Geotechnical Earthquake Engineering and Soil Dynamics V: Liquefaction Triggering, Consequences, and Mitigation

This paper discusses three aspects of nonlinear dynamic analysis (NDA) practices that are important for evaluating the seismic performance of geotechnical structures affected by liquefaction or cyclic softening: (1) selection and calibration of constitutive models, (2) comparison of NDA results using two or more constitutive models, and (3) documentation. The ability of the selected constitutive models and calibration protocols to approximate the loading responses important to the system being analyzed is one of several technical factors affecting the quality of results from an NDA. Comparisons of single element simulations against empirical data for a broad range of loading conditions are essential for evaluating this factor. Critical comparisons of NDAs using two or more constitutive models are valuable for evaluating modeling uncertainty for specific systems and for identifying modeling limitations that need improvement. The utility of an NDA study depends on the documentation being sufficiently thorough to facilitate effective reviews, advance best practices, and support future reexaminations of a system’s seismic performance.

The influence of the non-liquefied crust that overlies a liquefied deposit on the severity of surficial liquefaction manifestations has been noted for several decades. In 1985, Ishihara proposed a generalize relationship relating the thicknesses of the non-liquefied crust and of the liquefied stratum to the severity of surficial liquefaction manifestations. Although subsequent studies using data from multiple earthquakes give credence to Ishihara’s relationship, the implementation of the procedure is tenuous for all but the simplest of profiles. In an effort to overcome issues with implementing the Ishihara relationship, new procedures have been proposed for predicting the severity of surficial liquefaction manifestations. The efficacies of two of these procedures are currently being assessed in a study using unique case history data from the 2016, M_w5.7 Valentine’s Day earthquake that impacted Christchurch, New Zealand. Preliminary results from this study show that both procedures yield predictions that are in accord with field observations. However, the final results from the ongoing study are expected to more fully assess the efficacies of these procedures.

Dense granular columns are commonly used to reduce the risk of liquefaction or its associated ground deformations in gentle slopes through: 1) enhancing drainage; 2) providing shear reinforcement; and 3) densifying and increasing lateral stresses in the surrounding soil during installation. However, the independent influence and contribution of these mitigation mechanisms on excess pore pressures, accelerations, and lateral and vertical deformations is not sufficiently understood to facilitate reliable design. This paper presents the results of a series of dynamic centrifuge tests to evaluate the influence of dense granular columns on the seismic performance of gentle slopes (e.g. <5%). Granular columns with greater area replacement ratios (A_r about 20%) were shown to be highly effective in reducing seismic settlement and lateral slope deformations, owing primarily to expedited dissipation of excess pore water pressures rather than shear reinforcement. The influence of granular columns on accelerations in the surrounding soil depended on the column’s A_r and drainage capacity. The insight from these experiments aims to improve our understanding of the mechanics of liquefaction and lateral spreading mitigation with granular columns.

This paper presents a semi-coupled method to predict the occurrence of liquefaction in the city of Urayasu caused by the 2011 Tohoku-Pacific Ocean earthquake (Mw=9.0). The presented method is a combination of a linear equivalent computation and a formulation for computing plastic volumetric strain based on the works of Byrne (1991) and Wu (1996). Further, this approach was conducted to simulate the seismically induced soil liquefaction and volumetric settlements before triggering of two representative soil models: one representing a non-liquefied ground model chosen from the old town of Urayasu and another a liquefaction-prone ground at the reclaimed area of the city. Fully undrained behavior was assumed during calculation, so the effect of seepage are not taken into account and then results are restricted to the shaking duration before triggering liquefaction. The computed zones of liquefaction or non-liquefaction were compared and in a good agreement with observations. Predicted values of settlements are in a good agreement with in-situ measured values at non-liquefied locations.

The importance of predicting ground deformation in loose, saturated granular soils has been widely recognized for a reliable evaluation of liquefaction damage. A procedure is proposed in this paper for the evaluation of post-cyclic consolidation settlements, as a result of volumetric strains induced by the dissipation of excess pore pressure. A stress-based model is first adopted for generating the excess pore water pressure in 1D free-field conditions, allowing for an effective stress analysis according to a loosely coupled approach. Then, the post-cyclic settlement is simply calculated integrating the vertical strains. To this aim, by considering a well-documented case history in which an extremely small settlement was observed upon seismic excitation, soil stiffness is estimated on the basis of either CPT data or shear stiffness decay curve, to show the effect of modelling hypothesis on the results. Both approaches result into a value of the settlement close to the observed one and much lower than that calculated using a well-established empirical procedure.

Noting that injection of expanding resin mixtures into the ground has long been used worldwide for the re-levelling of buildings that have suffered foundation deformations, the aim of this study was to examine whether the same process could also be a viable form of ground improvement, primarily for liquefaction mitigation. A controlled field investigation was carried out into the efficacy of this ground improvement technique, in which three test sites were selected for the construction of resin-injected test panels. The pre- and post-improvement soils at all three sites were assessed using a variety of methods. The results of the study show that significant improvements in soil density and stiffness were achieved, thus demonstrating that resin injection is a viable ground improvement method in the right soil conditions.

Shallow-founded structures on liquefiable deposits are often rendered unserviceable due to excessive settlement and rotation incurred during earthquakes. Recent research has exposed the mechanisms by which large strains occur in liquefied soils, leading to foundation’s excessive settlement and rotation. The development of these mechanisms may be retarded with appropriate mitigation. Prefabricated vertical drains (PVDs) are a popular and economical mitigation measure, which have been shown to limit permanent structural settlement and tilt. However, while the popularity of these drains continues to grow, there are concerns regarding their usage in dense urban environments. This paper compares the performance during centrifuge tests of an isolated 3-story model structure, an identical structure mitigated with a network of PVDs, and a pair of dissimilar structures separated by 3 m, with and without PVDs around the smaller structure. Results show that PVDs should be designed with great caution in dense urban environments, due to their potentially negative impact on the performance of adjacent unmitigated structures.

Standard penetration test (SPT) data is typically adjusted for energy transmission effects using energy measurements at the top of the drill string. This energy ratio (ER) measured at the ground surface is assumed to also be applicable at the sampler at the bottom of the drill string. However, several studies since the 1970s have shown that there is appreciable energy attenuation along the drill string. For deep borings, the accumulated energy loss can be significant, and if ignored, may lead to incorrect conversion of the measured blow counts to energy-normalized N₆₀ values. As a result, the computed N₆₀ values may be erroneously high. This paper presents the results of a study on Stone Canyon Dam, a 225-foot high zoned earthfill dam located in Los Angeles, California. Downstream portions of the dam rest on up to 75 feet of alluvium consisting of interbedded gravely sands and clays. Liquefaction evaluation of the alluvium was based on side-by-side SPTs and cone penetration tests (CPTs) to depths of up to 200 feet. Initial SPT-based and CPT-based liquefaction analyses were found to produce drastically different cyclic resistance ratio (CRR) values for the alluvium. However, this discrepancy between the SPT- and CPT-based CRR values was largely resolved when energy loss along the SPT drill string was taken into account.

Cyclic liquefaction is a phenomenon in which soil deposits formed by loose and saturated sands, lose their shear resistance due to dynamic loads, as earthquakes, sometimes causing the collapse of structures. This phenomenon is a hazard in different cities of the world, as is the case of Santiago de Cali (Colombia), a city that has large alluvial deposits due to its seven rivers. Currently, the liquefaction potential is evaluated by field tests that can be expensive and difficult to perform in densely urbanized areas. Due to this drawbacks, this paper proposes a simple and versatile methodology to evaluate the liquefaction potential in large areas based on the vulnerability index ($K_g$) proposed by Nakamura in 1997 (hereinafter Nakamura Methodology). This methodology consists of calculating the vulnerability index using the fundamental frequency ($F_g$) and amplification factor ($A_g$) determined using HVSR method. The $K_g$ index quantifies the probable damage in the soil produced by earthquakes, and for sand deposits, $K_g$ can be correlated with liquefaction potential. For this research, the liquefaction potential was evaluated using Nakamura methodology in 34 points of Santiago de Cali, within an approximate area of 12 km². The results were compared with those of SPT and CPT tests performed in the same area. The methodology showed a match of up to 82% compared to conventional methods. In conclusion, Nakamura methodology is a viable alternative for liquefaction hazard mapping of large areas and could be very useful in places where there are limited funds or the conditions for more robust testing are not available.

Recent experiences (e.g. New Zealand and United States) show that blast tests can cause liquefaction, inducing undrained cyclic loading, generating excess pore water pressure, and consequently producing real deformations. This paper introduces the results obtained from a research project on blast induced liquefaction, performed at a trial site located in Mirabello (near Ferrara, Italy), a little village strongly affected by liquefaction during the 2012 Emilia-Romagna earthquake. The liquefiable soils in this area contained significantly more fines than in previous blast tests. In particular, the objectives of the experiment were: (i) to evaluate liquefaction induced settlement of Apennine silts and sands, and Po River sands; (ii) to determine downdrag forces on pile foundations; (iii) to measure change in shear wave velocity/shear modulus with excess pore pressure; and (iv) to measure change in soil properties with time after liquefaction. This paper illustrates the results acquired before, during, and after the blast test using in-situ tests, pore pressure transducers, profilometers, topographical surveys, and a test micropile instrumented with strain gauges.

Centrifuge model tests were conducted to evaluate the potential for earthquake-induced soil liquefaction mitigation via microbially mediated denitrification (dissimilatory reduction of nitrogen). Desaturation by denitrifying bacteria is the first stage of a two-stage process to increase earthquake-induced liquefaction resistance referred to as microbially induced desaturation and precipitation (MIDP). In the test described herein, denitrification was induced in Ottawa F-65 sand on the 1-m radius centrifuge at the University of California, Davis NHERI/CGM centrifuge. The degree of saturation was monitored during testing by monitoring soil moisture content. After monitoring indicated that significant desaturation had been induced at 1 g by an enriched denitrifying microorganism culture, the model was accelerated to 80 g. Saturation measurements during spin-up provided insight on the influence of steady state pore pressure on biogenic gas desaturation. Cyclic loading during the ultimate centrifuge acceleration of 80 g demonstrated desaturation via MIDP has potential to mitigate earthquake-induced soil liquefaction.

A set of centrifuge model tests were conducted to evaluate the mitigation of earthquake-induced liquefaction of loose Ottawa sand treated with microbially induced calcite precipitation (MICP). The bio-mediated MICP process can facilitate the formation of cemented bonds at particle-particle contacts, which can induce large changes in soil engineering properties including small-strain stiffness, volumetric behavior, and strength. In this study, two models were constructed in a flexible shear beam container and tested using the 1-m radius centrifuge at the University of California, Davis. The liquefaction triggering behavior of a lightly cemented (calcite content≈1.2% by mass) MICP treated model is compared to an untreated model of the same initial relative density. Models were subjected to multiple sinusoidal shaking events during which accelerations, pore pressures, and surface settlements were monitored. Shear wave velocity was measured prior to each shaking event and miniature cone penetration soundings were performed at select times during testing.

The effect of soil interlayering on the measured cone penetration resistance was examined in a layered soil model tested on a 9-m radius centrifuge. The soil profile consisted of a layer of sand between overlying and underlying layers of low plasticity clayey silt. The sand layer thickness varied from 0 to 240 mm (model scale) along the length of the model. The sand was loose with a relative density of 44% on one side of the model, and dense with a relative density of 88% on the other side. The clayey silt had a plasticity index (PI) of 6 and over-consolidation ratio (OCR) of about 1.5. Multiple cone penetration soundings were performed along the width and length of the model using cone penetrometers with diameters of 4, 6, and 10 mm. The model construction procedure, data processing, and cone penetration testing results are described.

The Fraser River Delta in Metro Vancouver is one of the most populated areas of significant seismic hazard in Canada. The low-lying delta is underlain by deep unconsolidated Holocene deposits of fluvial sands, deltaic silty sand, and marine silts and clays. Seismic ground motion amplification and liquefaction are well-recognized hazards that present significant challenges for infrastructures. Spurred by the 1989 Loma Prieta earthquake, British Columbia embarked on an ambitious seismic retrofit program of older bridges in the region, including many highway bridges in the Fraser Delta underlain by liquefiable soils. Seismic upgrade of existing bridges presents special risks and challenges, particularly related to site constraints and the need to maintain the bridge in service during remedial works. This paper summarizes the key issues and challenges faced in geotechnical seismic upgrades of bridges in the Fraser Delta in the last two decades. Examples of various ground improvement methods used in several seismic retrofit case histories are described and lessons learned outlined.

In this study, a series of numerical studies were carried out to compare the seismic performance of columns that have different rigidity in liquefaction mitigation. For this purpose, experimental results of two dynamic centrifuge test models were investigated through numerical analysis using an effective stress based model UBC3D-PLM. In order to evaluate the performance and effectiveness of columns that have different rigidity on the reduction of liquefaction risk, test models improved with stiff granular columns and soil-cement columns were analyzed numerically. The results of the numerical analysis were evaluated by going through the shear stress sharing mechanism that generate between granular column, soil-cement column, and surrounding liquefiable soil. For this purpose, shear stress reduction factor (K_G) and the ratio of shear strain (γ_r) were computed and the obtained numerical results showed that high modulus columns did not reduce seismic shear stress in the surrounding soils and the shear strain ratio between soil and column was found less than 1.0 which stands in for the shear strain compatibility assumption.

The author has published specialized techniques to interpret cone penetrometer test (CPT) data for prediction of liquefaction for over 30 years, starting in 1984. These techniques have been updated numerous times, most recently in 2015. In all these cases, the technique generated contours of liquefaction resistance ratio on the CPT soil characterization chart using field and laboratory data. The problem is that the shapes of these contours are always too difficult to represent by even the most complex of equations. The author has developed software to predict liquefaction resistance using this approach, and this technique has been verified by other researchers using point-to-point data comparisons. This update reflects important observations made in the last three years. The approach indirectly accounts for all soil types and relative strength consistencies without requiring the equivalent clean sand approach.

Liquefaction susceptibility of loose sand subjected to different magnitudes of principal stress rotation under different initial stress states is evaluated. Hollow cylinder torsional shear tests were carried out on water pluviated Fraser River sand specimens hydrostatically consolidated to two different effective confining stresses. Cyclic loading was applied with constant amplitude cyclic deviator stress, but along stress paths which impose different magnitudes of principal stress rotation. Both axisymmetric and approximately plane strain loading conditions were simulated by fixing the intermediate principal stress parameter, $b_{\sigma }= \left({\sigma }_2-{\sigma }_3\right)/\left({\sigma }_1-{\sigma }_3\right)$ at 0 or 0.4 respectively. Tests results show that weakest cyclic resistance is manifested when the principal stress direction is cyclically rotated between $\pm 45°$ with respect to the axis of deposition. Irrespective of the different initial effective confining stress and intermediate principal stress parameter, the combination of high shear stress on the horizontal plane and the orientation of the plane of maximum shear stress with the bedding plane are found to be responsible for the weakest cyclic resistance.

The liquefaction potential index (LPI) was found to significantly overpredict the severity of observed liquefaction for a large subset of case histories compiled from Canterbury, New Zealand, earthquakes. One potential cause for these overpredictions is the presence of non-liquefiable capping and interbedded strata with high fines-content and/or plasticity that suppress surficial liquefaction manifestations. Herein, receiver-operating-characteristic analyses of compiled Canterbury, New Zealand, liquefaction case histories are used to investigate LPI performance as a function of the soil-behavior-type index averaged over the upper of 20 m (I_c20) of a profile; I_c20 is used to infer the amount of high fines-content, high-plasticity strata in a profile. It is shown that generally: (1) the relationship between computed LPI and the severity of surficial liquefaction manifestations is I_c20-dependent; and (2) the ability of LPI to segregate cases on the basis of observed manifestation severity using LPI decreases with increasing I_c20. In conjunction with previous studies, these findings support the need for an improved index that more adequately accounts for the mechanics of liquefaction triggering and manifestation.

The K_σ correction factor has been routinely applied in the evaluation of liquefaction resistance of soils subjected to an overburden stress level other than 100 kPa. Its current prediction form is a function of, primarily, the overburden pressure and the relative density of the soil. Hence, its application is limited to free-field level ground condition. This study focuses on the effect of sustained shear stress, which exists in the soil layer under non-free-field or sloping ground condition, on the K_σ correction factor. The findings are largely based on a comprehensive cyclic triaxial test database with the effective confining pressure and the sustained shear stress ranged up to 500 kPa and 200 kPa respectively. The test program was performed for very loose to medium dense sand. The results show that K_σ is strongly dependent on the amount of sustained shear, in which the effect is exaggerated in more contractive soil. The observations are well supported by other test data in the literature. The study sheds light on the potential need of updating the prediction form of the K_σ correction factor.

Failure of superstructure resting on shallow foundation is one of the most catastrophic phenomena occurring due to liquefaction during earthquake. The present paper presents the 3D numerical modelling of shallow foundation resting on liquefiable soil under earthquake loading. The benchmark model simulation has been simulated first to obtain the dynamic behavior of a loose sand deposit with a surface footing. The responses of this model treated with stone column improvement under the same seismic loading has been analyzed and compared with the response of benchmark models (BM), focusing on the evaluation of the strengthening effect of soil columns and its effect on the behavior of the remediated soil deposits. Acceleration base input excitation of El Centro earthquake is applied to each model to monitor the displacements, liquefaction potential, and excess pore pressures (EPP). Based on the response of the model, the relative effectiveness of stone columns as mitigation measure can be gauged. PLAXIS-3D finite element software is used for the analysis. A significant reduction in EPP and settlement are visible with the use of stone column as remedial measures.

The results of a series of nonlinear finite element simulations are presented to demonstrate the effects of base motion variability on liquefaction-induced lateral spreading of mildly sloping grounds. The analyses are part of the 2017 Liquefaction Experiments and Analysis Project (LEAP-2017). A key objective of LEAP is to generate a database of reliable centrifuge experiments to facilitate the validation of the current state-of-the-art analysis tools for the modeling of soil liquefaction. In the context of centrifuge experiments, the sources of uncertainty in the base motion are the magnitude and frequency content differences between the target and the achieved base motions. While centrifuge experiments are conducted with diligence, achieving the target base motion is restricted by the machine limitation and the presence of noise. A series of Monte Carlo (MC) simulations are performed to investigate the effects of the base motion variability. Base motion statistics are obtained as a function of frequency capturing the variations observed in the spectral accelerations of the achieved base motions. The simulations presented in this paper consider the cases of homogeneous soil as well as spatially variable soil conditions. The results of the MC simulations are used to shed light on how small deviations of the achieved base motion from the target motion might influence the response of liquefiable ground. Sensitivity of the computed soil responses to these deviations are assessed by considering the variations of excess pore pressures, lateral spreading, and ground surface settlements.

The dynamic cone penetration test (DPT) developed in China has been correlated with liquefaction resistance in gravelly soils based on field performance data from the Mw7.9 Wenchuan earthquake. The DPT consists of a 74 mm diameter cone tip driven by a 120 kg hammer with a free fall height of 100 cm. To expand the data base, DPT soundings were performed at the Pence Ranch and Larter Ranch sites where gravelly soil liquefied during the 1983 M_w6.9 Borah Peak earthquake. DPT testing was performed using an automatic hammer with the energy specified in the Chinese standard and with an SPT hammer. Comparisons suggest that standard energy corrections developed for the SPT can be used for the DPT. In general, the DPT correctly predicted liquefaction and non-liquefaction at these two test sites. Liquefaction resistance from the DPT also correlated reasonably well with that from Becker penetration testing (BPT).

The fabric of sandy soil samples retrieved by conventional “so-called” undisturbed sampling techniques is prone to disturbance, which may lead to test results that over or underestimate liquefaction resistance. Although there is an option to use a reconstituted sample as a substitute for the undisturbed sample, it is difficult to recreate the in situ soil fabric for the reconstituted sample. In order to investigate the effect of soil fabric on the liquefaction resistance of Toyoura sand, a series of undrained cyclic triaxial tests and shear wave velocity measurements were performed, and then the results and those in previous literature were analyzed. Finally, a method for evaluating liquefaction resistance based on the in-situ and laboratory-measured shear wave velocities is proposed.

The influence of content and plasticity of fines in sands on field liquefaction assessment remains still somewhat unclear in geotechnical practice. In field liquefaction assessment, corrected standard penetration number (N₁)₆₀ or cone penetration resistance (q_c) were used in general to evaluate the liquefaction resistance of sands with fines. (N₁)₆₀ or q_c are generally related to the relative density of sands. In this study, stress-controlled cyclic simple shear tests were performed on sands with up to 10% plastic and non-plastic fines at different relative densities. The results demonstrated that at the same relative densities, sands with fines (up to 10 % content) have lower liquefaction resistance than clean sands. Furthermore, the effect of plasticity and the content of fines diminishes at loose states of sand specimens with low fines content. In sands with plastic clays, the relative density has less effect on liquefaction resistance when the content of fines increases from 5% to 10% when compared with sands with non-plastic silt. To conclude, further research is needed to correlate these experimental findings to the field liquefaction assessment of sands with fines.

The paper discusses the effect of seismic preshaking history on liquefaction of silty sand soils using case histories in California, as well as experimental data. The case histories are: (a) the response of the Wildlife site in the Imperial Valley to the 2010 El-Mayor Cucapah earthquake (M_w = 7.2, a_max = 0.15g); and (b) the response of the Treasure Island Fire Station (F.S.) site in the San Francisco Bay area to the 1989 Loma Prieta earthquake (M_w = 6.9, a_max = 0.16g). Both sites are similar in almost every respect except for their seismic history. The magnitude and intensity of the corresponding earthquakes were also very similar at both locations. While Treasure Island F.S. did liquefy during the shaking, Wildlife did not and was far from liquefaction as indicated by piezometers at the site. The experiments conducted in this research were a crude simulation of the seismic history of the sites. From the results of the experimental simulation as well as the field case histories, it is concluded that preshaking by previous earthquakes is the most probable explanation of the higher liquefaction resistance exhibited by the Wildlife site and other sites in the Imperial Valley of Southern California.

The moment magnitude (M_w) 7.8 Kaikōura, New Zealand, earthquake triggered relatively few cases of liquefaction and related phenomena (e.g., lateral spreading) despite the large magnitude of the event. Cases of severe liquefaction manifestation were confined to localized areas proximal to waterways near the township of Blenheim, in the north-eastern corner of the South Island of New Zealand. The occurrence and non-occurrence of liquefaction within the wider Blenheim area is shown to closely correspond with fluvial geomorphology and associated depositional setting of the sediments. Herein, the distribution of liquefaction within the region is detailed in the context of the geomorphological influences and sedimentologic controls. This work highlights the influence of geomorphic variability on the occurrence of liquefaction, with the aim of improving the assessment of liquefaction hazards for future events worldwide.

The M_w 7.8 Muisne-Pedernales, Ecuador, earthquake on April 16, 2016, produced significant damage through the northwest coast of Ecuador. Important state infrastructure was seriously affected, including one of the major seaports in the country, the Port of Manta, located in the city that bears its name. This paper focuses on the geotechnical exploration, site characterization, and analyses performed for the marginal wharf's embankment of the port, where clear evidence of liquefaction was observed. Laboratory results showed that the embankment was composed of gravelly sand and sandy gravels, which limited the applicability of common exploration techniques, such as SPT and CPTu tests. Subsequently, liquefaction triggering evaluations were performed based on V_s and the Chinese dynamic penetration test (DPT), which can account for the large gravel content in the soils. The results of this evaluation are presented and compared with field observations showing that these methods correctly identified the potential for liquefaction in the embankment.

The 2016 Muisne, Ecuador, earthquake caused widespread soil liquefaction and structural damage as affected sites were subject to peak ground accelerations of greater than 0.4 g. Liquefaction was observed in the silty sand foundation soils adjacent to the recently constructed Briceño Bridge embankment site as evidenced by sand boils and liquefaction ejecta. However, the nearly 1 km long Briceño embankment, whose foundation materials were reinforced by relatively widely spaced rammed aggregate pier ground improvement elements, exhibited only minor damage after shaking. Post-earthquake data indicated that the ground improvement works provided densification of the SM foundation soils. This densification was not sufficient, however, to fully explain the stability of the embankment, an achievement that necessitates the presence of other stabilization mechanisms. This paper is of particular significance because it presents proposed stabilization methods that reach beyond densification in difficult to densify soils such as silty sands and sandy silts.

Microbial induced calcium carbonate precipitation (MICP) has been shown to be an effective method to improve soil strength and stiffness, and one of the most promising applications of MICP is for liquefaction mitigation. In this paper, three silica soils with varying grain sizes were treated by MICP to a moderate level of cementation and their resistance to liquefaction is improved significantly compared to uncemented loose soil. The degree of MICP improvement appears to be influenced by the grain size and shape of the loose untreated soil. The level of excess pore water and cyclic resistance of the untreated and MICP treated soils is presented. The different response to cyclic loading after MICP treatment is compared with respect to different grain size and shape. The results show that reduction in grain size and increase in angularity result in a higher resistance to liquefaction after MICP treatment.

Liquefaction phenomenon observed in the Canterbury earthquake sequence, peaking during the Christchurch earthquake in February 2011, has raised questions over the effect of previous earthquakes on liquefaction susceptibility of soil. Typically, the increase of relative density due to post-liquefaction reconsolidation has been reasonably considered to increase liquefaction resistance. However, some recent laboratory experiments have provided new evidence that large shear strains during past liquefaction events may have adverse effects on the liquefaction susceptibility of sand. This study aims to investigate the effect of previous liquefaction occurrences on the current liquefaction susceptibility of granular materials under cyclic loads, through both centrifuge shaking table experiments on sand and 2D DEM (discrete element method) numerical simulations on circular particles. During both the physical experiments and the numerical simulations, the granular material repeatedly undergoes “cyclic loading and reconsolidation” cycles to examine the change in liquefaction susceptibility due to previous liquefaction history. Shear strain generated during previous liquefaction history is shown to have an opposite effect on liquefaction susceptibility compared to the densification due to reconsolidation, and could increase liquefaction susceptibility.

The selection of the appropriate input motion plays a crucial role in advanced numerical simulations of cyclic liquefaction. In the present paper, a numerical model validated against the evidence of soil liquefaction phenomena observed during the Emilia 2012 earthquake in Italy is used to assess the influence of the ground motion selection process. The numerical model has been developed using the finite element code OpenSees and adopting the pressure dependent elasto-plastic multi-yield constitutive model for the liquefiable layers. The model has been first calibrated by using laboratory and in situ tests from the San Carlo site. Then a large set of strong motion records has been used and results have been classified based on liquefaction occurrence. The most influent parameters for design purposes have been identified through a cluster analysis. The results underline that the selection criterion should be based on the estimation of expected ground motion parameters rather than on a pure spectral matching criterion.

This paper investigates the key parameters that influenced the settlement of a case study building on liquefiable soil in Adapazari (Turkey) during the 1999 Kocaeli earthquake. Ground movements in Adapazari caused large devastation, largely attributed to liquefaction of low plasticity silty soil layers underneath buildings on shallow foundations. The case study soil profile was well characterized by in-situ testing as well as laboratory tests from the Adapazari area. This allowed several different estimates of the building settlement to be obtained through different methods and through a variation in upper and lower bound estimates of the soil parameters. The different methods and different soil properties resulted in a wide range of estimates from 0.004 m to 1.6 m for the building settlement, compared to the observed in-situ value of 0.9 m. Even though the results were varied, the estimation of the liquefied strength of the soil appeared to be a key parameter for the settlement of the case study building. A detailed study with the PLAXIS finite-element software and UBC3D-PLM constitutive model, provided a consistent estimate of the final settlement of 0.9 m compared to the in-situ value. However, the limitation due to the enforced ‘undrained’ conditions during the dynamic phase of the analyses may have resulted in an inaccurate simulation of the pore water pressure and subsequently could have influenced the estimation of settlement. The modeling of the liquefaction settlements under free-field conditions was also considerably less than the re-consolidation settlements that were obtained through simplified procedures, suggesting that the re-consolidation settlement under the foundation was not modelled accurately. The present paper focuses on the assessment of the settlements due to earthquake-induced liquefaction as part of the research being conducted within the European project LIQUEFACT.

Liquefaction-induced lateral spreading during the 2011 Christchurch earthquake in New Zealand was severe and extensive, and data regarding the displacements associated with the lateral spreading provides an excellent opportunity to better understand the factors that influence these movements. Horizontal displacements measured from optical satellite imagery and subsurface data from the New Zealand Geotechnical Database (NZGD) were used to investigate two distinct lateral spread areas along the Avon River in Christchurch. These areas experienced displacements between 0.5 and 2 m, with the inland extent of displacement ranging from 100 m to over 600 m. Existing semi-empirical displacement models tended to under estimate the displacements at one site, but better predicted displacements at the other. The integrated datasets indicate that the areas with more severe and spatially extensive displacements are associated with thicker and more laterally continuous deposits of liquefiable soil. In some areas, the inland extent of displacements is constrained by geologic boundaries and geomorphic features, as expressed by distinct topographic breaks. In other areas the extent of displacement is influenced by the continuity of liquefiable strata or by the presence of layers that may act as vertical seepage barriers. These observations demonstrate the need to integrate geologic/geomorphic analyses with geotechnical analyses when assessing the potential for lateral spreading movements.

Semi-empirical models based on in-situ geotechnical tests are the standard-of-practice for predicting soil liquefaction. Since the inception of the “simplified” cyclic-stress model in 1971, variants based on various in-situ tests have been developed, including the cone penetration test (CPT). More recently, prediction models based solely on remotely-sensed geospatial information were proposed. These “geospatial” models aim to rapidly predict liquefaction using data that is free and readily available. This includes: (1) common ground-motion intensity measures, which can either be provided in near-real-time following an earthquake, or predicted for a future event; and (2) geospatial parameters (e.g., among many, ground slope; distance to water bodies; and compound-topographic-index), which collectively are used to infer characteristics of the subsurface without in-situ testing. However, the predictive capabilities of geotechnical and geospatial models have not been rigorously compared, which could elucidate techniques for model improvement, and which would inform modeling approaches that statistically coalesce predictions from multiple models. Accordingly, this study assesses the relative efficacy of liquefaction prediction-models based on geospatial vs. CPT data using 9,731 case-studies compiled from recent earthquakes in New Zealand. While the top-performing models are CPT-based, the geospatial models perform surprisingly well considering the relative time- and cost-requirements underlying the predictions. Although further research is warranted, the findings of this study suggest that geospatial models have the potential to provide valuable predictions of liquefaction occurrence and consequence. Towards this end, several thrusts of ongoing investigation are outlined.

In the 2016 M_w7.8 Kaikoura (New Zealand) earthquake, widespread liquefaction occurred in the reclaimed land of CentrePort, Wellington, in end-tipped gravelly fills and hydraulically-placed sandy fills. The liquefaction resulted in global (mass) settlement of the reclamations and lateral spreading, which was particularly pronounced along the edges of the reclamation. The large settlements and spreading-induced movements affected container terminals, wharves, and buildings constructed on deep and shallow foundations. There are several interesting aspects of this important case history including the performance of reclamations of different soils (gravelly soils and sandy soils), reclamations of different age (from 40 years to over 100 years) and methods of construction (end-tipping and hydraulic placement), wharves subjected to spreading of liquefied fills, and buildings on deep and shallow foundations affected by liquefaction-induced settlement or/and lateral spreading. The paper combines detailed observations from damage inspections, results from comprehensive CPT investigations, respective analyses, and interpretation to emphasize key aspects of this important case study on the liquefaction of gravelly soils.

Liquefaction probability curves for three Quaternary sand deposits near Charleston, South Carolina, are derived in this paper. The derivation is based on seismic cone penetration test profiles for 33 sites in the Holocene beach to barrier-island deposits (Qhs); 24 sites in the 33,000- to 85,000-year-old Silver Bluff terrace (Qsbs); and 91 sites in the 70,000- to 130,000-year-old Wando Formation (Qws). Liquefaction potential is expressed in terms of the liquefaction potential index (LPI). The effect of aging processes on liquefaction potential is considered through a correction factor based on the ratio of measured shear-wave velocity (V_S) to estimated V_S. The liquefaction probability curves are expressed as functions of peak ground acceleration, magnitude, and probability that LPI is greater than or equal to a threshold value for surface manifestation of liquefaction at level ground sites. The results indicate that Qhs exhibits the highest probability for a given level of ground shaking, followed by Qsbs and then Qws.

One of the most valuable liquefaction case histories is that of the Wildlife Liquefaction Array (WLA) in the November 23–24, 1987 Elmore Ranch and Superstition Hills earthquakes. The WLA site contained a vertical array with accelerometers at the ground surface and at 7.5 m depth, and six piezometers. Five of the piezometers were in a layer of loose, silty sand that extended from about 2.5 to 6.5 m depth and the sixth was in a layer of silt at 12 m depth. The groundwater table was approximately 1.5 m deep. On November 23, 1987, the Elmore Ranch (M6.2) earthquake produced moderate shaking but no surficial evidence of liquefaction at the WLA site. The next day, the Superstition Hills (M6.6) earthquake produced surficial evidence of liquefaction in the area of the strong motion instruments. Recorded pore pressure measurements, however, were considered by some to be inconsistent with the recorded ground motion measurements leading to published discussions that described significant technical disagreements in the geotechnical earthquake engineering literature. Recently developed time-frequency analyses and nonlinear effective stress site response analyses can offer new insight into the WLA response during the Superstition Hills event. This paper reviews the conditions and measurements at the WLA site, and the arguments that support and criticize the accuracy of the pore pressure measurements. It then presents time-frequency and site response calculations that clarify the response at the WLA array, and illustrate how such calculations can be used to better interpret liquefaction at vertical array sites.

Lateral spreads caused by liquefaction of underlying soils present a significant hazard to both overlying infrastructure such as roads, bridges, and buildings as well as subsurface infrastructure including pipelines, utilities, and tunnels. Assessing lateral spread displacements and spread extents using traditional field methods (i.e. manual measurement of cracking associated with lateral spreads) is time consuming, labor intensive, and often ambiguous. Remote sensing techniques such as optical image correlation (OIC) can provide a way to quickly measure and identify lateral spread displacements and spread extents at much lower cost. In this paper, OIC is applied to lateral spread areas in Kaiapoi, New Zealand, associated with the 2010 Darfield earthquake. Results are compared with ground survey measurements showing good agreement with these quantitative measurements. Specifically, the maximum displacement calculated using OIC matches the ground survey measurements within 0.4 m, and the failure mode and extent of the lateral spreads inferred from OIC are consistent with those obtained in the ground survey.

The waterfront Neptune Marina Development is located in Marina Del Rey Harbor, Los Angeles County, California. Four, 4-story wood-framed apartment buildings, covering a footprint of approximately 10 acres, are currently being constructed over a 2-level concrete subterranean podium parking structure, adjacent to the waterfront harbor. The current site soils are subject to liquefaction under the design earthquake. Should the site soils liquefy, there is the potential for structural damage due to foundation bearing failure and excessive settlement. In addition, the presence of continuous liquefiable soil layers poses lateral spreading hazard under a design earthquake event. The areas susceptible to lateral spreading are located toward the marina seawall. Working closely with the project team at the early conceptual stages of the project, Hayward Baker, Inc. (HBI), designed and built a comprehensive ground improvement program to mitigate the sites liquefaction and lateral spreading hazards, to provide adequate soil bearing capacity and settlement control under footings, as well as the temporary excavation supports. This paper will provide the detailed site investigation, liquefaction and lateral spreading analysis, the ground improvement design by vibro stone columns and soil mixing, the construction procedures, and QA/QC program with the lab and the field verifications tests.

This paper presents the organizational structure (schema) of the next-generation liquefaction (NGL) relational database. The schema describes the tables, fields, and relationships among the tables, and provides an important resource for users who wish to interact with the database by writing queries. Structured relational databases are not commonly utilized in the natural hazards community, where file repositories are more commonly used and often called “databases”. This paper also discusses what a relational database is, and why this approach was adopted for the NGL project.

Performance-based earthquake engineering analysis methods are desirable in geotechnical engineering due to their ability to incorporate the various uncertainties associated with complex problems and their ability to frame responses in terms of a targeted hazard. However, such methods are difficult to implement on routine projects without the use of specialized software because the many iterative probabilistic calculations required. Prior research has introduced a simplified approach of using performance-based reference parameter maps in conjunction with correction equations that account for site-specific geotechnical and topographical conditions. This paper addresses the development of such performance-based reference parameter maps for the calculation of post-liquefaction volumetric strains and settlements in the free field. A study is performed and presented to assess the optimized grid spacing for a performance-based volumetric strain analysis and development of a strain performance-based reference parameter map. Recommended grid spacing values as correlated to USGS seismic hazard maps for PGA are provided to produce a maximum of 0.15% strain error in interpolating the reference strain for the simplified performance-based analysis approach.

Estimation of post-liquefaction settlement is critical to robust design within a performance-based earthquake engineering framework. However, most research in soil liquefaction has focused on triggering of liquefaction rather than post-liquefaction response. In addition, the liquefaction and post-liquefaction response of gravelly soils is not as well understood as sandy soils because of the challenges of testing large particle size materials in the laboratory. In this study, a prototype large-size cyclic simple shear (CSS) device was utilized to evaluate the post-liquefaction response of gravel-sand mixtures. Test materials included 9 mm subrounded pea gravel, 8 mm angular crushed limestone (CLS8), subrounded Ottawa C109 sand, and mixtures of gravel and sand at varying percentages. Constant volume cyclic simple shear tests were performed to assess post-liquefaction volumetric strain (settlement). The effects of particle angularity, relative density, and mixture percentage on post-liquefaction volumetric strain were investigated. The results show that post-liquefaction volumetric strain is dependent on particle angularity. As particle angularity increased, the post-liquefaction volumetric strain decreased for the uniform materials (Ottawa C109 sand, pea gravel, and CLS8). The subrounded pea gravel and subrounded Ottawa C109 sand had similar post-liquefaction volumetric strains, indicating that uniform gravel and uniform sand can have similar values of post-liquefaction volumetric strain if their particle morphology is similar.

The most popular procedure for evaluating liquefaction potential is heavily based on peak ground acceleration (PGA). Not only is it an inadequate representation of such a complex phenomenon, but also limited to a binary classification (liquefaction or no liquefaction) without providing associated information on the damage caused by liquefaction. This paper presents an experimental study on identifying an optimum parameter that can potentially predict lateral spreading during a liquefaction event. To identify this optimal parameter, the missing dimension of time is brought into consideration with the use of evolutionary ground motion intensity measures (IMs). It is essential to identify the time that liquefaction is initiated within a given ground motion to accurately predict the lateral displacement because the damage predominantly occurs after liquefaction initiation. The experimental program using cyclic simple shear tests evaluate the responses of sands subjected to undrained transient loadings with a wide range of spectral and temporal characteristics. Out of the four IMs considered in this study, the post-initiation cumulative absolute velocity (CAV₅) is identified as the most efficient IM for predicting liquefaction induced displacements.

Soil liquefaction was one of the major causes of damage in the 1886 Charleston earthquake, one of the largest historic seismic events in the eastern United States. Paleoliquefaction studies indicate that at least seven, large, prehistoric earthquakes have occurred within the last 6000 years in the South Carolina Coastal Plain (SCCP). The U.S. Geological Survey (USGS) has marked the SCCP as a region with high seismic hazard. In this study, data from cone penetration tests obtained from six test locations at two sites with a history of paleoliquefaction, Sampit and Gapway, were used for site-specific evaluation of liquefaction potential. Liquefaction evaluation methods, that include both deterministic and probabilistic approaches, were used to evaluate the current liquefaction susceptibility based on the most recent estimation of prehistoric earthquake magnitude and peak ground accelerations. Results indicate that, at these two sites, the factor of safety against liquefaction ranges from 0.4 to 0.7 and liquefaction probability is greater than 83%. Moreover, results obtained from probabilistic approaches were compared with deterministic approaches and it was found that except for one test location, the equivalent minimum probability of liquefaction for the soil deposits was about 16%.

This paper addresses the quantitative evaluation of ageing effects on liquefaction resistance of sand on the basis of experiences during recent earthquakes. Because of the practical needs of the people, emphasis is placed on the ageing in the recent 1000 years that have not been studied in details. Many uncertainties in case histories led to the idea of upper and lower bounds of ageing that successfully indicated the progress of ageing. The interest was extended to archaeological aspects and the site of an ancient man-made island, which had been missed for centuries, was identified. One of the very important causes of ageing is identified to be dislocation of sand grains and the minor cyclic shear during earthquakes promotes ageing. Therefore, seismically active countries enjoy increase of liquefaction resistance more rapidly than inactive countries.

A sequence of earthquakes causing surface manifestations of liquefaction struck the Canterbury region in New Zealand in 2010–2011. Satellite and aerial imagery is readily available after an earthquake and can be used efficiently and quickly to map liquefaction occurrence with the use of automated classification techniques. Automated techniques applied to satellite multispectral imagery could allow for a minimal compromise between time and accuracy and result in a reliable map of liquefaction surface affects for an area. In this study, multispectral satellite imagery after the February 22, 2011, Christchurch earthquake has been used to identify liquefaction occurrence following the event. The image has multi-spectral bands with approximately 2 m spatial resolution in addition to the panchromatic band with approximately 50 cm spatial resolution. An aerial image taken after the event with 10 cm spatial resolution has been used to establish ground truth. In this study, the effectiveness of supervised and unsupervised classification methods have been evaluated in classifying different types of liquefied materials on the ground level. Liquefied materials have been classified into different categories such as wet and dry sands and a confusion matrix has been generated to evaluate the method accuracy. The K-means method is used for unsupervised classification and the minimum distance method and artificial neural network (ANN) algorithm have been used for supervised classification. ANN outperformed the other classification methods; however, all methods had difficulty in discriminating the “liquefied dry materials” and “liquefied wet materials.”

This study uses an earthquake energy attenuation model to reevaluate a SPT-based liquefaction case history database collected by previous researchers. The necessary information such as mechanism of faulting, average shear wave velocity of the top 30 m of the sites, and site-to-source distance were added to the database. The earthquake demand energy (EDE) of liquefied and non-liquefied sites were calculated and then plotted versus corresponding SPT-N values. Based on the results, a SPT-based liquefaction triggering curve was suggested that can be used in the practice. The mathematical expression of the curve makes the possibility of using SPT-N value as an alternative to the soil capacity energy, which its determination is associated with some complications and difficulties for engineers. Finally, the proposed liquefaction triggering curves are compared with the previous correlations, which are based on mainly the same case histories, and the slight differences between them are discussed.

Seismic performance of earthen dams subject to potential liquefaction hazard is of concern in both seismic risk assessment and design mitigation for dams. Seismic deformation analyses are increasingly utilized in evaluating and mitigating these hazards. It is important to assess different types of analytical approaches by means of their application to well-documented field case histories. The performance of the Lower San Fernando Dam (LSFD) during the 1971 San Fernando earthquake is a well-documented seismic performance field case history involving a liquefaction-induced flow slide failure of the upstream side of the embankment. The dam was in close proximity to the fault rupture during the M_W≈6.7 event, and it was strongly shaken. The dam had 11.1 m (36 ft) of freeboard prior to the earthquake, and the nearly catastrophic upstream slide reduced freeboard to within approximately 1.5 to 1.8 m (5 to 6 ft) from overtopping. This unusually well-documented field performance case history has been foundational to the development of the field of modern seismic dam engineering, and to the inception of the U.S. national seismic dam safety programs ongoing today. It is important to periodically assess the strengths and reliability of analytical models and approaches employed in engineering analyses of these types of critical facilities. The LSFD case history was re-analyzed in these current studies by means of a series of fully nonlinear seismic deformation analyses, employing (1) a suite of four different analytical models, (2) a suite of three cyclic pore pressure generation models, and (3) a suite of three post-liquefaction residual strength relationships.

Seismic performance of earthen dams subject to potential liquefaction hazard is of concern in both seismic risk assessment and design mitigation for dams. Seismic deformation analyses are increasingly utilized in evaluating and mitigating these hazards. It is important to assess different types of analytical approaches employed by means of their application to well-documented field case histories. The performance of the Upper San Fernando Dam during the 1971 San Fernando earthquake is a well-documented seismic performance field case history involving moderate deformations; a range of deformations of significant interest for both risk assessment and for engineered mitigation of such risks. The dam was in very close proximity to the fault rupture during the M_W≈6.7 event, and it was strongly shaken. It sustained maximum horizontal displacements of approximately 2.7 m (9 feet) at the downstream toe, and 0.9 m (3 feet) of vertical crest loss. This case history was re-analyzed in these current studies by means of fully nonlinear seismic deformation analyses, using (1) a suite of four different analytical models, (2) a suite of three cyclic pore pressure generation models, and (3) a suite of three post-liquefaction residual strength relationships to assess the accuracy and reliability of these types of analyses.

A simplified methodology is proposed for the computation of elastic response spectra at liquefiable sites, based on the widely used equivalent linear site response analysis method. The target response spectrum is obtained by frequency dependent interpolation between the elastic response spectra for “non-liquefied” and “liquefied” ground, the former computed for the native soil properties, without excess pore pressure buildup, and the latter for the liquefied soil properties. Apart from the structural frequency, the interpolation coefficient is also related to the factor of safety against liquefaction. In this way, the proposed methodology indirectly takes into account the initial segment of the seismic excitation, until the onset of liquefaction, which is of cornerstone importance for the evaluation of the overall ground response at sites with low and moderate liquefaction potential. Comparisons with results of nonlinear numerical analyses and field case studies verify the potential of the proposed methodology for preliminary design purposes.

Natural gravelly soils liquefied from time to time during previous earthquakes were all well-graded and of much higher dry density than poorly-graded sands, though their S-wave velocity V_s and SPT N-value were as low as liquefiable sands. A series of systematic pressure chamber tests conducted on a set of granular soils indicate that N and V_s of well-graded gravels are correlated with soil densities quite differently from those of sands and highly dependent on grain-size curves or uniformity coefficient C_u. In laboratory triaxial tests, CRR-values for well-graded gravelly soils corresponding to the initial liquefaction, for axial strain ε_DA=5%, are not much different from those for poorly-graded sands of the same relative density. On the other hand, post-liquefaction residual strength for larger strains, ε=20% for example, is obtained about 10 times larger than poorly-graded sands indicating that gravelly soils are much more resistant to post-liquefaction large-strain failures. CRR~N correlations for gravelly soils developed by triaxial tests on intact samples and liquefaction case-histories are discussed in view of the basic laboratory test results.

Previous studies have shown that ureolysis driven calcite precipitate improves the behavior of granular soil under unidirectional principal loads. However, the behavior of these improved soils can be very different under dynamic loading conditions. This study aims to understand the behavior of bio-cemented sand under cyclic simple shear tests. Calcite was precipitated by inducing urease enzyme rich bacteria in clean fine sands (particle size<425 μm). Improvement in the cyclic resistance of sand was assessed through cyclic-direct simple shear (CDSS) tests. Samples were prepared by dry pluviation method in a stainless steel column of 71 mm diameter and 20 mm height, with an opening on the top and bottom cover. Bacterial suspension was injected in to the saturated soil sample followed by cementation solution containing equimolar mixture of urea and calcium chloride solutions. Load controlled cyclic tests were conducted on the above treated samples. The samples were then tested under constant volume, which helped co-relate the vertical stress change to its excess pore water pressure. The sample failed when shear strains exceeded 3.75 % or when the pore pressure ratio reached unity. Shear strains governed the failure in most samples. The test results showed that the cyclic resistance increased and the pore pressure ratio decreased with the increase in precipitation.

For many dams, the behavior of cohesionless and cohesive soils under cyclic loading plays a key role in their seismic stability. Such is the case of Gatun Dam, which retains Gatun Lake, the main waterway of the Panama Canal. The dam was built by hydraulic filling between 1907 and 1914 and is founded on deep, geologically young deposits of granular and fine-grained soils. This paper presents a case history on the seismic evaluation of the dam. The study involved recently developed procedures for evaluating the susceptibility and potential for liquefaction of soils, for characterizing the residual strength of liquefied soils, and for evaluating the potential for cyclic softening of clayey soils. The paper describes the dam and foundation characterization; the analyses of liquefaction potential, cyclic softening, and seismic deformations; and the assessment of expected dam performance. Key issues in the implementation of the analysis procedures and considerations involved in assessing the seismic stability of the dam are also discussed.

Liquefaction-induced settlement of buildings on shallow foundations requires evaluation of shear-induced, volumetric-induced, and ejecta-induced ground settlement mechanisms. There are few procedures for estimating the shear-induced component of building settlement due to liquefaction. Nonlinear dynamic soil-structure interaction (SSI) effective stress analyses can capture this mechanism. A simplified procedure was developed that provides building settlement estimates consistent with those calculated from the dynamic SSI effective stress analyses of a variety of buildings for generalized subsurface conditions. The liquefaction-induced building settlement (LBS) index is employed to characterize more complicated ground conditions. The proposed simplified procedure is applied to several field case histories and shown to provide estimates of liquefaction-induced building settlement consistent with those observed. Recommendations for evaluating building performance at liquefiable sites are shared.

The seismic failure of an old quay wall, built in 1932, located in the Port of Iquique during the Mw 8.2 April 1, 2014, Pisagua, Chile, earthquake is studied. A geotechnical exploration, including boreholes, CPT soundings, and surface-based geophysical techniques, was performed to characterize the materials forming the wharf. The data and field observations are then analyzed and used to develop a geotechnical model of two cross-sections modeled in Plaxis2D, to conclude which is the most likely failure mechanism under a dynamic analysis. The results are compared against the deformations measured in a number cross sections of the wharf, suggesting that the failure can be explained by a combined effect of the dynamic increase in lateral earth pressures and the base rotation of the confining walls due to the lack of proper compaction of the rockfill prism below the confining walls.

A novel approach to screen for soil liquefaction potential utilizes the stress history profile obtained from cone penetration tests (CPT) to differentiate contractive versus dilative behavior within the context of critical state soil mechanics (CSSM). Contractive soils are liquefaction prone, whereas dilative soils are not. Specifically, the profiles of apparent yield stress (σ_p') and yield stress ratio (YSR) with depth are determined from the net cone resistance (q_net) and CPT material index (I_c) from an established relationship for sands, silts, and clays. When the in situ YSR falls below a threshold value established by CSSM (i.e., YSR<3), the soil is contractive, whereas otherwise it is dilative. Both static flow and cyclic liquefaction can be assessed in this manner. Four case studies showing the YSR triggering threshold are shown in comparison with standard procedures.

Effective-stress nonlinear dynamic analyses (NDA) were performed for a large-diameter reinforced concrete (RC) pile in multi-layered liquefiable sloped ground. The objective was to assess the effects of earthquake duration on the combination of inertia and liquefaction-induced lateral spreading. A parametric study was performed using input motions from subduction and crustal earthquakes covering a wide range of motion durations. The NDA results showed that the pile head displacements increased under liquefied conditions, compared to nonliquefied conditions, due to liquefaction-induced lateral spreading. The NDA results were used to develop a displacement-based equivalent static analysis (ESA) method that combines inertial and lateral spreading loads for estimating elastic and inelastic pile demands.

Inverse analyses were used to evaluate the degree of partial drainage occurring during dynamic shaking of liquefying soil profiles in a set of centrifuge model tests. Three tests were performed using the 9-m radius centrifuge at the UC Davis Center for Geotechnical Modeling on saturated Ottawa sand models with initial relative densities of 25, 43, and 80%. Models were subjected to multiple sinusoidal shaking events with acceleration amplitudes ranging from 0.03 to 0.55 g. Densely spaced pore pressure transducer arrays provided profiles of pore pressure generation and dissipation; inverse analyses of the pore pressure data were used to obtain volumetric strain profiles during shaking and dissipation. Surface settlements computed by integrating the volumetric strain profiles are compared to surface settlements measured from linear potentiometers. The magnitude of the volumetric strains due to partial drainage and their potential effects on liquefaction responses are discussed.