribute to the understanding of the accuracy of soil k-function estimates using predictive models based on SWRC, the manuscript presents a comparison between the predicted k-function using a SWRC based model and direct measurements of unsaturated hydraulic conductivity of a sandy soil. The soil k-function was inferred using the Mualen-Van Genuchten model, with SWRC data obtained from pressure extractor and free drainage column tests. The soil unsaturated hydraulic conductivities were obtained experimentally using steady-state unsaturated centrifugal flow at different increased gravities. Results comparison indicates that the predicted k-function using a SWRC based model can differ considerable from the actual unsaturated hydraulic conductivities measured experimentally. The difference between the predicted and measured magnitudes of hydraulic conductivity increases as the volumetric water content decreases. The analysis indicates that the accuracy in predicting the soil k-function relies on defining properly the soil residual volumetric water content. Inferring the soil k-function using the residual volumetric water content magnitude obtained from pressure extractor tests, leads to an underestimation of the actual soil k-function, while inferring the soil k-function using the residual volumetric water content obtained from evaporation tests leads to an overestimation of the actual soil k-function.

Unsaturated flow modelling of landfill hydraulics is a potentially valuable tool for the prediction of leachate discharge rates, design of leachate control systems and simulation of biodegradation in engineered landfills. Such an approach requires information on the moisture retention and hydraulic properties of waste, usually in the form of empirical functions such as those presented by van Genuchten (1980). However, there is a lack of information on moisture retention and conductivity properties. Recently, Kazimoglu et al. (2003, 2005) reported on moisture retention in MSW and presented a laboratory-scale method by which retention properties can be obtained. As with conventional inert soils, laboratory determination of unsaturated hydraulic conductivity is difficult and recourse is often made to analytical methods. The aim of this paper is to compare the unsaturated hydraulic conductivity of MSW obtained using Passioura's (1976) one-step outflow test method with predictions using van Genuchten's model. Good agreement is observed between the predictive model and experimental method for unsaturated hydraulic conductivities at low moisture contents but there is poor agreement at high degrees of saturation. This latter discrepancy is attributed to the difficulty of measuring retention properties of large pores at low suctions and the applicability of the van Genuchten model to such a material.

The final cover layer of municipal solid waste (MSW) landfill is considered an important matter in reducing gas flow generated by organic matter decomposition process. Landfill gas (LFG) liberation may cause a number of environmental impacts, basically due to the presence of methane and carbon dioxide in its composition. A program of air permeability tests was carried out in order to investigate the cover layer efficiency in mitigating gas percolation at Muribeca's Landfill, Recife-PE, Brazil. This paper presents experimental techniques developed to study gas flow through compacted unsaturated soils by using a flexible membrane permeameter. The results allowed to determine the air permeability parameters related to molding water content, density, particles distribution, water saturation, and matric suction.

The use of TDR methods allows for the determination of the moisture profiles and the soil electrical conductivity profiles. Associated to the model of van Genuchten describing the water characteristic curve, the unsaturated hydraulic conductivity is computed. It is found that the leachate hydraulic conductivity is little greater than the water hydraulic conductivity. It is also found that the solute convection rate and the infiltration rate yield to the same value. A long time infiltration is necessary before the application of the Darcy law to determine the coefficient of permeability. The two models used to describe the hydraulic conductivity are in agreement. The influence of the leachate is not on these models. The dispersion coefficient becomes constant after 100 days of infiltration. The analytic solution of the convective-dispersive equation is a good approach of the solute concentration profile.

This paper compares hydraulic conductivity values obtained by laboratory and field tests. Two unsaturated gneiss residual soils are tested. In the laboratory, the water retention curves and the saturated hydraulic conductivities of the soils got from constant and falling head permeameters were used. In the field, the hydraulic conductivities were obtained using the Guelph permeameter. As it is well known, these techniques are limited to high hydraulic conductivity soils. The paper describes the testing methodology, the test results and comparisons between laboratory and field results. At the end, some conclusions are outlined.

Soil-water characteristics curve (SWCC) is a fundamental constitutive relation for the explanation of the engineering behavior of unsaturated soil. The study of SWCC is helpful for reducing the time and cost of unsaturated soil testing for engineering purposes. Some properties of unsaturated soils, such as shear strength and permeability, can be predicted from it. After a brief introduction to osmotic technique and vapour phase method, the SWCCs of Shanghai soft soil have been obtained and some basic characteristics of the SWCCs have been analyzed. Coefficient of permeability is an important parameter for unsaturated soils, but the wide range in coefficient of permeability has been proven to be a major obstacle in analyzing seepage problems. Childs and Collis-George proposed a model for predicting the coefficient of permeability based on the random variation of pore size. It was improved several times and proved to be effective. Based on this model, the relative coefficients of permeability of Shanghai soft soil are predicted from the soil-water characteristic curves. Results show that unsaturated permeability of Shanghai soft soil changes nonlinearly with suction.

The main problem in unsaturated flow modeling is the costly, tedious and time consuming process of hydraulic conductivity function measurement. Fredlund divided the assessment of unsaturated soil properties into a hierarchical system involving four levels. This paper shows the procedures suggested and conducted to estimate a lean-clay-hydraulic-conductivity function at three hierarchical levels. The hydraulic conductivity function was obtained under levels 2, 3 and 4. Level 1 (i.e. direct measurement) was not considered. Level 2 was divided in Level 2a and Level 2b, measurement of matric suction with pressure plate and filter paper, respectively. The most common equations, algorithms and estimation methods were analyzed, selected and applied. It was observed Level 2b analysis is the best for the lean clay studied.

Geotechnical engineers frequently encounter environmental problems involving unsaturated soil. Geotechnical problems related to hydraulic conductivity are affected by both air flow and water flow. Decision of coefficient of permeability for each of air flow and water flow has increased significantly. In this study, the air-water permeability testing apparatus is modified for determining hydraulic conductivity of double fluid flow. The soil-water characteristic curve of sand-clay mixture soil was determined by desorption using high air entry disk and glass desiccator. Vapor pressure technique to control relative humidity environment is effective for determining the water content in high soil suction. Soil-water characteristic curve can be used for interpretation of hydraulic conductivity. The coefficient of air permeability is greater than that of water in unsaturated sand-clay mixture soil.

The paper presents an experimental study carried out on a clayey material used in the construction of the core of the Sikkak earth dam (N. W. of Algeria). The characterization of the volumetric behavior under the effect of suction is highlighted on drying-wetting paths. The material is used under slurry form and compacted at different initial water content. Oedometric paths were studied in order to determine the potential of collapse under various initial densities.

The influence of fines (percent < 0.075 mm) on the properties of arid surface sands was examined by an extensive laboratory testing program. This was initiated after chemical analysis revealed that the fines are calcareous, softer and contain more than 30 percent of carbonates and sulphates. Jahra sand — a local natural silica sand with 4.6 percent fines was employed in this program. The amount of fines introduced in this sand varied up to 30 percent. Tests included basic properties, and Modified Proctor tests. Direct shear and consolidation test specimens were prepared at a relative density of 60 percent for all sands with different fine contents. CBR tests were performed on samples prepared to maximum density and optimum moisture content. The results indicate increased strength and maximum dry density up to 20 percent fines followed by a sharp decrease in strength with 30 percent fines. Compressibility remained unchanged and increased sharply as the fines content increased from 20 percent to 30 percent. The influence of the degree of saturation on the compaction characteristics and the CBR values of the test soils are examined.

This paper presents a geotechnical characterization of an unsaturated soil from the "Barreiras Formation", in Pernambuco state, Brazil. The investigated area encompasses a slope under erosive process located in the forest reservation of Horto de Dois Irmão, in Recife. The purpose of the study was to elaborate an executive project for the recuperation of the eroded area. The geotechnical investigation program presented, borehole with Standard Penetration Test (SPT), hydraulic conductivity test, undisturbed and disturbed sampling, laboratory testing for soil-water characteristic curve, assessment of soil erodibility, direct shear test both conventional and with controlled suction. The results were in accordance with the type of soil and the literature. The proposed solution consisted of an alteration in the slope geometry, recomposition of vegetation cover and implementation of a superficial drainage system. A slope stability analysis was performed considering some suction effect on the shear strength of the soil.

Light-weight soil is mainly used effectively as recycling technology of the construction surplus soil in Japan. Foam composite light-weight soil proposed in this study is a mixture of clay, cement, and foam. As a porous material, the wet unit weight of the foam composite light-weight soil is decreased by the usage of foam from 7 kN/m³ to 13 kN/m³ compared to a normal soil which is around 18 kN/m³. In this study the mechanical properties of unsaturated foam composite light-weight soil are investigated. Laboratory test with initial suction, isotropic consolidation, consolidated drained and constant water triaxial compression tests are carried out on unsaturated foam composite light-weight soil samples. From the study carried out, it is learned that initial suction do not exist, and collapse phenomena do not occur when the isotropic stress is below consolidated yield stress for this kinds of soil. The authors also found that the stress path and the failure criterion of this peculiar unsaturated foam composite light-weight soil are best described with an independent effective strength formula.

This paper presents water retention data of a nonwoven polyester geotextile obtained from a controlled flow capillary cell. The results were evaluated and compared to conventional hanging column test results. A good correlation was found between the two test methods. The hydraulic properties of the geotextile were determined by inverse modeling using the modified van Genuchten equations. The modified van Genuchten model was found to be in close agreement with the experimental data.

Expansive clays exhibit high potential for volume change because of changes in soil moisture. It is estimated that the annual cost of damage to facilities built on expansive clay in the United States exceeds $9 billion. Because of the physico-chemical nature of expansive clay behavior, the use of chemical agents to improve their behavior is common. The research reported involved development of a viable method to simulate injection of chemical agents into soils and evaluation of the results. The agents included in this study were lime, a potassium based agent and a group of what are called ionic agents. The clay utilized was weathered from the Eagle Ford Shale formation in north central Texas. An injection treated layer of the clay, using one of the differing chemical agents, was placed between a relatively wet layer and relatively dry layer of untreated clay, to test each chemical's affects on moisture movements. The sealed cylinders were allowed to have what moisture movements during a fifteen week period. The cylinders were opened and the soil layers tested for water content, dry unit weight, soil suction, and swell potential. Cylinders of untreated clay, constructed in the same way, but not injected were also analyzed. The injection process developed is believed to have simulated field injections well. The results of moisture movement, soil suction, and swelling potential provided insight into how these chemical agents can affect moisture movements and soil behavior when injected in this clay.

This paper aimed to study the behavior of soil-water characteristic curve SWCC (studied range 50–325000 kPa) for three expansive soils selected from Mosul city. The effect of lime treatment was also studied on SWCC. The studied variables were soil type, compaction effort, compaction state, lime percent and curing time (up to 150 days). A prepared computer program by the authors was used to fit the obtained experimental data depending on Gardener, Van Genuchten, Fredlund & Xing equations. Test results show that all SWCCs have S-shape curve, and the water holding capacity was increased due to the increase in clay fraction, compaction effort, compaction state in the sequence of (wet of optimum, optimum, dry of optimum), lime percent and curing time. The program shows that Fredlund and Xing equation best fit the experimental data.

In general, a capillary barrier system is constructed using a fine-grained soil overlying a coarse-grained soil. Experimental results of two 1-D capillary barrier models are presented in this paper. The first capillary barrier model was constructed using gravelly sand as the coarse-grained layer and the second capillary barrier model was constructed using geosynthetic material as the coarse-grained layer. The geosynthetic material used in this study is Polyfelt Megadrain 2040, which consists of a matrix of polypropylene monofilaments sandwiched between two layers of polypropylene geotextile. Fine sand was used as the overlying fine-grained layer for both capillary barrier models. Below the gravelly sand and the geosynthetic material, fine sand layers were placed to represent the original soil which is protected by the capillary barrier model. Several infiltration tests followed by drying processes were applied to the capillary barrier models. Measurements of pore-water pressures (PWP) and volumetric water contents along the capillary barrier models were used to study the behavior of the models. The test results showed that both gravelly sand and geosynthetic material were able to create capillary break. Both capillary barrier models were able to maintain the PWP above and below the coarse-grained layers at negative values. The geosynthetic material was found to be more effective than the gravelly sand to be used as the coarse-grained layer in a capillary barrier system due to the fact that the PWP head above the geosynthetic material was lower than that above the gravelly sand layer. Therefore, the unsaturated permeability of the fine-grained layer above the geosynthetic material was lower than that above the gravelly sand layer. As a result, water did not infiltrate easily into the soil layer above the geosynthetic material.

A thermal probe based on the hot wire method was used to measure the thermal conductivity of MX80 bentonite specimens compacted at different dry densities and water contents. These tests evidenced the influence of dry density, water content, saturation degree, mineralogy and volume fraction of soil components on the thermal conductivity of the soil. A simple method was finally proposed allowing the prediction of the thermal conductivity of compacted bentonite. Its relevance was evaluated.

This paper presents results of a series of laboratory tests aimed to investigate the behavior of thermal properties of unsaturated tropical soils. Two different soils were used: a sandy-clay (SC) and a clayey-sand (CL). The thermal parameters, thermal conductivity and heat capacity were study for both soils. Laboratory tests were performed using the thermal needle. A new methodology was developed to determine the heat capacity of soils. The results have shown that thermal conductivity and specific heat are not constant for a given soil but varies with gravimetric water content, degree of saturation and temperature. Results showed that that the thermal conductivity varies exponentially with gravimetric water content and degree of saturation whereas specific heat varies linearly with gravimetric water content and temperature.

The paper reports an experimental investigation on the influence of suction, stress path and temperature on the mechanical behaviour of compacted bentonite. The tests have been performed in oedometers in which suction is controlled by means of sulphuric acid solutions of different concentration. The cells were placed in thermostatic baths that allowed testing at 20, 40, 60 and 80 degrees C. Two different stress paths were followed: loading under constant suction followed by wetting under constant load, and wetting under a low load followed by loading of the quasi-saturated sample. The results show that the stiffness of the bentonite, when loaded at hygroscopic conditions, is very high in the range of temperatures tested. Nevertheless, the temperature increase involves a certain reduction in the size of the elastic domain, both in saturated and unsaturated conditions, and consequently an increase of the compressibility of bentonite. From the experimental results, it seems that the deformation of bentonite is more dependent on the stress and suction path than on temperature.

In order to study the Structure-damage evolution and thermal-mechanical characteristics of unsaturated expansive soil, a multi-function triaxial apparatus for soil was developed in recent years. The new equipment has the functions of both CT-triaxial apparatus and temperature-controlled triaxial apparatus. The structure and characteristics of the new equipment were introduced. A number of CT-triaxial test were conducted, and the properties of the structure-damage evolution of undisturbed expansive soil during loading were obtained. A series of triaxial tests with temperature-control for unsaturated remolded expansive soils were carried out, and the test results show that temperature has great effect on the mechanical behavior of the soil, such as strain, strength and soil-water characteristics curve.

The objective of this research is to evaluate the characteristic of air permeability in order to study the water seepage characteristic of unsaturated soil considering the pore air behavior. In this paper, a new method of air permeability test which able to repetitively control the suction to adjust the degree of saturation of an unsaturated soil sample has been developed. The relationships among air permeability coefficient, degree of saturation and suction have been evaluated. The proposed apparatus is able to measure the air permeability coefficient in the range of 10^–2 to 10^–6 m/s. The developed apparatus is able to measure the air permeability with respect to the degree of saturation and suction and also able to obtain the moisture characteristic curve simultaneously. The new method resulted in appropriate data by comparing it with past experimental data in other papers.

The soil-water characteristic curve SWCC has become one of the most readily available experimental means for estimating fundamental engineering properties of unsaturated soils for a wide range of matric suction states, including permeability, volume change, and shear strength parameters. Numerous laboratory techniques have been developed for accurately assessing the SWCC of unsaturated soils, from filter paper technique to the more sophisticated pressure plate extractor devices. However, the majority of these techniques and devices allow for the testing of unsaturated soils under unknown or zero-confinement conditions, resulting in SWCC data that do not correspond to realistic in-situ stress states in the unsaturated soil mass. Moreover, recent advances in SWCC testing using oedometer and triaxial setups may prove costly and time consuming. This paper describes an attempt to develop a modified pressure plate extractor (MPPE) device for assessing the SWCC of unsaturated soils under anisotropic stress sates, including details of its main components and the step-by-step assembling process. The MPPE device features independent control of the net radial confinement and net vertical pressure. Results from a series of SWCC tests on poorly-graded sand (SP) and low-plasticity clay (CL), for different values of K_o ratio, that is, the net radial confinement to net vertical pressure ratio, are presented. Results show a paramount influence of the net radial confinement, and hence the initial K_o value, on soil's air-entry value and residual volumetric water content. Additional data and further refinement of the developed MPPE device, which builds upon a well known technique, will prove very useful in assessing key engineering properties of unsaturated soils under different K_o stress states.

A laboratory test method has been developed to determine the relationship between vertical movement and water content of an expansive soil. The test method involves cyclically wetting and drying of the soil under a normal compressive load to develop the relationship between vertical movement and water content. Testing is accomplished using a conventional oedometer with a modified loading cap that allows for forced air circulation to accelerate the shrinking phase. Accelerated shrinking both reduces testing time to less than a week and results in cracking patterns in the surface of the clay specimen that are similar to those observed in the field. Test results are presented for specimens of expansive clay from central Texas. Specimens followed the same path of shrinking and swelling after three to four swell-shrink cycles, indicating that effects of sample disturbance can possibly be overcome with cycling. The measured vertical movements for a given change in water content were compared with those predicted by the empirical Potential Vertical Rise (PVR) method that is used commonly in practice. For specimens where the PVR method predicts small movements (less than 25 mm to 50 mm), the measured movements were greater. For specimens where the PVR method predicts large movements (greater than 25 to 50 mm), the measured movements were smaller. This result is consistent with comparisons of PVR predictions with field-measurements. Therefore, this test method shows promise in that it is both practical and potentially effective at accurately predicting vertical movements in expansive soils.

This paper presents a new triaxial testing device for unsaturated soils. It describes the main features of the triaxial system including double wall cell, pressure/volume controllers, frictionless volume change device, automatic flushing device, and software. The use of a double wall cell eliminates almost all of the volume change compliance errors inherent to single wall cells while the frictionless volume change device accurately measures the change in sample volume. An important feature presented is the automated flushing device which removes diffused air during unsaturated soil testing. Diffusion causes a small amount of air to escape through high air entry ceramic disks during testing. The diffused air accumulation beneath the ceramic disk affects water flow and volume change measurements. In addition, large air pockets created over time can reduce the contact area between the water and the ceramic, which may eventually cause de-saturation of the ceramic disk. Therefore, diffused air should be regularly eliminated to make accurate water content measurements. The accessories of the new diffused-air flushing device include a computer-controlled ball valve and a water level sensor. The flushing device can be integrated into other unsaturated soil testing systems such as the direct shear apparatus, and soil-water characteristic devices. The software provides sophisticated algorithms for automating all testing procedures including diffused air flushing.

Swelling pressure is an important parameter used in the design of drilled pier foundations by engineers practicing in the state of Colorado. The confined swelling pressure is also used in some heave prediction methods, in assessing swell risk for construction, and may be used in analysis of drilled pier movement. Measurement of confined swelling pressure can be a labor-intensive process or involve lock down of the load arm when using conventional oedometer equipment. A computer-controlled device was developed to apply additional load in small increments when swelling was detected. The control was set to limit volume change in the sample to a fine limit and the load device was directed by the control mechanism to maintain constant sample volume. The paper presents measurements of swelling pressure measured using both procedures to allow comparison and correlation with the more common load-back procedure.

The knowledge on the thermo-mechanical behaviour of compacted bentonites at unsaturated state is helpful for the conception of high-level nuclear waste disposals where this material is used as an engineered barrier. In the present work, a new suction and temperature controlled isotropic cell was developed. This cell allows to perform thermo-mechanical tests under high pressure (p = 0.1 – 60 MPa), at different temperatures (T = 20 – 90 degrees C) and different suctions (s = 4 – 500 MPa). After a careful calibration of the system, some tests were carried out to study the thermo-mechanical behaviour of compacted MX80 bentonite, which is one of the reference materials for engineered barriers. The results obtained were in good agreement with those found in the literature. This confirmed the good functioning of the new cell.

Creep plays an important role in the estimation of long-term behaviour of cohesive soils. It is, among other parameters, related to the viscosity of the soil, hence rate-dependent. Apart from the degree of saturation and the initial void ratio, the stress rates as well as isotropic / deviatoric stress states determine the behaviour. A partially saturated kaolinite is selected as the test material. The creep tests are performed in a triaxial apparatus with tensiometer-measured suction. To combine the viscous and the elasto-plastic behaviour, a new triaxial apparatus performing deviatoric triaxial tests has recently been developed, which is described in detail in this paper. Selected results of the triaxial and creep tests are presented. The creep tests show good agreement with results from the literature and may be directly used to formulate an elasto-visco-plastic constitutive law for this type of soil.

Resilient properties are assumed to account for the plastic deformation of subsoils and hence are used in the characterization of subsoils, while designing flexible pavements. Characterizations based on resilient strains result in poor characterizations of select fine soils, which experience considerable amounts of permanent or plastic strains under repeated loads. Hence, a test procedure to measure plastic strains of soils was formulated using a repeated load triaxial test device. Several plastic strain tests were performed on Kaolinite clays compacted at different moisture contents. This paper which presents these test results address the effects of compaction moisture content, dry unit weight, confining pressure and deviatoric stresses on plastic strains of clays. The accumulated plastic strains at two different numbers of repeated loading cycles, 10 k and 50 k cycles, have been studied. The repeated loading response on the resilient modulus properties over the same number of cycles is also analyzed and addressed.

The rock water retention curves (RWRC) of two limestones, namely; Tuffeau and Sebastopol that are commonly used in the construction of several monuments in France are determined using modified experimental procedures. The long term objective of this research is to extend the principles of unsaturated soil mechanics to interpret the engineering behavior of these two soft limestones. As a first step, the equations proposed by van Genuchten (1980) and Fredlund Xing (1994) for fitting the soil-water retention curves (SWRC) were tested for fitting the RWRC. The study suggests that there is a good comparison between the measured and the fitted RWRC using the commonly used SWRC equations.

The filter paper technique has been used in the geotechnical community for the determination of soil suction. Depending on the sort of contact existing between both, the filter paper can measure either matric or total suction, absorbing water until equilibrium is reached. By means of the calibration curve, one may infer suction values in the soil from the water content of the filter paper. In order to obtain this calibration, several equipments can be used to impose a known suction on the filter paper. This article presents calibration data of five batches of Whatman No. 42 filter paper, obtained through the use of pressure plate, suction plate and NaCl solutions. It has been noticed that one of the batch presents a calibration curve which differs from calibrations proposed in the literature. In order to better visualize the effect of the difference between calibration curves, information on the retention curve of a residual soil of gneiss is presented, using the calibration obtained in this work. Results suggest that when the filter paper method is used as suction quantifier, it is important to get at least one point of calibration, comparing it to the calibrations proposed in literature.

Suction control for testing unsaturated soils requires the use of high air-entry ceramic disks (HAED). HAEDs provide a medium for water in a soil specimen to migrate as required regardless of the applied pore air pressure. Even though the HAEDs are rated for a specific bubbling pressure, some air diffuses through the HAEDs, even at pressures below the rated capacity. Diffused air has a tendency to accumulate underneath the HAEDs resulting in an error in the specimen water content measurements. Diffused air has the potential to de-saturate the bottom of the HAED thereby retarding the flow of water through the disk. In this study, commonly used HAEDs were tested with the Fredlund soil-water characteristic curve (SWCC) device to measure the diffusion rate at different pressures. Three-inch diameter HAEDs rated at 100, 300, 500, and 1,500 kPa bubbling pressures were tested at different percentages of their rated pressures (e.g. 50 percent, 95 percent) to determine the diffusion rate. Several HAEDs from the same manufacturer were tested for each bubbling pressure. The hydraulic conductivity of the HAEDs was also measured at different testing stages to determine the effects of air accumulation underneath HAED disks. Diffused air volume rates are presented graphically. Conclusions and recommendations for the frequency of flushing are also presented. In summary, the amount of diffused air generated when using 1- or 3-bar, 5-bar, and 15-bar is: negligible, small, and large, respectively.

The retention curve is an important tool in the analysis of the behavior of unsaturated soils. Its result has been used in models to forecast soil shear strength envelope and in problems that involve water flow. Several types of equipments and techniques have been used for the determination of the soil water retention curve, such as the suction plate, the pressure plate and the filter paper. In this article a study is presented on the equilibration time for the suction imposed on the pressure plate apparatus. Through the use of this equipment, the axis translation technique may be applied, imposing matric suction on soils specimens assembled on the ceramic disk. Usually, suction equilibration time is determined by the observation of the variation of the water level in a burette connected to the ceramic disk. However, it is not always possible to determine the instant of full stabilization. From a gneiss residual soil, specimens have been statically compacted, at different region of the compaction curve: the dry branch, the optimum point, and the wet branch. Equilibration time has been overseen by means of the observation of the water level in the burette, and by the variation of the gravimetric moisture content of the soils specimens. Suction was also measured using a high capacity tensiometer. The results seem to demonstrate the effect of the water exchange between the ceramic disk, the soil specimens and the internal environment of the pressure plate, and it justifies the asymptotic behavior generally observed.

A new suction probe, the WF-DU tensiometer, has been developed by Wykeham Farrance in collaboration with Durham University. The performance of the new suction probe is assessed in terms of the minimum negative value of pressure that can be measured by the probe before cavitation. A series of tests performed on four different units of the same design of tensiometer is presented and the results are discussed in relation to the procedure adopted for the saturation of the probe prior to the tests. Other aspects of the tensiometer response are interpreted by making reference to similar experiments published in the literature. All tensiometers used in this work are calibrated in the positive range of pressure. The extrapolation of such calibration to the domain of negative pressures is validated by using the tensiometer to measure known values of suction.

The paper presents an experimental investigation of the effects of temperature on the response of strain-gauge based tensiometer for direct measurement of high suction. These tensiometers are often manufactured with a stiff measurement diaphragm and are therefore sensitive to temperature. The effect of temperature on strain gauge voltage output is difficult to predict and manufacturers suggest to directly measuring thermal effects gauge by gauge. The paper shows that measurement errors associated with temperature change may be significant and suggest possible approaches to correct temperature-induced errors.

Theoretical and experimental studies have shown that the tensile strength of water is much greater than 100 kPa, the cavitation pressure of water. On this basis, a miniature tensiometer was developed in this study to measure high matric suctions in the residual soils of the sedimentary Jurong Formation of Singapore. Various factors that may affect the performance of the miniature tensiometer were evaluated. This paper presents the development of this miniature tensiometer and the evaluation results. The effects of pre-pressurization, assembly procedure, pressure reversal, equilibrium time, suction sustaining time, and restoration after suction exceed the air entry value of the ceramic disk were examined.

The paper presents an elasto-plastic model describing the behaviour at yielding of partially saturated soils. This model assumes that the soil remains saturated until the air entry value is exceeded and its behaviour is controlled by effective stresses. Beyond this particular point, the soil becomes unsaturated and the model then makes use of two independent stress variables: "modified net stress" σ^* = σ – u_a + χ^* (u_a – u_w) and matric suction. An experimental programme was performed in suction-controlled apparatus on reconstituted clay from Ste-Rosalie, Canada. The agreement between observed and computed results is satisfactory and confirms the possibilities for the model to reproduce the most important features of partially saturated soil behaviour.

In this paper, a fully coupled elasto-plastic model is presented for the cyclic analysis of unsaturated soils based on the bounding surface theory. The bounding surface is defined in terms of effective stresses using a simple and versatile function. A modified radial mapping rule is adopted enabling prediction of a range of cyclic loading paths including drained and undrained, isotropic and deviatoric. The model captures saturated as well as unsaturated behavior in a unified manner.

The elastoplastic model BExM for swelling unsaturated soils was implemented in a finite-element code (Code_Bright). The model validation is then checked through the numerical simulation of suction-controlled oedometric tests made on three different swelling soils. The simulation results of these tests showed that the numerical model is able to correctly reproduce the experimental data.

In this paper, aspects of modelling geotechnical problems for a class of hard soils and soft rocks, soft argillaceous rocks, are addressed. The model developed by Vaunat and Gens (2003), dealing with typical features of cemented geomaterials, is introduced first. Afterwards, additional developments are proposed to tackle suction induced effects. Finally a simulation of the excavation and ventilation of a 6m diameter shaft in a Callovo-Oxfordian mudstone has been performed using a computer code, Code Bright, in which the new model has been implemented. The presented model appears to offer a possibility to simulate the time dependent damage and fracturing induced by suction cycles often observed in this type of geomaterials.

This paper deals with the application of the theory behind the pseudo strain concept to unsaturated soil behavior and testing methods leading to the calculation of pseudo strain and dissipated pseudo strain energy. Drained monotonic triaxial and undrained cyclic triaxial tests at constant suction were conducted on silty sand to investigate the effect of viscous resistance during testing. The monotonic triaxial test results analyzed both with and without the use of the pseudo strain concept were compared to the critical state based elasto-plastic model. The pseudo-strain test results conform more closely to the Alonso et al model. The cyclic test results were used to calculate the dissipated pseudo strain energy associated with accumulating plastic pseudo-strain. Based on calculated dissipated pseudo strain energy, a dissipated pseudo-strain deviatoric power failure theory for unsaturated soils under earthquake loading is proposed. The proposed theory shows the conditions under which the soil will fail either by cyclic mobility (soil acts as a solid) or by liquefaction (soil acts as a liquid).

This paper deals with constitutive modeling of soil behavior in saturated and non-saturated conditions. It presents an extension of the effective stress principle that, together with small adaptations made in an elasto-plastic model developed for dry and saturated conditions, allows for modeling soil behavior on non-saturated conditions, including the decrease of volume during wetting stress-paths (collapse). Besides discussing theoretical aspects, the paper presents results of triaxial tests performed on a young residual soil in saturated and non-saturated conditions, shows how to obtain new parameters included in the model and presents comparisons between experimental and analytical results obtained with the constitutive model.

In this paper, a constitutive model is presented to reproduce the behaviour of hard soils and soft rocks (HSSR) that exhibit an intrinsic expansive behaviour. The model focuses on argillaceous soft rocks and includes suction effects. The soil or rock is considered as a composite material made of an expansive argillaceous matrix and bonding agents that provide the effect of structure. Both components are endowed with their own constitutive law: Barcelona Expansive Model is used to represent the behaviour of the matrix; the damage model presented for the bonds. The model for the composite material is then derived on the basis of energy considerations. In a first part of the paper, the model is formulated. In a second part, its performance is described. Results show the capability of the model to reproduce features such as moduli and strength degradation during loading and elastic rebound above the initial compression line during unloading.

As part of a study to develop a new theoretical framework for unsaturated soils that incorporates both shear strength and volume change behaviour due to changes in soil suction and applied stress, a new four zone shear strength surface has been proposed. This model encompasses the shear strength behaviour at low and high values of both suction and net stress or effective stress for unsaturated and saturated soils respectively. The model is validated by data reported in the literature and obtained from large diameter triaxial tests on limestone gravel.

A non-linear incremental constitutive model for unsaturated loess is proposed in this paper on the basis of Fredlund's nonlinear total constitutive law and Domaschuk-Valliappan model (K-G model) for saturated soils. The proposed model can describe the stress-strain-strength behavior and moisture change of remolded unsaturated loess. The influence of matric suction on the strength and deformation characteristics of unsaturated loess is considered in this model. Totally 14 parameters are involved in this model. When the matric suction approaches to zero, this model is degenerated to Domaschuk-Valliappan model. The method of determining all parameters is introduced in detail. The test required are (a) the isotropic compression test by controlling matric suction s constant, (b) the triaxial drained shear test by controlling cell pressure and matric suction s constant, and (c) the isotropic shrinkage test by controlling net mean stress p constant on unsaturated soils. From the test results, all parameters can be determined directly and can be utilized in the numerical calculation. The model is used to predict matric suction changes in a drained triaxial test, good agreement is obtained.

Experiments were undertaken using a suction controlled oedometer to determine parameters for reconstituted South Carolina Kaolin, for use in the so called "Barcelona Basic Model", a constitutive model for describing the behaviour of partially saturated soils. The oedometer employs the osmotic method of suction control and incorporates a semi-rigid cell wall to enable the determination of lateral stresses in the soil sample throughout testing. Two series of five tests were undertaken. In the first series of tests, suction was held constant and vertical stress was increased incrementally. In the second series of tests, vertical stress was held constant and suction was increased incrementally. Unload-reload steps were also undertaken for selected tests in both series. The results from the tests were used to determine the stiffness parameters λ(s) in response to changes in net mean stress, and λ_s in response to changes in suction. The elastic stiffness parameters, k for changes in net mean stress, and k_s for changes in suction, were determined from the limited unload-reload data collected. Clear trends were evident in the values of the stiffness parameters, and conclusions are drawn regarding the significance of the trends.

In the present paper, the formulation based on averaging theory is developed for simulating unsaturated soils as a 3-phase nonlinear elastic media. The model is implemented in a new finite element code. This code is able to use different expressions given for state surfaces and soil properties. Comparing the experimental measurements of drainage from a sand column with the results of this model verifies its validity.

Rate sensitivity is an important characteristic of geomaterials for both saturated and unsaturated soils. However, many constitutive models for unsaturated soil have been constructed within the framework of the rate independent theory. The present study addresses an elasto-viscoplastic constitutive model which considers the effect of suction for unsaturated clayey soil and a soil-water-air three-phase coupled analysis using the elasto-viscoplastic model. The proposed constitutive model adopts the average skeleton stress for the effective stress from the viewpoint of the mixture theory. Hence, it has become possible to construct a model for unsaturated soil starting with a model for saturated soil by substituting the average skeleton stress for the effective stress and introducing the suction effect into the constitutive model. Furthermore, the collapse behavior, which is brought about by a decrease in suction, is described by the shrinkage of the overconsolidation boundary surface, the static yield surface, and the viscoplastic potential surface. A numerical analysis for multiphase materials is conducted within the framework of a continuum mechanics approach through the use of the theory of porous media. The theory is a generalization of Biot's two-phase mixture theory for saturated soil. A soil-water-air three-phase coupled finite element method is developed in the present study using the governing equations for multiphase soil based on the non-linear finite deformation theory. The average skeleton stress is defined as the difference between the total stress and the average pressure of the two fluids and is used in the proposed elasto-viscoplastic constitutive model. A van Genuchten (1980) type of equation is employed as the constitutive equation between the liquid saturation and the suction pressure. Numerical simulations of unexhausted-undrained compression with different strain rates are conducted under plane strain conditions, and the applicability of the proposed method is evaluated with respect to strain localization and the effect of suction.

Numerous authors have described issues of numerical oscillation and slow or inaccurate convergence in finite element analyses of transient non-linear phenomena. Karthikeyan and Tan have discussed numerical issues such as the effect of element size and time step in analyzing transient flow through unsaturated soils. Unsaturated flow problems are highly non-linear because of the variations of soil hydraulic conductivity and water content (soil-water characteristic curves) with matric suction. Edgers and Nadim (2003) analyzed the pore pressure response of soils during a precipitation event that triggered a debris flow in Norway. They used the computer program SEEP/W for a two-dimensional finite element analysis with one directional flow. Their numerical analyses showed reasonable agreement between the time computed for the development of positive pore water pressures, approaching full hydrostatic, and the observed elapsed time to failure for the debris flow. This paper follows up on these earlier studies by analyzing the effects of time step, element size, and element type (4- vs. 8- noded elements) on the numerical solutions to the Edgers and Nadim (2003) case study. This paper shows that for this case study, uncertainty in the assumed hydraulic conductivity is more important than inaccuracies caused by numerical oscillations and slow or inaccurate convergence.

The physical meanings of the directional derivative of the constitutive surfaces are discussed. The signs of these parameters are reexamined by using the measured constitutive surfaces for an unsaturated expansive soil and an unsaturated collapsible soil. Based on the discussion, parameters studies are performed to numerically investigate these parameters' influence on the coupled consolidation of unsaturated soils. It is found that there is also a Mandel-Cryer effect in unsaturated expansive soils and a reverse Mandel-Cryer effect in unsaturated collapsible soils. Six parameters influencing the Mandel-Cryer effect in unsaturated soils are discussed.

The paper presents the MUSE project to academics and practitioners who have an interest in the area of unsaturated soil mechanics. The MUSE ("Mechanics of Unsaturated Soils for Engineering") project is a major research and training network funded by the European Union involving six European universities and five industrial partners. Further details can be obtained visiting the MUSE website (2005). The project started in December 2004 with a planned duration of four years. It aims at undertaking a comprehensive programme of research on unsaturated soils divided in the following four main tasks: laboratory testing, constitutive modelling, numerical analysis and application to practical problems.

This paper investigates the effect of various surface and subsurface site conditions on the predicted time rate of heave for expansive soils. A review of the migration of the wetting front for two project sites in the Denver Area is presented. Seepage analyses using VADOSE/W software (Geo-Slope, 2005) was conducted to evaluate the rate at which the wetting front will migrate under various boundary conditions. Heaving associated with change in water content of two soil profiles was computed to obtain the time rate of heave for these profiles. Poor grading and drainage are shown to have a large effect on the time rate of heave. The depth of wetting and its effect on the amount of heave to be expected in the lifetime of the structure are presented.

In this paper, a two-dimensional finite element model has been developed to investigate the effects of a leaking underground water pipe on the behaviour of a stiffened concrete slab on expansive soils. Two different cases have been studied. In the first case, a leaking pipe is located outside the footing, 0.6m away from the perimeter of slab. The second case considers a leaking pipe under the slab at a distance of 1m from the edge beam. For both cases, the pipe is assumed to be 0.6m below the ground level. The surface of the soil that is not covered by the slab is subjected to seasonal variation of moisture. The numerical analysis indicated that a leaking underground water pipe can cause more severe distortion to the footing than the seasonal climate changes.

Design of slab-on-grade foundations on expansive soils requires that the geotechnical engineer determine the maximum differential heave and the distance in from the edge of the slab over which the differential heave takes place (PTI, 2004). Currently the most commonly used method relies upon semi-empirical methodologies developed using observations primarily in the Southwestern United States. This methodology utilizes broad climatological data based on the Thornthwaite Moisture Index. Heave prediction methodology is based on the suction compression index and equilibrium suction states. This paper takes a different approach and utilizes the actual climatological and irrigation data. Oedometer methodology is used for heave prediction. Seepage analyses are conducted to evaluate wetting patterns over time beneath slab-on-ground along different vertical planes. The manner in which differential heave takes place over time and the pattern of heave development from edge-heave to center-heave is investigated. Maximum differential heave and edge moisture variation distance are plotted as a function of time. It is shown that in arid and semi-arid environments, as irrigation is applied around the structure, edge-heave initiates and governs over the life of the structure. This pattern, however, will change due to accidents such as broken pipes or wetting from other sources. The maximum differential heave and the edge moisture variation distance computed for sites with various drainage conditions using the methodology presented in this paper could be significant higher than those calculated using the PTI method.

Residential foundations such as slabs-on-ground are often constructed on unsaturated, expansive soils and have to resist deformation associated with external loads and changes in matric suction in the soil. Residential foundations are lightly loaded, and therefore the displacements are mostly due to changes in matric suction. Changes in matric suction can occur as a result of variation in climatic conditions, change in depth of water table, water uptake by vegetation, removal of vegetation or the excessive watering of a lawn. The theoretical context for the analysis of slabs-on-ground placed on expansive soils involves the numerical modeling of moisture flow through the soil and the associated deformations associated with changes in the stress state. The theory associated with moisture movement and volume change has been formulated and widely accepted as part of unsaturated soil mechanics. This paper presents the use of the finite element method to address this complex problem. Various aspects of the separate water flow and deformation aspects will be presented. The combination of stress analysis and seepage analysis is presented. An analysis of the final estimated deformations provides insight into this complex process.

Relationships between swell potential and soil index properties are commonly used in soil design. It was found that the Expansion Index as determined by the modified Arizona method, EI_AZ, correlates very poorly to both Plasticity Index., PI, and percent passing US sieve number 200, P₂₀₀, when considered separately. The product of these two parameters, however, wPI, improves the correlation from R² values of 0.41 and 0.27, respectively, to 0.67.

The interaction of the soil and the slab foundation resting on it has been investigated by many researchers. The models used can only handle the mechanical behaviors at the soil-structure interface while the volume change of the soil is a fully coupled hydro-mechanical stress (consolidation) problem. As a result, all of these methods can not realistically simulate the moisture-dependent transient interaction between the soil and the slab. In this paper, coupled hydro-mechanical stress jointed (contact) elements are proposed to simulate the soil structure interaction between the soil and slabs, and the coupled consolidation theory is used to simulate the volume change of the soil. The proposed technology is not brand new. However, it provides a way to simulate the soil-structure interaction in a more comprehensive and realistic way. The slab can be of any shape and the ground soil can be either saturated or unsaturated. A real soil, completely characterized by tests on samples taken from a site in Arlington, Texas is used to demonstrate the proposed technology.

A two-dimensional shrink-swell model is presented in the paper to address the problem of moisture flow and associated soil deformation due to seasonal cycles of wetting and drying in expansive soil mass. The numerical model performs the surface vertical movement prediction in a manner of sequentially coupled flow-displacement analysis approach. The theoretical formulations of moisture flow and stress-strain components of this model are first described, while the log-linear Mitchell equation is employed for the moisture diffusion computation and elastic Young's Modulus E is expressed as a function of matric suction h_m and net mechanical stress δ_m. An example problem of surface swelling-shrinkage calculation is given to illustrate the proposed shrink-swell model and the use of a finite element computer program FLODEF which is developed based on this model.

In order to define correlation of shear strength with matric suction which further depends on water contents for unsaturated soils, it is recognized by examining four-grain packed system in this paper that in additional to the conventional matric suction distributed on contact area of water, there exists a normal tension force which is distributed along the saturation line on grain surface due to surface tension while the air-water interface overlaps on particle surface. The line-distributed tension can be transformed equivalently into a distributive tension stress which is defined as tension suction and will pose the similar effect on shear strength and deformation as that of matric suction. According to the actual contact states of air-water interface with the surface of soil particles, the unsaturated soil with the characteristics of continuous water with continuous air (CWCA) can be further classified into the overlapping and non-overlapping unsaturated soil. The tension suction appears only in the unsaturated soils of discontinuous water with continuous air (DWCA) and overlapping continuous water with continuous air (CWCA). The thermodynamics and geometric conditions are considered to establish a set of nonlinear equations for defining the air-water interface shape between two mono-sized spheres. These nonlinear equations are solved by iterative procedure for different combinations of contact angles and water contents to obtain theoretical matric and tension suctions in a four-grain packed system with isolated pore water. Then the equivalents of matric and tension suctions are made in order to take both intensity and action area of matric and tension suctions on grain surface into consideration. For the unsaturated soils of DWCA with different contact angles, it is concluded that the ratio of tension suction to matric suction increases from a small value to 2.5 or 4 with increasing saturation degree. It is implied that the tension suction cannot be overlooked. The matric suctions falls down sharply with increase of saturation degree whereas tension suction is almost kept constant and is larger than matric suction from a saturation degree around 8 percent. The resultant suction of both matric and tension suctions in the meaning of equivalence varies slightly with saturation degree and may be taken as a constant which can well agree with the experimental observations that shear strengths almost keep constants with increasing matric suctions as soils approach to completely dry.

Aggregated soils, such as agricultural soils, are under consideration as representatives of structured soils with double structure, including macro and micro porosity. In this paper, the main characteristics in compression behaviour of these soils are first addressed. The possible mechanisms of structure degradation are presented with special emphasis on the role of microstructure. The experimental technique of neutron tomography followed by image processing is used to quantify structure degradation during compression. A state parameter is introduced to describe the soil structure degradation and an evolution law is proposed as a function of the associated plastic strains. This parameter is then used to develop a compression model formulation and to propose an evolution rule for the apparent preconsolidation pressure for aggregated soils.

Changes in volume and pore space induced by the shrink-swell behavior of clay minerals present a challenge to predictive modeling of hydraulic properties of clayey soils. We present a pore-scale framework that combines physico-chemical processes with pore-geometrical, hydrostatic, and hydrodynamic considerations toward prediction of constitutive hydraulic relationships for swelling porous media. Variations in pore space are modeled by considering the soil clay fabric as an assembly of colloidal-size tactoids with lamellar structure. The arrangement of clay tactoids and the spacing between individual lamellae are functions of primarily clay hydration state quantifiable via the disjoining pressure that is dominated by a large electrostatic repulsive component. Solution chemistry and clay type are also considered. Silt and sand textural constituents are represented as rigid spheres interspaced by clay fabric in two basic configurations of "expansive" and "reductive" unit cells. Bulk soil properties such as clay content, porosity and surface area serve as constraints for modeling pore-space geometry. Liquid saturation within the idealized pore space is calculated as a function of chemical potential considering volume changes due to clay shrink-swell behavior. Closed-form expressions for prediction of unsaturated hydraulic conductivity are derived from calculations of average flow velocities in ducts, parallel plates, and in corners bounded by liquid-vapor interfaces, and invoking proportionality between flux density and hydraulic gradient. Measurements of saturated hydraulic properties are used to evaluate model predictions and lay the foundation for upscaling considerations to sample and profile scales. Advanced flexible-wall permeametry is employed to measure permeability of clay-sand mixtures with mono- and divalent ionic solutions. Preliminary pore-scale model calculations show favorable agreement with measured permeability values for a wide range of clay contents and varying ionic solutions.

Interparticle repulsive pressures in clays are commonly determined from physico-chemical considerations. In this paper, the interparticle repulsive pressures due to double layer repulsion for several compacted clays were determined from the diffuse double layer theory. The interparticle attractive pressures were also determined following the van der Waals-Hamaker equation. For the compacted clays, the additional repulsive pressures at very close particle spacing were then determined by subtracting the calculated net interparticle double layer repulsive pressures from the experimental swelling pressures of the clays. The additional repulsive pressure was found to come into effect for a platelet spacing of 1.5 nm and smaller and was found to be very sensitive below a spacing of 0.7 nm. There appears to be a good agreement between the additional repulsive pressure and surface charge density of the clays.

Bulk constitutive relations for unsaturated soils should be based upon and not violate the fundamental micro-mechanical interactions that occur. Thus, knowledge of these basic interactions is required in order to fully derive or assess a given constitutive relation. In particular, at low water contents liquid bridges can be assumed to exist between pairs of particles; knowledge of the interactions due to these bridges of varying liquid volume is necessary. To this end, a micro-mechanical experimental method to determine the capillary forces between individual soil particles at varying liquid content is proposed. A rough glass sphere (160–325 μm diameter) is glued to a glass fiber cantilever (30 μm) while another is glued to a rigid capillary tube. A small drop of liquid is applied, and the capillary force is measured by the deflection of the fiber cantilever as the particles are brought together and separated. The position of the base of the cantilever is fixed, while the movement of the particle attached to the rigid capillary tube is controlled by a high precision micro-manipulator. Displacements are tracked using high-speed digital video microscopy under an inverted light microscope. Image processing is used to analyze the frames taken during the experiments. Experiments with unevenly-sized spheres compare well to known analytic results. Liquid volume and mean curvature of the meniscus as functions of filling angle are well described analytically. Forces measured range from 25 to 50 micro-Newtons depending on the liquid content. A force characteristic curve is presented to describe this dependence. Viscosity effects are as yet undetermined as are detailed measurements of the liquid-solid contact angle hysteresis.

Soils from lacustrine zones are initially subjected to a sedimentation process, this stage continues with desiccation and consolidation in a partially saturated state. When the soils are highly compressible, this occurs simultaneously with soil cracking. This is the case of the "Sabana de Bogotá". This zone is a great depth lacustrine deposit (600 m approximately) with a surface composed of highly compressible organic clays. These soils are affected by desiccation caused by the atmosphere, vegetation and subsurface water extraction. This process comes together with crack generation that may reach 7 m depth and 20 cm wide. Soil surface desiccation is a complex phenomenon where mass and heat interchange with atmosphere has a fundamental importance. In particular, some main factors that influence this process are: solar radiation, wind speed, temperature, precipitation, evapotranspiration, etc. Most of existing numerical model simulations for water flow in partially saturated soils consider the ground surface as a boundary with characteristics independent from atmospheric variables. In this paper a numerical model developed specifically to simulate the interaction between atmosphere and soil is presented. This model is applied to the conditions of the "Sabana de Bogotá".

Evaporative surface fluxes can be a dominant factor in sub-surface moisture distribution and movement, especially in dry climates. Evaporation depends on both soil characteristics and atmospheric conditions, and is non-trivial to accurately measure. Consequently, evaporative fluxes are often numerically estimated. The purpose of this paper is to evaluate the predictive capability of a widely used numerical approach, and provide a method for interpreting transient evaporative fluxes from simple moisture content measurements. A parameter estimation analysis of the van Genuchten parameters α, n and Ks (saturated hydraulic conductivity) using an inverse solution and volumetric soil water content data was developed. In this study rather than using water potential data, a Tikhonov regularization method was used to deal with the ill-posed nature of the problem. The water flux was simulated using the computer program UNSAT-H. The code was then combined with the program DAKOTA (Design Analysis Kit for Optimization and Terrascale Applications) for the parameter study. By minimizing the proposed objective function through iterative calculations, the resulted evaporation from the simulation, shows a good approximation to the measured evaporation on the field.

This paper explores issues related to the numerical simulation of moisture migration patterns associated with vegetation. Particular attention is given to prediction of soil suction variations in the unsaturated zone and in the vicinity of a single mature tree. A preliminary investigation is presented based on the use of Richard's Equation for unsaturated moisture flow incorporating a sink term. A numerical solution has been achieved via the finite element method for spatial discretisation along with a finite difference time-marching scheme. The model is first verified against a one-dimensional simulation for a uniform crop cover. An axisymmetric solution is then developed to represent water-uptake near an established tree. Field data and non-linear hydraulic properties have been sourced from independent published data. Numerical results have been validated by direct comparison to experimental/field data. The results obtained show a realistic moisture migration pattern can be achieved in this manner.

Bioengineering aspects of native vegetation are currently, and rapidly, being evolved to improve soil stiffness, slope stabilisation, and erosion control. Apart from the reinforcement effect, tree roots establish sufficient matric suction to increase the shear strength and stiffness of the soil. This paper looks at the way, vegetation influences soil matric suction, shrinkage, and ground settlement. A mathematical model for the rate of root water uptake that considers ground conditions, type of vegetation and climatic parameters, has been developed. Based on this proposed model, the distribution of moisture and the matric suction profile adjacent to the tree are numerically analysed. The model formulation is based on the general effective stress theory of unsaturated soils. Field measurements taken from literature published previously are compared with the authors' numerical model. The predicted results calculated using the soil, plant, and atmospheric parameters contained in the numerical model, compared favourably with the measured results, justifying the assumptions upon which the model was developed.

The simplified energy balance method for estimating soil-water evaporation using infrared thermometry requires the use of assumptions often found to be invalid. Without these assumptions, the method is too expensive in labor or equipment to be applied at remote locations, and eliminates the possibility of using remote sensing to collect the necessary data. In this study, a method is described that eliminates the necessity of using these assumptions, while retaining model simplicity and versatility. The proposed method uses an optimized fuzzy learning from example (OFLFE) algorithm to develop the model. The OFLFE algorithm develops a fuzzy rule base using a learning from example approach, fine tunes the rule base using a recursive least squares approach, and independently optimizes the rule base using constrained nonlinear optimization. K-fold cross-validation was used to test the model. The results indicate that the fuzzy model can predict evaporation better than the existing simplified energy balance method, with a reduction in root mean squared error (RMSE) of nearly 50 percent.

2-D finite element analyses are used to study how, during rainfall, the changes in negative pore-water pressures in a slope are controlled by the hydraulic properties of the soil and the initial conditions within the slope. For a "coarse grain soil", the zero pore-water pressure contour moves gradually from the toe upslope with time, while in a "fine grain soil", the rate of suction loss is nearly the same all along the slope. These different infiltration patterns lead to different potential failure surfaces and failure mechanisms. Shallow translational sliding is observed along the slope composed of fine grain soil while local rotational sliding occurs at the toe of the slope in the coarse grain soil model. A numerical back-calculation of a landslide in Norway is also presented. The 38 degree slope failed in November 2000 after an extremely wet fall. The slope is computed to be stable under normal rainfall conditions but to fail when subjected to the rainfall of 2000.

Numerous research studies are being carried out to examine the behaviour of unsaturated slopes subjected to rainfall. In many cases the research is a consequence of transportation infrastructure that has shown instabilities triggered by specific rainfall events. This paper presents results of generic modeling to assess the impact of the thickness of a weathered crust on the susceptibility of a typical highway embankment slope subjected to rainfall induced instability. Engineered embankment fills and underpass cut slopes will undergo stress relaxation after construction. This relaxation near the surface and exposure to the environment can result in development of a secondary structure with decreased strength and increased permeability. The extent of this weathered zone depends on the geometry of the engineered slope and the nature of the soil. The model developed in this work presents generic modeling results that examine the importance of the weathered zone by examining the sensitivity of some generic slopes with varying weathered zone thickness under specified rainfall events. The results provide some insights as to the importance of characterizing the characteristics of the weathered zone for this class of problem. Specifically the results show that the hydraulic properties are more important in terms of assessing the susceptibility of slopes to this mode of failure as opposed to the strengths. Results also show that in this generic modeling the dissipation of suction was more important that any mounding of the groundwater level.

Progress made over the past decade on the direct measurement of suction has clarified the design requirements of high capacity tensiometers and the procedures surrounding the rather difficult process of tensiometer saturation. However, many challenges still exist relating to the direct measurement of matric suction; particularly in the interpretation of suction data. In this paper, we present a case study of tensiometer interpretation derived from apparent confusing observations of matric suction which had been observed in a centrifuge model test of an unsaturated slope. Using additional observations of tensiometer behaviour in the laboratory, the apparent confusing nature of this data is explained, and lessons are drawn regarding good practice in both model tests and field instrumentation.

Centrifuge model tests were conducted for a rockfill dam model (central core zone type) in order to investigate behavior of dam during first reservoir filling and the typical test result was simulated by using a consolidation analysis method coupled with an elastoplastic model for unsaturated geo-materials. The main behavior obtained from the centrifuge tests was that (1) large amounts of settlements due to wetting within the upstream rock zone were monitored, (2) the crest of dam at first moved toward upstream and then returned toward downstream, and (3) differential settlements between core and rock zones occurred and cracks along the dam axis were measured. The behavior of the dam model during reservoir filling was simulated. The simulation results showed that (1) the settlements calculated agreed roughly with those measured and the settlements due to wetting could also be estimated, (2) the horizontal movements calculated were fairly larger than those measured, and (3) the cracks monitored in the model test occurred at the places where tension stresses were calculated in the simulation. Thus, our method presents an effective approach to analyzing the behavior of the fill-type dams during reservoir filling periods.

The green house effect has brought change in climate all over the world. In Japan many slope failures due to rainfall often occur in the rainy season. Therefore, it is necessary to investigate the mechanical behavior of unsaturated soil in order to clarify the slope failure mechanisms due to heavy rainfall. In this paper a soil tank test on sandy soil called Shirasu was carried out to investigate the failure mechanism of an unsaturated slope due to the increase in degree of saturation. The numerical simulations using 2D unsaturated-saturated seepage analysis and slope stability analysis is then performed to compare with the experiment in order to prove the validity of numerical simulations at the model scale. It is found from our research results that the proposed models can be used to simulate the seepage behavior of unsaturated Shirasu and to estimate the change in safety factor of the slope with elapsed time due to rainfall.

This paper shows the results of a fully coupled analysis of the construction of an earth fill compacted with an unsaturated soil subjected to structural collapse. Initially, the formulation of the problem is presented based on the finite element method. Then, the constitutive relations necessary to describe the behavior of the phases of an unsaturated soil are described. Emphasis is given on how to appropriately impose the boundary conditions related to the water flow continuity equation during the construction period. An enhanced limit equilibrium approach is used to compute the safety factor, using the stress and pore pressure fields previously obtained in the finite element analysis. The results show that uncoupled analyses can be against safety in such cases, because they neglect load transfer mechanisms that can only be appropriately simulated with a coupled solution.

Unsaturated soils are three-phase porous media consisting of a solid skeleton, pore liquid, and pore gas. The overall behavior of unsaturated soils is influenced not only by these three bulk phases, but also by the interfaces between them. The coupled mathematical equations governing the dynamics of unsaturated soils can be derived based on the theory of mixtures with interfaces (TMI). Solution of these equations requires large computational resources. A framework-based parallel finite element code was developed for efficiently solving such large problems. Simulation of an unsaturated soil embankment subjected to earthquake loading is presented to show the capabilities of the computer code developed. A reduced set of governing equations, neglecting the relative accelerations and velocities of fluids, is also solved to study the effects of these quantities on the overall dynamic behavior of unsaturated soils. The reduced set of equations gives comparable results to that of the complete set of equations for a low permeable soil considered in this study.

There are about 210,000 small earth dams in Japan that are important water resources for irrigation. Collapse of these dams induced by seepage due to heavy rainfalls occurs almost every year, since most of the dams are old and constructed of sandy soils with low dry densities. In order to investigate the behavior of the embankment dams during seepage, a large-scale model test was carried out. It was found from the test results that the embankment model showed a complex behavior combining shear deformations with saturation collapse. A simulation of the model test was conducted using saturated and unsaturated consolidation analyses with an elastoplastic model for unsaturated soils. The analysis simulated not only pore water pressures, deformations but also conditions of the failure of the embankment. Thus, the simulation results indicated that the method of analysis used here was suitable for estimation of behavior of embankment dams during seepage.

By analyzing the work input in unsaturated granular materials, Houlsby (1997) obtained the simplified version of Bishop's equation as the effective stress equation for unsaturated soils. In his paper, Houlsby idealized the structure of unsaturated soils as a set of particles linked together by water meniscus. However, a real soil generally exhibits a bimodal structure where unsaturated as well as saturated packets of particles can be found. Based on the analysis made by Houlsby (1997), but now considering a bimodal structure of the soil, it is possible to establish an equivalent stress that accounts for the influence of suction on the strength of unsaturated soils. This equivalent stress has the same structure as Bishop's equation. However, this analysis shows that Bishop's parameter χ depends not only on the degree of saturation of the unsaturated fraction of the soil but also on the volumes of the saturated and unsaturated fractions.

Currently many researchers consider that only one stress state variable and two stress state variables are needed for determining the behavior of saturated and unsaturated soils, respectively. A thermodynamic analogue to the consolidation of soil is used to demonstrate the need of a minimum of two stress state variables for both saturated and unsaturated soils. Some topics related to the use of stress state variables such as constitutive surfaces and constitutive laws and the governing differential equations for saturated and unsaturated soils are discussed. The stress state variables needed to describe the volume change behavior of both saturated and unsaturated soils are discussed.

The significance of the thermodynamic potential and extensive variables is outlined and leads to identification of the role and use of enthalpy in a multi-phase material comprising solid particles, liquid and gas. This is presented in terms of triaxial stress conditions. Enthalpy considerations are then used to develop equations that describe the equilibrium stress regime in an unsaturated soil and demonstrate the dependence on the relative volumes of the phases. The analysis supports and expands on previously published work.

The fundamental considerations and assumptions in laboratory equilibrium tests on soils in the triaxial cell are examined using the principles of virtual work and thermodynamics. Compliance with the laws of thermodynamics is essential and the first law of thermodynamics necessitates that the work equation, thus the stresses, may be derived from the thermodynamic potential. At equilibrium the thermodynamic potential is a minimum and it is shown that this can be written in a similar form for both isotropic and anisotropic loading conditions but for the latter the mean stress replaces the isotropic pressure. The significance of the extensive variable terms making up the thermodynamic potential is also described. The analysis is applicable to soils at any degree of saturation.

Four closed-form water-retention curve functions were validated using the previously measured water retention curve of 4 sand fractions and their 6 two-component mixtures. Out of the models of Gardner (1958), Fredlund - Xing (1994) and the simplified and non-simplified models of Van Genuchten (1980), the non-simplified model of Van Genuchten model fitted with the smallest error to the tested soils.

A simple formulation of capillary hysteresis with internal state variables is developed. The new model describes all types of (primary, secondary, and higher-order) scanning curves within the main hysteretic loop of the moisture retention curve. Provided that the main wetting and drying curves are analytically obtained, the model requires only one additional parameter as input. It is shown that the model predictions agree well with the experimental data on various soil types available in the literature. The proposed model is easy for implementation into a computer code that is used in the analysis of immiscible two-phase flow through porous materials.

Soil water characteristic curves (SWCCs) relate matric suction and water content in unsaturated soils. SWCCs exhibit hysteresis during drying and wetting cycles and the behavior is path dependent. An elastoplastic model is developed using the bounding surface plasticity theory to predict the SWCCs including the hysteresis. The SWCCs predicted for a fine Ottawa are compared with measured results obtained using an automated parallel miniature pressure cell system and good comparisons are achieved. Influence of key model parameters on the predicted behavior is illustrated.

This paper deals with the evaluation of a model for the soil water retention curve based on fractal geometry. The model departs from the proposal by Tyler & Wheatcraft (1990) which assumes that grain size distribution could be modeled considering the fractal principles, that is, self-similarity and dimension. In this paper, the original formulation is substituted by an alternative that considers the pore distribution in the soil. The adapted model was tested against some results from different soils and provided a good fit, however the fractal dimensions obtained did not show any relationship with soil texture.

A twofold model that simulates the porous structure of soils is presented. The model considers that the voids in the soil mass are composed of only two kinds of elements: the sites or cavities and the bonds or throats. To each one of these elements, a proper size distribution is assigned according to the pore size distribution data. Based on these assumptions and making use of the probability theory, the equations corresponding to the main hysteresis loop during wetting or drying are established. Finally, these results are compared with experimental results including the main hysteresis loop of the soil-water characteristic curve and the pore size distribution of the soil.

The water retention curve (WRC) (or soil-water characteristic curve SWCC), representing the relationship between volumetric water content and suction, is essential to describe water flow through unsaturated soils. However, the laboratory (or field) measurements of the corresponding parameters can be time consuming and difficult to perform in some instances. This is particularly the case when both drying and wetting paths need to be assessed. An interesting alternative, at least at the preliminary stage of a project, is to estimate the WRC using a predictive model based on the basic geotechnical properties of the media. However, hysteresis effects are not typically taken into account by most existing predictive models. The authors propose in this paper an extended version of the modified Kovács (MK) model that integrates the hysteresis of the WRC, along different paths. Model predictions are compared to experimental data for different granular soils. The predicted results show good agreement for the drying curve and for both the main and primary wetting curves.

An analytical solution is obtained for transient heat and moisture transport in unsaturated bentonite buffer in the Laplace transform domain, a numerical inversion scheme of Crump's method is used to obtain the time-domain solution. The coupled effect of thermally driven moisture transport is especially investigated because of its importance to alter the flow field in clay-rich and low-permeability bentonite buffer used in radioactive waste repository. The practical background is based on the case of an engineering bentonite buffer placed in a drift excavated in rock in the context of underground disposal of high-level radioactive waste.

The design of repositories for the isolation of radioactive waste generally contemplates the use of expansive clays as a buffer material. The FEBEX bentonite is the material adopted to build the compacted clay barriers of the FEBEX project (ENRESA, Spain). To study the Thermo-Hydro-Mechanical (THM) behavior of the FEBEX bentonite an experimental program has been carried trying to reproduced, as close as possible, the actual repository conditions. In field conditions the clay barrier will be hydrated (due to the water coming from the host rock) under confined conditions and will also undergo processes of heating (induced by the heat-emitting waste), up to maximum of 100 degrees C (this is according to the Spanish concept for the isolation of high level radioactive waste). Based on experimental data, constitutive models have been developed to explain and reproduce the observed THM behavior of the FEBEX bentonite. The constitutive laws incorporate key aspect to model the complex behavior of this highly expansive material, such as, large swelling under wetting, yielding, stress path dependency, clay-fabric changes, among others. However, an unplanned episode in the mock-up test has provided the opportunity to analyze the behavior of the FEBEX bentonite under thermal conditions quite different from the design ones. The mock-up test is a large scale heating experiment, which has provided relevant data to advance in the knowledge of the THM behavior of expansive clays. In this test the barrier is hydrated uniformly from all around the perimeter with an applied water pressure of about 0.5 MPa. Simultaneously, the barrier is heated maintaining a constant temperature of 100 degrees C at the contact between heaters and bentonite.

A large in-ground experiment, the Isothermal Test (ITT), was conducted in the Underground Research Laboratory to examine how a large mass of dense sand-bentonite buffer would take on water from a surrounding rock mass. The ITT was installed in a 1.24-m-diameter by 5-m-deep borehole in granite. The bottom two metres were filled with an in situ compacted sand-bentonite material. This material was capped by a 1.25-m-thick mass of concrete and was left undisturbed for 6.5 years. During this period, the pore water pressures in the buffer and the surrounding rock were continuously monitored. Numerical simulations were conducted using the commercially available finite-element program CODE_BRIGHT in an attempt to model the evolution of the pore water pressure in the unsaturated sand-bentonite buffer and the surrounding rock. The influence of the excavation-induced damage zone on the evolution of the hydraulic response in the sand-bentonite material and the rock was also examined. The water seepage from the rock to the borehole and the pore water pressure in the rock and the buffer were compared with the measured data.

For isothermal and non isothermal behavior of multi-phase porous media, a suction-based mathematical approach has been developed by the first author since 1990. In this frame of approaches, the "independent variables" are chosen as state variables and the strong coupling effects are considered via "state surface" notion. Temperature-dependent state surface formulations are drived for variation of void ratio and degree of saturation within porous media for the thermal cases. The coupling effects of temperature and moisture content on deformation of skeleton, and the inverse effects are included in this model in a complete manner. In this paper, after a brief description of thermohydromechanical behaviour of unsaturated media, the extension of THM models to thermo-hydro-chemo-mechanical (THCM) models are discussed for the general conditions. Theoretical and numerical developments for this complex formulation are presented. The integration of new development in θ_STOCK software developed previously by the first author has been described. Some validation and application cases are presented to illustrate the correctness of the new developments and the efficiency of finite element θ_STOCK code.

A granular mixture of highly compacted bentonite pellets has been proposed as a suitable impervious engineered barrier for waste isolation in deep underground storage for nuclear waste. A demonstration project (Engineered Barrier, "EB" Project) was set up to investigate the feasibility of the concept including both, a real scale test and a comprehensive laboratory test program of the pellet mixture. The aim of the laboratory test program was to characterize the hydro-mechanical response of the buffer material under conditions similar to that occurring in the real scale test. A special suction controlled oedometer cell was designed and built for the purposes of the project. It was found that specimen behaviour is controlled by the rate of application of moisture at the specimen boundaries. Two extreme cases have been investigated: either a "slow" or a "fast" application of suction changes. Measured results have been interpreted within the framework of a double structure plasticity model. Away from equilibrium, the capillary suction prevailing at the large macropores, is different from the microstructural suction, which controls pellet expansion (or shrinkage). A local transfer of water between both levels of structure and the associated changes in macro and microporosity explains the nature of observed phenomena.

A method for predicting capillary flow from a water table up into an overlying unsaturated soil is presented. The method is based on Philip's equation of vertical upward infiltration. Three parameters, b₀, b₁ and b₂, which are functions of moisture content, are crucial for the predicting equation. These parameters can be obtained by inducing capillary flow in three soil columns for different periods of time. Three soil samples were tested: two samples were sandy loam and the other was sandy clay loam. The experiments were performed using columns of dry soil and inducing capillary rise from the bottom ends. Five periods of capillary flow were conducted for each soil sample. The data from three periods (1, 2, and 4 hours for sandy loam and 1, 2, and 4 days for sandy clay loam) were used to calculate parameters b₀, b₁ and b₂. The data from two other periods (3 and 6 hours for sandy loam and 3 and 6 days for sandy clay loam) were used to test the accuracy of the method. The period of 3 hours (or 3 days) gave better agreement than that of 6 hours (or 6 days). The 3-hour period tests were interpolated while the 6-hour period tests were extrapolated.

Cavity expansion theory has long been used to interpret cone penetration test results in saturated soils. The pressure required to expand a spherical cavity approaches a limiting value at large strains, and this limiting value is thought to be related to cone penetration resistance. However, there is very little experience in interpreting the cone penetration test in unsaturated soils using cavity expansion theory. Practitioners often use correlations developed for saturated soils which may lead to non-conservative misrepresentations in estimated soil parameters. This paper gathers the results of cone penetration tests in two unsaturated quartz sands and highlights the similarities between spherical cavity limit pressure and cone penetration resistance. In some cases the effects of suction causes both the cone penetration resistance and cavity limit pressure to be doubled. The contribution of suction to effective stress is back-calculated using these results.

This paper presents a simplified analytical model for desiccation cracking of clay layers in laboratory tests. An analytical solution is presented for the maximum tensile stress developed at the mid section of the layer. As the desiccation progresses, the clay layer will crack at the middle when the maximum tensile stress exceeds the tensile strength of the clay. For further desiccation, the analytical solution is recursively applied to model the subsequent subdivision of the clay layer through formation of secondary and tertiary cracks. The model was applied to predict the laboratory desiccation tests of long-thin clay layers that produced parallel cracking. The non-linear material properties of clay and the basal interface were represented by empirical correlations with the moisture content. The free shrinkage strain was found to be linearly correlated to moisture content reduction. The model was able to predict the development of cracks during desiccation reasonably well. In particular, it was highlighted that once a crack initiates, subsequent subdivision of the layer takes place rather rapidly until a number of cracks that is in equilibrium with the test conditions has emerged.

This paper provides a summary of nineteen empirical procedures or techniques that are available for predicting or estimating the shear strength of unsaturated soils. While six of these procedures use the soil-water retention curve (SWRC) as a tool; the remainder thirteen procedures are mathematical formulations for the prediction or estimation of the shear strength of an unsaturated soil. A comprehensive analysis is undertaken to predict the variation of shear strength with respect to matric suction of twenty soils using the six prediction equations that use SWRC. The studies presented in this paper highlight the suitability of these equations for successful predictions of the shear strength of the unsaturated soils.