Geo-Congress 2023: Geotechnics of Natural Hazards

This paper presents the results of liquefaction hazard analyses for four slopes along a deep-water canyon. The slopes are composed of mainly clay-like soil; however, the top 2–7 m of three of the slopes contains clayey-sand layers that were found to be susceptible to liquefaction. As a result, liquefaction triggering and consequences were evaluated by performing (1) three different CPT-based liquefaction triggering analyses; (2) 1D nonlinear effective stress analyses with three different pore pressure generation models; and (3) 2D nonlinear effective stress analyses using an advanced constitutive model. The results highlight the benefit of performing advanced analyses calibrated with site specific data. Simplified CPT triggering analyses were overly conservative, while 1D effective stress analyses calibrated with index parameters predicted no liquefaction compared to the 1D and 2D analyses that were calibrated with stress controlled cyclic DSS tests.

Microbially induced carbonate precipitation (MICP) is a bio-mediated ground improvement technique that can increase the liquefaction resistance of loose granular deposits. Previous researchers have acknowledged the possible amplification of ground motions due to the improvement in soil stiffness from MICP-treatment. To further elucidate the response of MICP-treated soils to cyclic loading and its impact on ground motion characteristics, an understanding of the shear modulus reduction and energy dissipation as a function of shear strain is required. Modulus reduction and damping (MRD) curves are a key input for site response analysis, which evaluates the influence of a local soil deposit on the characteristics of seismic waves. We investigate the influence of the coefficient of uniformity (C_u) on MRD curves of sands treated to a moderate level of cementation. The dynamic properties of the untreated and MICP-treated sands were evaluated through resonant column testing. Results indicate that an increase in C_u leads to greater reduction of the MICP-treated normalized shear modulus curve compared to that of the untreated soil. Additionally, the minimum calcium carbonate content and MICP uniformity was observed to influence the reduction behavior shown in MRD curves obtained from resonant column testing. These preliminary findings are important for seismic analysis and design because the C_u of MICP-treated soils and the MICP-treatment uniformity influence dynamic soil properties and will likely affect the response of MICP-treated soils to cyclic loading; however, further research is needed to fully understand the observed behavior.

The vibrations induced in the ground due to the propagation of high amplitude pressure waves generated during blasts pose a serious danger to structures on the ground and underground. In this paper, we explore the attenuation behaviour of sand subjected to blast waves generated in a vertical shock tube driven by compressed helium gas. Peak pressure, decay time, and impulse of the blast wave profile have been used to compare different blast profiles generated in the laboratory. Pressure sensors are embedded at a number of locations in the test section to note the pressure profiles in the air and sand column. In order to study the effect of compaction on attenuation, sand compacted at different relative densities is subjected to blast waves of various overpressures, and the attenuation factor is estimated. This study shows that shock tubes can be effectively used to understand the behaviour of attenuation characteristics of sand under blast loads.

Lateral spread case histories in the NGL database differ in a key manner from the majority of liquefaction case histories compiled by NGL; lateral variability is critically important for lateral spreading. This project examines the Heber Road case history that has primary CPT subsurface data and uses geostatistical tools to study this case history in terms of layer continuity and overall variability for each case history. Different valid realizations of the subsurface are compared to conventional lateral spread prediction methods and measured lateral displacements. No lateral displacement sections are also studied for the site. This work shows (1) that not all viable geostatistical realizations will result in conditions for lateral spread despite the measured occurrence and (2) that geostatistical tools can be essential tools for helping understand if continuity conditions exist at a site for lateral spreading conditions to be present if sufficient spatially distributed data of acceptable quality exist.

This study presents an investigation of the cyclic and post-cyclic behavior of a sensitive saprolite soil. Saprolite soils are formed by the chemical decomposition of rock to soil particles that remain in place. Sensitive saprolites have an open boxwork structure that when worked by means of excavation, placement, or compaction lose this structure and become very wet. The authors were not able to find studies discussing cyclic and post-cyclic behavior of these unusual soils. The soils in this study classify as high-plastic silt (MH) based on their USCS classification with a plasticity index ranging between 18 and 36, a fines content between 52% and 97%, and a saturated water content of between 50% and 65%. Laboratory tests include stress-controlled cyclic direct simple shear (DSS), post-cyclic reconsolidation, static and post-cyclic monotonic DSS, and consolidation tests. The characterized stress history shows these soils have an overconsolidation ratio (OCR) between 1.5 and 2.0. The results of this testing program indicated that the cyclic resistance ratio (defined as shear stress normalized by initial vertical effective stress) to reach 3% single amplitude shear strain in 15 cycles was 0.29. Post-cyclic shear tests showed a strength reduction between 12% and 22% interpreted at 10% shear strain. Post-cyclic reconsolidation tests showed volumetric strains between 1.0% and 1.5% for specimens with residual excess pore pressures between 70% and 85%. The results of the laboratory cyclic tests in this study are compared to other fined-grained soils reported in literature.

When modeling seismic site effects, many studies have focused on the importance of vertically characterizing the shear wave velocity (Vs) profile. However, the importance of laterally characterizing the Vs variability has generally been either overlooked or oversimplified. As such, little is known about the spatial area that influences seismic site effects, and the state-of-the-practice continues to quantify these effects by combining results from multiple one-dimensional (1D) analyses, such as Vs randomization. In this study, we investigate the spatial area influencing site response at the Treasure Island Downhole Array (TIDA) by (1) performing numerical two-dimensional (2D) ground response analyses (GRAs) and (2) combining multiple 1D GRAs using all 1D Vs profiles sampled from the 2D models. We use a site-specific, large-scale, Vs model that has been developed to a depth of 150 m over the entirety of Treasure Island. By comparing 2D and combined 1D GRA predictions from different lateral extents to recorded earthquake ground motions at TIDA, we systematically highlight the large spatial area that influences site response, which extends to distances as much as 1 km from the TIDA sensors.

This article studies the effect of overburden pressure on liquefaction behavior of saturated Ottawa F65 sand using centrifuge experiments as well as numerical simulations. A series of centrifuge tests were conducted simulating a 5 m layer of Ottawa sand having different relative densities and subjected to overburden effective pressures of 1 atm and 6 atm. The objective was to study the pore pressure response of sand to base acceleration under low and high overburden pressure. The sand layer had a bottom impervious boundary and a top pervious boundary, approximating an idealized field condition. The analysis shows that sand during liquefaction is not fully undrained but rather partially drained. It was also found that partial drainage was more significant at 6 atm than at 1 atm due to the increased coefficient of consolidation, c_v at higher overburden pressure.

The coal ash collected from the disposal site contains majorly silt size particles that are highly prone to liquefaction under seismic loading conditions. The current study investigates the cyclic and post cyclic shear response of coal ash under different densities (86%, 90%, and 95% MDD). Additionally, the influence of overburden stresses (50 kPa, 100 kPa, and 150 kPa) has been investigated for the coal ash specimens prepared at 95% MDD. The results concluded that the cyclic resistance ratio (CRR) was found to be increased with an increase in the density of the specimen. However, CRR was obtained to be decreasing with an increase in the overburden stresses due to higher positive pore pressure evolution at large overburden stresses. The rapid shear strain accumulation was observed at a lower number of loading cycles for specimens subjected to higher overburden stresses. The post cyclic shear strength was obtained to be higher as compared to the specimens with no cyclic history (never experienced earthquake loading in the past).

A reactive transport model is presented, which can be used to predict the distribution of substrates and products for microbial induced desaturation and precipitation (MIDP). MIDP is an emerging bio-mediated ground improvement method, which can be used to improve the cyclic resistance of the soil underneath existing structures. The process involves the injection of a substrate solution containing calcium acetate and calcium nitrate into a targeted soil layer. Indigenous nitrate-reducing bacteria produce biogenic gas (nitrogen and carbon dioxide), calcium carbonate minerals, and biomass. The gas gets trapped in the pore space, desaturates the soil, and can reduce pore pressure build-up during cyclic loading, which increases the undrained shear strength of loosely packed granular soils. Precipitation of calcium carbonate minerals can cement the soil particles and improve shear strength. In this study, a numerical model is developed using COMSOL Multiphysics, which couples the biochemical reactions to a reactive transport model and can be used to relate changes in substrate and product concentration to changes in porosity, degree of saturation, and hydraulic conductivity. The model is used to evaluate how the treatment performance is affected by different treatment variables, such as flow rate, treatment time, and injected substrate concentrations, and environmental conditions such as treatment depth and grain size distributions.

The effects of shaking duration on liquefaction-induced settlements have been studied in the past, but there is a lack of experimental data in the literature. Therefore, a series of moderate scale 1-g shake table experiments are conducted to evaluate the influence of shaking duration with the same input motion in liquefiable soil layers. First, a 1/5 scaled model is prepared based on a large-scale shake table experiment conducted recently at the University of California, San Diego. The scaled model consists of three soil layers with relative densities of 50% for the top crust, 30% for the middle liquefiable, and 85% for the bottom dense layers. Second, a shallow foundation is seated over the crust layer, which consists of unsaturated sandy soil. Third, the models are subjected to harmonic input motions with different shaking durations (i.e., 8, 12, and 15 s). Based on the prototype test, the expected settlement of the foundation was 25 mm; however, the measured settlement was 60 mm. The foundation and free field settlement increased from 7 mm to 157 mm and 3.5 mm to 14 mm when the shaking time was increased from 8 to 15 s.

Characterization of the liquefaction strength of sandy soils is essential in modeling geotechnical engineering problems involving liquefiable soils. This paper investigates the liquefaction strength of Ottawa F65 sand through an extensive series of undrained, stress-controlled, cyclic direct simple shear (CDSS) tests performed at different densities, overburden pressures, and static shear stresses prior to cyclic shearing. The relatively large number of CDSS tests is used to develop an artificial neural network (ANN) model in order to predict Ottawa F65 liquefaction strength for densities and loading conditions that are not available in the experimental results. The predictive capability of the ANN model is assessed using blind predictions of the cyclic strength in new CDSS tests for a relative density and vertical effective stress not available in the training data set. Afterwards, CDSS tests under similar conditions were carried out. The comparisons of the predictions with the experimental results suggest that the ANN model provides reasonably good and conservative predictions of the soil liquefaction strength and shows sensitivity to changes in vertical effective stress, soil density, and cyclic stress ratio.

Ground improvement is often used in engineering practice to mitigate the effects of liquefaction in the built environment. Prior centrifuge studies using bio-cementation (microbially induced calcite precipitation, MICP) have demonstrated increased liquefaction triggering resistance and reduced associated settlements by treating the entire soil model. One additional centrifuge test that treated discrete MICP zones within the soil model also showed improvement, but the complex interactions between the treated zones made it difficult to elucidate how treatment depth impacted system performance. This paper presents 1-m centrifuge experiments performed at the Center for Geotechnical Modeling at the University of California, Davis, to investigate how performance improvement relates to the dimensions and cementation level of a discrete MICP treated zone. Models were constructed with Ottawa F-65 sand at an initially loose state (D_R ≤ 50%) with different treatment zone dimensions. In one model, the entire soil body was cemented (18.9 × 39.6 × 8.22 m in prototype scale), while in the second and third models an 8 × 8 m plan area was treated to a depth of 8.22 m (100% depth), and 3.05 m (37% depth). The stiffness levels for the treated zone were selected and tracked using shear wave velocity with a target of ΔV_s = 300 m/s relative to an untreated model. In addition, a baseline uncemented model constructed at the same relative density was tested for comparison. All models were subjected to eight shaking events. Shear wave velocity measurements tracked changes in biocementation integrity through the progression of the shaking events. System performance during shaking was monitored using accelerometers, a linear potentiometer, and pore pressure transducers. Cone penetration soundings were used to measure the initial model state and track the model evolution. All models treated with MICP showed improved liquefaction resistance when compared to untreated models, which varied with MICP treatment zone dimensions.

Soil liquefaction evaluation is critically important in site seismic analyses. The presence of fines-dominated materials such as clayey and silty soils has been reported in many of the occurred liquefaction cases. Despite this observed fact, there are limited studies analyzing the presence of such materials in liquified sites compared to sandy sites. One of the practical methods to analyze such sites is to use Finn-Byrne formulation for pore water pressure buildup calculation along with other constitutive models such as simple Mohr-Coulomb. Recent advancements in constitutive models with more sophisticated PM4Silt and PM4Sand models allowed a more rigorous liquefaction analysis of such sites. In this study, these two modeling approaches are applied to Wildlife Liquefaction Array (WLA) monitored site. The WLA site profile consists of about 2.5 m of surface silty soil, 4.5 m of silty sand, followed by 5 m of silty clay, and a thick silty layer at the bottom. The liquefaction analyses of the WLA site under the 1987 Superstition Hills earthquake were conducted using these two approaches and compared to observed liquefaction. The results of the study showed that both approaches, despite significant differences in their theory and input parameters, have the capacity of predicting excess pore water pressure (EPWP) generation in the non-plastic silty layers. However, the PM4Silt model was able to produce a better prediction of EPWP variations in the low-plasticity, non-liquefiable deep silty layer during the earthquake.

Seismic prone regions undergoing major infrastructural progress require site-specific or region-specific dynamic soil characteristics for efficient earthquake resistant design of structures. Dynamic soil characteristics are traditionally represented using the strain-dependent dynamic properties (low strain shear modulus—G_max; modulus reduction variation—G/G_max; and damping ratio–D variation with shear strain—γ) over wide strain range. This study investigates the dynamic response of two soft cohesive soils sampled from the northeastern part of India, a highly active seismic region. Three advanced element testing techniques (bender element, torsional resonant column, and dynamic simple shear) have been adopted to assess the cyclic response at various testing conditions, including confining pressure, cyclic strain amplitude, loading cycles. The experimental data obtained from independent testing are combined together to have a comprehensive soil behavior over a wide range of shear strains (0.0001% to 5%). The experimentally obtained data have also been compared with traditionally adopted curves and found that a significant variation is noted in the dynamic response of tested soils. It is concluded from this study that mere utilization of empirical data is not feasible for dynamic analysis in the case of soft cohesive soils.

The undrained cyclic shear behavior of low-plasticity, fine-grained soils transitions from liquefaction to cyclic softening over a narrow range of plasticity index (PI). The uncertainty associated with characterizing the cyclic behavior of low plasticity silts has a significant impact on seismic vulnerability studies in the Pacific Northwest region of North America and other parts of the world. This study presents the results of laboratory cyclic shear tests on a low plasticity silt from a major infrastructure project in the Portland, Oregon, area. The soils in this study are defined as silty sands (SM) to sandy silts (ML) using USCS classification methods, with PI ranging between nonplastic (NP) and 5. The cyclic behavior of the soil is examined in the context of published relationships and data on similar soil types. The effects of pore water pressure during cyclic loading on post-cyclic reconsolidation strains and post-cyclic shear strength degradation are explored and compared to published studies on similar soils of the region. The results of this study contribute to the body of knowledge for cyclic behavior of low-plasticity silts and reduce uncertainty in design.

Predicting the response of soil profiles during infrequent events such as earthquake remains challenge in geotechnical engineering. The presence of soft soils adds more complexity to this system as the soil experiences a highly nonlinear response during strong ground motions. Nonlinear (NL) and equivalent linear (EQL) site response analyses are typically used for site-specific response analysis of soil, with NL analyses being preferred for soft soil sites under strong ground motions. The main objective of this paper is to use the results of a centrifuge model experiment together with one-dimensional NL and EQL site response analyses to investigate the seismic response of a soft soil profile of low plasticity. The centrifuge model was shaken through different shaking events, and the results from centrifuge experiments are compared with total stress ground response simulations performed using equivalent-linear and nonlinear analyses. The one-dimensional ground response simulations are performed using DEEPSOIL for EQL and NL site response analyses.

In this study, an experimental investigation is conducted to determine the soil-pile separation lengths and frequency-amplitude responses of a single pile under rotating machine induced coupled, i.e., horizontal and rocking vibrations. Forced vibration tests are performed in the field on the single pile made of a hollow steel pipe of length 5.8 m and an outer diameter of 0.166 m to determine the frequency-amplitude responses for four different eccentric moments. The frequency-amplitude responses display nonlinearity of the soil-pile system. The soil-pile separation lengths for all the eccentric moments are deduced by measuring the contact pressure between the pile and soil. The nonlinear responses characteristics of the soil-pile system is also obtained from the continuum approach analysis using the experiential soil-pile separation length and frequency-amplitude responses. From the analysis, the variation of the parameters responsible for the nonlinear responses of the piles with depth and eccentric moment are determined.

Permanent ground deformation induced by seismic faulting is a severe hazard for buried transmission pipelines. Over the years, many studies have attempted to investigate the factors that influence the performance of buried pipelines subjected to earthquake faulting. Some of the influential factors are normalized pipe thickness with pipe diameter (t/D), normalized burial depth with pipe diameter (H/D), fault offset rate, soil density, and soil moisture content. A better understanding of the effect of these factors or discovering new influential factors can lead to a more appropriate design of buried pipelines in the future. This study presents results from a series of experimental tests designed to investigate the influence of pipe thickness on pipe behavior while the outer pipe diameter remains constant. In contrast to previous studies that investigated the t/D factor by changing both pipe diameter and pipe thickness, this experimental research explores the independent effect of pipe thickness on soil-pipe interaction. The experimental tests were carried out in a split-box apparatus modeling the reverse faulting under 1 g conditions. Two copper pipes with thicknesses of 0.7 and 1.0 mm and the same outer diameter of 15 mm were used as the model pipes. A scaling factor n = 25 was used for the model pipes to represent a 600 mm diameter API 5L prototype steel pipe with wall thicknesses of 6.35 mm and 9.53 mm, respectively. The model pipes were buried in a medium dense sandy layer and subjected to 180 mm reverse fault offset. The results show that the thickness of the pipe has a significant effect on the maximum bending strain generated along the pipe. For the thinner pipe, it was shown that the required fault offset for buckling the pipe is about 30% less than the thicker pipe, indicating a safer response for pipes with a larger thickness value.

This study focuses on the load-deformation response of a heavily loaded rocking footing on sand reinforced by soil-cement columns via centrifuge-model scale shake table tests. In a previous study involving the same footing on unimproved ground, the footing exhibited overturning failure and excessive settlement. To mitigate this, the present study aims to control the kinematics of the footing via strategic ground improvement while maintaining the enhanced energy dissipation characteristic of a rocking footing. The ground improvement technique consists of soil-cement mixed columns, which were cast externally and then placed in the model sand layer during pluviation. Shake sequences are applied to both determine the frequency response of the soil-structure system and induce controlled yielding in the improved soil. The addition of paired soil-cement columns at the ends of the footing was found to significantly reduce settlement during rocking and preserve recentering throughout demanding shaking events.

Sites located in the Sacramento-San Joaquin River Delta region of Northern California typically have peaty organic soils near the ground surface. These soils are characteristically soft, with shear wave velocities as low as 30 m/s, and may be present to depths as great as 15 m. These unusually soft geotechnical conditions will produce significant site effects in the event of a strong earthquake, and are not expected to be reliably predicted using existing ergodic site amplification models. This study presents site characterization data related to stratigraphy, shear wave velocity (V_S), and microtremor-based horizontal-to-vertical spectral ratio (mHVSR) surveys collected across the delta and archived as part of a project to develop a delta-specific site response model. Spatial variations of time-averaged shear wave velocity 30 m (V_S30) across the delta region are mapped and compared to other site parameters. An analysis of 14 sites with HVSR and V_S profile data shows examples of sites with and without mHVSR peaks. Sites without peaks also lack layer boundaries with large impedance contrasts within the range of the V_S profile. Sites with peaks have such boundaries in some but not all cases.

Results from laboratory tests have shown that effective confining pressure and non-zero initial static shear stress can affect a soil’s cyclic resistance ratio (CRR). To account for these effects, the factors K_σ and K_α are used to modify the CRR, respectively. However, there is no consensus on the use of existing K_σ and K_α models. To facilitate the validation of existing models and the development of new models, we compiled results from 63 published studies and organized them in a digital data set. The compiled data set contains 3,493 CSR vs. N (cyclic stress ratio vs. number of cycles) points, 960 CRR values, 231 K_σ and 326 K_α values that span a wide range of soil type, relative density, initial stress, fines’ content, and specimen preparation method. This data set greatly extends the amount of data available for evaluation and development of K_σ and K_α models and is provided for use in next generation liquefaction (NGL) models.

This paper presents the development of base case shear-wave velocity (V_S) profiles for computing seismic amplification factors for reference site to local site conditions in the state of South Carolina. As part of an ongoing statewide hazard mapping study for the South Carolina Department of Transportation, the reference site conditions were characterized using time-averaged, shear-wave velocity in the top 30 m (V_S30) of 1,183 to 1,518 m/s for different general rock types in the South Carolina Blue Ridge and Piedmont (SCBRP) provinces, and 345 and 385 m/s for younger tertiary sediments and older sediments, respectively, in the South Carolina Coastal Plain (SCCP) province. In this paper, a total of 189 V_S profiles are compiled from various project reports and used to characterize the residual soil and saprolite that overlie partially weathered rock in the SCBRP and the quaternary soil that overlie tertiary and cretaceous sediments in the SCCP. The resulting median V_S profile ranges from 253 m/s near the ground surface to 402 m/s at a depth of 30 m for residual soil/saprolite, and from 170 m/s near the ground surface to 215 m/s at a depth of 25 m for quaternary soil. The transitions from soil to reference site conditions are also modeled. To capture the epistemic uncertainty in V_S and variations in depth to the reference site conditions, two sets of base case V_S profiles are proposed for the SCBRP and SCCP.

The substructure method, which is a widely used method in engineering practice to assess soil-structure interaction (SSI), suggests the motion at the foundation (FM) is usually controlled by the simultaneous contribution of kinematic SSI, where the stiffness of a massless foundation deviates the FM from free-field motion (FFM), and inertial SSI, where the structure’s inertia further modifies the FM. The kinematic SSI effects is estimated by calculating the ratio of response spectra (RRS) between FM and FFM using simplified procedures that were calibrated semi-empirically using earthquake databases mainly from California only limited to horizontal translational motions. This study evaluates empirical vertical translational RRSs based on recorded data at five nuclear facilities in Japan. The results illustrate that simplified design guidelines for horizontal RRSs tend to match the shape of the foundation vertical translational RRSs but require an amplitude modification. In conclusion, the comparisons establish the need for a dedicated simplified procedure for vertical translational RRSs.

Modulus reduction and damping (MRD) curves are essential in estimating the response of soils to earthquake ground shaking in seismic site response analysis. Microbial-induced carbonate precipitation (MICP) is a biological grouting method, with demonstrated effectiveness in shear strength improvement and potential for liquefaction mitigation. However, the effect of MICP on seismic site response has rarely been studied, and the MRD formulation for MICP-treated soil (MRD_MICP) applicable to numerical site response analysis has yet to be developed. To overcome this hurdle, (1) we provide a best-fitting MRD formulation of MICP-treated sand for varying cementation levels at a constant confining pressure based on laboratory dynamic tests, and (2) we used the proposed formulation in simulating a one-dimensional nonlinear site response analysis for a hypothetical MICP-treated site. We assumed a 35 m-deep soil column composed of saturated, homogenous, and poorly graded sand that will be locally treated with MICP a few meters above the assumed bedrock for the site response analysis. Outcrop ground motions with weak and strong shaking intensities were used as input at the bedrock level. Nonlinear site response analysis showed a reduced pore water pressure (PWP) ratio at the MICP-treated layers. However, the PWP ratio and ground motions at the surface layer (that remained untreated) experienced a rebound and amplification, respectively. Optimum cementation levels that minimize the amplitude of surface ground motions and reduce the PWP ratio at the soil layers were different when considering different levels of ground shaking. Hence, it is important to further elucidate the relationship between levels of cementation, hazard consistent ground motions, and potential site effects.

In this study, the seismic response of shallow foundations resting on a liquefiable soil layer is modeled using a coupled smoothed particle hydrodynamics (SPH)-discrete element method (DEM) scheme. In this framework, the soil deposit is represented by an assembly of DEM particles, and the fluid domain is lumped into a set of SPH particles carrying local fluid properties. The averaged forms of Navier-Stokes equations dictate the motion of the fluid-particle mixture, and the interphase forces are estimated using well-known semi-empirical equations. A saturated soil-foundation system with an average contact pressure of 50~kPa was created using the coupled scheme. The foundation block was composed of a collection of DEM particles glued together by high-stiffness bonds. No-penetration boundary condition was applied to all sides of the foundation block to allow for fluid-foundation interaction. The model was subjected to a strong base acceleration, and the response was analyzed and compared to the free-field. The obtained results indicate that ground settlement in the soil-foundation system mostly originated from co-seismic deviatoric deformations, while volumetric strains were the main contributing factor at the free-field. In addition, unlike the soil-foundation system, a large part of the free-field settlement occurred post-shaking.

In engineering practice, one-dimensional seismic site response computer programs were used to estimate the spectral responses at the ground surface, assuming level ground conditions. However, in reality, the ground surface and subsurface layers can be inclined, which may affect the propagation of earthquake waves through soil layers. This study aims to conduct two-dimensional (2D) site response analyses to develop slope adjustment factors to update the site coefficients computed from one-dimensional (1D) analyses for Charleston, SC. Eleven shear wave velocity profiles were generated considering the mean shear wave velocity (V_s) profile, mean ± 1σ (σ = standard deviation) of the natural logarithm of V_s, and variation in the thickness of quaternary soil layers. The modulus reduction and damping curves were developed based on the region-specific relationships. Four earthquake time histories that represented site-specific seismic hazards were selected. Finally, 308 2D finite element models were developed and analyzed considering seven ground inclinations (0^o, 1^o, 2^o, 3^o, 4^o, 5^o, and 6^o), 11 V_s profiles, and four earthquake time histories. The results show significant differences between 1D and 2D analyses, justifying the importance of considering sloping ground conditions in seismic analysis to capture the realistic ground motion characteristics at the site. Based on the results, a set of slope adjustment factors were developed for various average shear wave velocities at top 30 m (V_s30) and ground inclination (θ) cases. These slope adjustment factors can account for the mild sloping ground (θ≤6^o) effect by multiplying it with site factors obtained from the 1D analysis for their respective combination of V_s30, ground inclination, and spectral period.

This paper infers shear wave velocity profiles and soil dynamic properties from acceleration recordings at the instrumented Hayward I-580/I-238 bridge site in California where two geotechnical downhole arrays are located. The geotechnical arrays include three-component accelerometers at the surface and various depths down to the depth of 90 m. A total of 15 earthquakes were recorded during 2011–2022 ranging from Mw3.4 to 6. The cross-correlation method was employed to determine the average shear wave velocity profiles. The estimated shear-wave velocity profiles show agreement with the results of corresponding P-S suspension logging data. In addition, the uncertainty envelopes for the shear wave velocity profiles were generated. Furthermore, the shear modulus and damping properties were calculated, which showed a more linear behavior at the east downhole array compared to the west downhole array, which was consistent with the sub-surface soil information. The damping ratios increased with shear strain and were generally less than 1.2%.

The fidelity of the surface wave inversion is greatly influenced by the efficiency and accuracy of the forward analysis. Existing forward modelling approaches generate only theoretically calculated dispersion curves without any information regarding the modal amplitude. The present study proposes a semi-analytical forward modelling approach to determine the dispersion image from a 1D profile similar to that produced by multichannel field data. An analytical solution is used in the horizontal direction, and finite element discretization is applied in the vertical direction. The infinitely extended elastic half-space is efficiently modelled by using perfectly matched discrete layer (PMDL) elements. The higher-order finite element discretization and the use of PMDL elements significantly reduce the computational cost. The proposed method calculates the vertical displacement responses at different receiver positions on the surface in frequency-space domain. The complex frequency-space domain response is transformed into seismogram data and velocity spectra.

The purpose of this study is to compare results of site response analysis with the centrifuge model response of gravelly soils. A series of centrifuge modeling tests were performed at 50-g centrifuge gravitational acceleration to simulate a uniform soil profile subjected to a 1-Hz sinusoidal shaking motion with peak amplitudes ranging from 0.01 to 0.10 g. The soil tested in this study is a gravel-sand mixture of 50% gravel and 50% sand by weight, prepared in a loose condition. Dynamic soil properties of the material were assessed based on several methods, including resonant column, centrifuge bender element, and dynamic response of centrifuge model tests. Equivalent linear site response analysis was performed using DEEPSOIL to analyze the response of the same soil column as the prototype soil modeled in the centrifuge. An overall comparison of results shows that the equivalent linear site response analysis does not provide a good agreement of shear stress and shear strain for the loose gravelly soil observed in the centrifuge tests. At the shaking amplitude of 0.05 and 0.1 g, the shear strain appears to be comparable, but the equivalent linear method underestimated the shear stress. For the shaking amplitude of 0.01 g, the equivalent linear method overestimates surface response acceleration at frequencies higher than 2 Hz due to the variation of shear modulus reduction curve. For the shaking amplitude of 0.05 and 0.10 g, the equivalent linear method underestimates the surface response acceleration because the shear modulus decreases as the number of cycles increases.

Variations between foundation-level and free-field ground motions can be mainly attributed to the kinematic soil-structure-interaction (SSI) effects. Current design guidelines such as ASCE 41-17 include a simplified procedure to estimate the kinematic SSI effects for regular-sized buildings with footprint sizes less than 260 ft and embedment depth shallower than 20 ft. Therefore, modifying such equations for the design of buildings with large foundations and embedment depths is of great importance, especially when safety-critical structures are involved. A recent study at the University of Nevada, Reno (UNR) has recently proposed modifications to the ASCE 41-17 equations taking into account the unique features of such buildings using a data set of earthquake motions recorded at five instrumented sites in Japan. The finite element method (FEM) is an efficient numerical tool to take the kinematic SSI effects into account. However, in engineering practice, it is often the case that the actual recordings at the location of the site are not available to validate the FE models against. This study benchmarks the results of the FEM, the simplified code-based formulations, and the modified simplified equations against the actual recordings to show the suitability of the modified simplified equations for validating the developed FEM models. For this purpose, a three-dimensional (3D), nonlinear (NL) SSI model of a well-instrumented site located in Japan, including a building with a large foundation and embedment depth, is developed in the time-domain code LS-DYNA. Both low-intensity and design-level records from the recordings at the site are applied to the model. It is demonstrated that the FEM captures the kinematic SSI effects appropriately. It is also shown that the modified simplified equations show good agreement with the results captured via FEM. It is recommended that for buildings with large footprints and embedment depths, the modified simplified equations are used instead of current simplified procedures to validate the SSI models in case the actual recordings are not available for validation purposes.

In this paper, we propose a machine learning-based approach for predicting foundation settlement on liquefiable soils improved through ground densification. The model considers variations in the properties of the soil profile, foundation, 3D structure, mitigation design (in terms of densified depth and width), and ground motion. A numerical data set from 770, 3D, fully coupled, effective-stress, finite element analyses was developed initially with a statistically determined range of input parameters (through quasi-Monte Carlo sampling). The numerical models were themselves calibrated and validated with centrifuge model studies. Subsequently, the numerical database and an additional 15 centrifuge experiments were used to train a gradient boosting model (tree-based, supervised, machine learning method, GB) for predicting foundation’s settlement. In general, the data-driven GB model could better predict settlement compared to the classical regression model (by about 14%). This is because the non-functional form model could better capture the nonlinear trends in permanent foundation settlement as observed in the numerical and experimental database. However, when evaluating a very limited existing field case history database in the literature, the data-driven GB model only slightly improved the settlement predictions compared to the regression model. This is because the GB model cannot take the impact of model features on foundation settlement in a continuous manner (due to the inherent shortcoming of a decision-tree framework in GB), leading to a dramatic increase in model uncertainty when the input parameters are outside the ranges considered in the database. The insight from the presented GB model aims to guide the development of future data-driven predictive models for a more reliable estimation of engineering demand parameters related to soil–foundation–structure systems.

Dynamic responses of the Lenihan and Austrian Dams in the 1989 Mw = 6.93 Loma Prieta earthquake were analyzed using the finite element program VERSAT; undrained response of the saturated dam fills was modelled using a total stress approach. The two dams were constructed in the early 1950s and comprised primarily of compacted low plasticity clayey sands and clayey gravels. Subjected to estimated peak accelerations of 0.44 g at Lenihan Dam and 0.55–0.6 g at Austrian Dam, the dams developed some to extensive longitudinal cracks on the dam faces and settled at the dam crest on average 0.25 m at Lenihan Dam and 0.76 m at Austrian Dam. This study demonstrates the robustness of the adopted approach in capturing the embankment dam’s key performance under large earthquakes. The nonlinear response history analyses indicate that the proposed S_u/σ'_m approach for calculating in situ undrained shear strengths can be a reasonably conservative approach for engineering analysis and design. The calculated dam crest settlements for both dams are in good agreement with the dam crest settlements observed in the 1989 earthquake.

Earthquake-induced slope displacements are typically estimated using a sliding block analysis. Sliding block approaches, for both rigid and flexible sliding masses, have been shown to roughly capture the observed seismic performance of slopes during previous earthquakes. However, these approaches incorporate simplifying assumptions that do not represent the full dynamic response of a slope. Two-dimensional (2D), nonlinear, total stress finite-element analyses were performed to identify the seismic deformation patterns for shallow (friction-dominated) and deeper (cohesion-dominated) critical sliding masses. The numerical parametric analyses consisted of four models with soil stiffness and strength parameters selected to isolate the impact of the slope seismic resistance (i.e., k_y) and the depth of the slip surface on seismic slope deformation patterns. Each model was analyzed in the context of how the ground motion intensity (in terms of peak ground velocity, PGV) affects the maximum permanent displacement along the surface of the slope, the profiles of permanent displacement and shear strain, and the stress-strain behavior at different depths. The results indicate that for the same k_y and same T_slope, the displacement is significantly affected by the depth of the critical sliding mass due to the distribution of static shear stress and soil’s shear strength along with the slope height. For the conditions considered, at small k_y the deeper sliding masses experience larger displacements due to the presence of localized straining at depth. Their slope displacements are about 1.2–2 times greater than their counterparts with shallow sliding surfaces, particularly when the PGV is greater than 20 cm/s. At large k_y, shallow sliding masses experience more distributed straining across the entire sliding mass depth (and hence slope displacements), leading to displacements about 1.25–2 times larger than the deep sliding mass models. These insights point to the importance of considering more advanced numerical simulations (as opposed to traditional sliding block analysis) when quantifying the seismic performance of a slope.

The paper presents a series of centrifuge tests of level sites consisting of a granular soil deposit, with and without an embedded structure subjected to various biaxial and uniaxial base excitations. The tests were conducted at RPI NEES centrifuge facility to assess the dynamic response characteristics of level deposits with and without an embedded structure under multidimensional conditions. Synthetic sinusoidal waves were used as base excitations to test soil models under biaxial and uniaxial shaking. A dense array of accelerometers was used to monitor the deposit response along with pore water pressure transducers. Two uniaxial tests and three biaxial shaking tests were conducted on three similar soil models, to study the effect of multidirectional shaking on the generation of soil liquefaction for loose and dense deposits. The observed acceleration and pore pressure are used along with non-parametric identification procedures to evaluate the corresponding dynamic shear stress–strain histories. The measured results along with the obtained time histories are used to shed light on the mechanisms of liquefaction occurring through the stratum under multiaxial shaking. It also shows the effect of soil relative density on soil behavior when subjected to biaxial shaking. The soil structure interaction (SSI) test showed significant differences in soil behavior at different locations beneath the footing when subjected to biaxial shaking. This difference is associated with coupling in stress–strain in both experiments compared to the more known uniaxial soil behavior.

In this study, the number of periods in acceleration response spectra and the characteristics of input motion and screening criteria on the representative earthquake time histories are investigated for the Charleston, SC, area. Uniform hazard spectra (UHS) obtained from probabilistic seismic hazard analysis (PSHA) and NEHRP provisions typically contain 10–22 spectral periods ranging from 0.01 to 10 s. Acceleration response spectra for recorded time histories generally have peaks and valleys between the spectral periods used for the UHS. To address this issue, the UHS was re-digitized by adding more spectral periods (50–300) using a linear interpolation technique in semi-log space. The seed time histories were selected from the NUREG/CR-6728 ground motion database and modified to match the re-digitized UHS using a time domain modification technique. The results showed that using re-digitized UHS for spectral matching minimizes the peaks and valleys. However, sensitivity analyses showed that the response spectra of the modified acceleration time history matched the UHS well when the number of re-digitization periods increased but resulted in a considerable distortion in the corresponding displacement time histories. The study provided insights into the effect of the number of spectral periods and criteria for screening seed time histories for spectral matching to generate site-representative time histories.

Shear modulus and damping ratio are considered essential soil parameters for ground response analyses under dynamic loading. The dependency of these two key parameters on the amplitude of induced strain for several types of soil has been studied in the laboratory and in situ tests under a limited set of controlled environmental factors, mainly soil confinement, plasticity, and moisture content. The present research investigates the influence of an additional potential critical factor, i.e., soil temperature. The thermo-controlled resonant column (RC) tests were conducted on low plastic clays (CL), and the effects of rising temperature on the soils’ dynamic parameters were thoroughly investigated, with particular emphasis on the shear moduli degradation and variations in damping ratios (D) vs. G/G_max under thermally controlled conditions. The results suggest that the ground temperature might not play a significant role in the assessment of suitable values of the shear modulus and damping ratio for ground response analyses.

Accurate estimates of ground motion intensity measures (IMs) at liquefaction case history sites are crucial as they are used on the demand side of the equation for regressing liquefaction triggering and consequence models. Therefore, it is desirable to have IM estimates that are accurate (i.e., unbiased), have defined uncertainty levels that reflect the available information, and that are consistently developed across sites in liquefaction databases. Many of the earthquakes that form the basis for current liquefaction triggering relationships occurred prior to 1999 and produced relatively limited numbers of ground motion records. By contrast, modern earthquakes in California, Japan, and elsewhere often produce large numbers of densely recorded ground motions, which enable the use of spatial interpolation to interpret shaking intensities at sites of interest. We present a method that uses Kriging to interpolate IM within-event residuals to estimate ground motions at liquefaction sites. By interpolating within-event residuals rather than IM values, the Kriging is more stable because systematic path effects and first-order site effects are accounted for.

Non-ergodic ground-motion models (NGMMs) have the potential of reducing the ground-motion aleatory variability significantly, which has a large impact on the seismic hazard, especially at large return periods important for critical infrastructure. This reduction in aleatory variability is accompanied by epistemic uncertainty in regions with sparse recordings or a systematic shift in the median ground motion in regions with dense recordings. Gaussian process regression (GPR)—with spatially varying coefficients for modeling the source and site systematic effects and cell-specific anelastic attenuation for modeling the systematic path effects—is a flexible and robust modeling technique used in this study for developing NGMMs. As part of this work, open-source computer tools and instructions have been developed to show the steps toward developing NGMMs in the GPR framework. Statistical software packages STAN and INLA are used and compared. The developed software packages were tested against synthetic data sets with known non-ergodic effects, and different implementations of the developed software were evaluated for scalability, universality, precision, and model complexity.

This paper summarizes ground motion models (GMMs) developed for the inelastic response spectra of shallow crustal events. The GMMs are developed based on a large database of recordings in the NGA-West2 database with magnitude between 3.3 and 7.9, and source-to-site distances less than 80 km. GMMs are developed for six displacement-ductility levels, two damping values, and two hysteretic behavior models. One of this work’s main novelties is that the model is developed for RotD50 inelastic response instead of geometric mean, making the results directly comparable to the models developed in the NGA-West2 project for the response of elastic oscillator. The ground motion model incorporates the effect of magnitude, focal mechanism, source to site distance, site characteristic, and basin effect on the inelastic response of structures. The model shows that the yield strength demand decreases significantly when ductility goes from 1 to 1.5 (moderate nonlinearity is allowed); however, the decrease is not as significant for higher ductility levels (3 or 4). This study also quantifies the range of applicability of the “constant displacement rule.” The effect of response nonlinearity on the aleatory variability of ground motions is also quantified.

Naturally cemented sands exhibit heightened mechanical properties compared to clean sands as a result of the action of crystal binders that strengthen the structure of such materials. Based on this observation, a myriad of ground improvement techniques such as bio-cementation and polymerization have been developed to artificially cement sands. Despite such techniques finding increasing applications, a handful of constitutive models remain available to capture their effects on the mechanics of sands. To advance the state of the art, this article describes a constitutive model for capturing the mechanics of cemented sands: SANISAND-C*. This work specifically focuses on a succinct overview of the mathematical formulation of such model and its application to the simulation of drained and undrained triaxial compression tests of cemented and clean sands. Results indicate that the model can realistically reproduce multiple effects caused by the natural or artificial cementation of sands, such as their increased strength and stiffness, changes in dilatancy, and the destructuration of cementation bonds upon mechanical loading. Accordingly, the model represents a powerful tool for computational analyses involving cemented sands for both scientific and engineering purposes.

This paper presents the results of a series of laboratory experiments performed to evaluate the erosion potential of silt and clay deposits that were subjected to repeated freeze-thaw cycles. The effects of climate change are more pronounced with several soil erosion, slope failures, and other mass-wasting event sites. To evaluate the erosion potential with cyclic freezing and thawing, soils from different geological formations with dominant silt and clay fractions were collected. Basic characterization of soils including unified soil classification system (USCS) classification, Atterberg limits, Proctor tests, and mineralogy were determined. The samples were prepared with three different moisture contents’ optimum moisture content (OMC) and ±5% OMC. Samples were subjected to freeze-thaw (F-T) patterns using a custom-built device that facilitates top-down freezing similar to in situ conditions. After 0, 1, 2, 10, and 20 F-T cycles, samples were tested for erosion potential using a modified erosion function apparatus for different flow conditions and molded moisture content. The coefficient of soil erosion, sediment loss, and volume of eroded soil was determined for the samples after each cycle. Test results and analyses depicted that the erosion potential of these soils significantly increased with an initial increase in F-T cycles. This paper highlights the vulnerability of the soils to erosion after thawing.

Wind erosion is responsible for the formation of various landforms and poses a significant risk to health, food security, and infrastructure. On Earth, the initiation of wind erosion is affected by the water content of the sediments, which can significantly increase the friction threshold velocity (FTV), which is the minimum wind velocity at the bed surface needed to initiate particle mobilization. Incorporating the effect of water on FTV is essential to assess problems such as sandstorm formation, topsoil removal in crops, or sand dune migration. In this study, a theoretical model is proposed to calculate the FTV in sands and horizontal surfaces using the principles of unsaturated soil mechanics. The sand matric suction is incorporated using soil water retention curves (SWRC). Two SWRCs are used resulting in simple functional forms that are easy to use. Data from the literature and new wind tunnel experiments are used to calibrate and validate the suggested FTV model. The results support the use of the proposed simplified equation as providing an adequate prediction of the FTVs of moistened sands.

Soil piping is the gradual and progressive erosion of soil grains, causing a void (open pipe) to form as water flows through the soil. In dam engineering, this type of internal erosion is often referred to as concentrated leak erosion and has been a cause of failure at multiple dams. Soil piping has also been observed in many landslides and contributes significantly to soil degradation in hillslopes and agricultural areas. Despite these many important impacts, there is still limited understanding of how soil pipes develop and progress and what factors control pipe stability. One of the significant challenges with analyzing soil piping, or concentrated leak erosion, is that it typically occurs in the vadose zone, where unsaturated conditions are present. However, most studies examining internal erosion have focused on saturated conditions, and few studies have examined the role unsaturated hydraulic properties (i.e., air entry value, matric suction, etc.) may play in the likelihood of internal erosion. Consequently, this study aims to explore the mechanisms controlling the erosion rate within soil pipes from the perspective of unsaturated soil mechanics. Bench-scale experiments were performed to examine the formation and progression of an eroded pipe in a small slope constructed at different water contents. Soil samples were also tested to measure its unsaturated hydraulic properties. The results show that the likelihood of pipe formation varies with the moisture content and, therefore, suction in the soil, as does the potential for pipe collapse. This demonstrates that unsaturated soil properties are key to understanding the formation and progression of piping in slopes.

The erosion function apparatus (EFA) is a laboratory device that can determine the relationship between the erosion rate and the hydraulic shear stress of a soil specimen tested; this relationship also estimates the critical shear stress required to initiate soil erosion. The computed hydraulic shear stresses from EFA testing are a function of the mass density of water, the mean water flow velocity, and a friction factor relating to the Reynolds number and the relative roughness of the soil sample after erosion. Since the water and flow geometry characteristics are well defined, the primary variable in computing shear stress is the soil surface roughness. There are a variety of methods available to measure the roughness; however, the location, accuracy, and resolution of the surface measurements can affect the final roughness value used in computing the corresponding hydraulic shear stress. This paper presents the results of different soil roughness measurement methods on five field samples after EFA testing and the impact those measurements have on the computed hydraulic and critical shear stresses. Findings suggest that measuring points across the entire sample (including points close to the perimeter) and taking readings every 0.3 points/cm² using a laser texture scanner or similar will result in a reasonable approximation of the erosion function of soils.

The Federal Highway Administration (FHWA) is developing NextScour, a scour research initiative to improve scour analysis and provide more accurate scour depth estimates for bridge foundation design. NextScour considers the two major components of scour: hydraulic loads and the erosional resistance of soils. Through an ongoing transportation pooled fund study, FHWA partnered with several state departments of transportation to provide soil and erosion testing services for bridge scour evaluations, which has resulted in several excellent case studies for the NextScour initiative. This paper summarizes the evaluations of two bridge replacement projects in Michigan (MI) and North Carolina (NC). The proposed bridge site in MI had a calculated scour depth that exceeded 11 m; however, a layer of medium-to-hard clay at depths of 4.6–6.1 m below the channel was identified in the geotechnical site investigation. The NC bridge site has a shallow medium to very stiff clay layer 1.2 m underneath the fine sand of the riverbed. As part of the subsurface exploration, samples of both clayey soils were collected and tested by FHWA to obtain the erosion resistance and other geotechnical properties. The resulting critical shear stress measurements suggested that both clay layers would be resistant to scour from the design floods, potentially saving costs on the foundation designs for both bridge replacements. This paper presents an overview of the NextScour initiative, discusses current and proposed methods to evaluate hydraulic loads and soil erosion resistance, details the two case studies, and provides a summary of the hydraulic and geotechnical testing conducted within the context of NextScour.

Riprap is commonly used to prevent erosion around different hydraulic structures such as riverbanks, shorelines, streambeds, bridge piers and abutments, and dams. When riprap is installed on the riverbanks, its stability is affected by the stream flow, slope angle, and material properties of the riprap elements. The magnitude of the slope angle affects the critical velocity and critical shear stress of the riprap rocks, which could endanger the stability of the riprap and cause the failure of the riprap layer. In recent years, a combination of computational fluid dynamics with the discrete element method, CFD-DEM, has proven its efficiency in the simulation of micro behavior of erosion around different hydraulic structures. In this research, using this tool, multiple numerical simulations are performed to study the effect of the slope angle on the critical velocity of individual riprap rocks. The effect of other parameters such as angularity of the rock elements is also studied. Results of the numerical simulations are summarized in a chart that shows the critical velocity and critical shear stress of individual riprap elements as a function of the slope angle.

Vegetation can be a natural and cost-effective bioengineering solution to improve the stability of hillslopes. The mountainous hilly terrain of tropical islands including Guam is plagued with numerous rainfall-induced shallow slope failures. Even several houses are located along coastal hillslopes, and as such, there is a dire need for finding eco-friendly solutions to protect the slope failures and loss of soil into oceans. In this study, soil hydrological and geotechnical properties including the soil water characteristic tests were conducted on native soil collected from hillslopes in southern Guam. Also, Pandanus tectorius, a native tropical plant, was selected, and its roots were tested for their tensile strength. It was found that the tensile strength offered by roots decreased with increasing root diameter. The best-fit exponential and power-fit relationship between root diameter and tensile strength were established. Root morphological characteristics were documented, and root bundle theory was used to translate the tensile strength offered by roots into additional root cohesion, i.e., root enhanced mechanical shear strength by soil-root matrix. Numerical simulations show that vegetation was able to provide a decent amount of stability through an increase in factor of safety via root water uptake and mechanical root reinforcement.

Power transmission and subtransmission lines in California often face unpredictable climate change, high seismic activity, and geohazards such as slope instability. One such line is a transmission line that traverses over 50 mi of mountainous terrain from Ventura County to Santa Barbara County. Most of the transmission line structures along this line are located in remote mountainous locations and are founded on sedimentary bedrock formations. Some are in landslide terrain, adjacent to mapped landslides, on steep slopes, or on ridgetops. The instability risk for these structures increases with significant precipitation and/or seismic events. Loss or damage to any of these structures due to ground deformation would result in disruption of service to all subtransmission and distribution lines in the service territory. Since the structures’ remote locations make prompt repairs impractical, preventative mitigation is often necessary. However, mitigation is expensive, so measured and targeted mitigation becomes essential. Through a geological desktop study and site visits, the locations of structures that are most prone to landslides were identified. Groundwater mounding modeling, field investigations, and other geotechnical and hydrologic studies were conducted at these sites. The project team estimated peak ground accelerations and spectral accelerations (PGA and SA), with 2% and 10% probability of exceedance in 50 years. They used the historic precipitation data to generate landslide-triggering precipitation criteria and create a monitoring tool to predict rainfall. The probability of exceedance depends on the project life cycle. The team first calculated the factor of safety (FOS) under saturated conditions due to precipitation, and then they calculated the seismic load. The seismic vulnerability was assessed, excluding the impact of storm events. The FOS was calculated based on the project life cycle, where line resiliency is determined using relationships between FOS and the probability of exceedance, considering storm recurrence intervals. In short, this transmission line is a single-point service interruption system prone to landslides, and due to its remote location, it is difficult, time-consuming, and expensive to fix. Mitigation measures were developed and implemented. The approach requires determining which locations pose the highest risk, defining the circumstances that require mitigation, and selecting efficient mitigation measures based on the structure’s resilience.

A geospatial database of over 24,000 slope failures across the state of North Dakota was developed. The database revealed that nearly 1,500 slopes failures were larger than 100,000 m² in area posing significant safety concerns and engineering issues. To better understand the influence of various conditioning factors, a statistical study was undertaken in this paper. The slope failures were found to be concentrated in western North Dakota primarily occurring in the Sentinel Butte Formation. Statistical analyses suggest that this is likely due to the steeper topography and the abundance of Sentinel Butte Formation in this region in comparison to eastern North Dakota. The number of slope failures were log-normally distributed with the slope inclination. The mean slope inclination on which slope failures occurred was found to be 9.5° with a standard deviation of 5.7°. Interestingly, despite the presence of high salt concentrations in the soils underlying the majority of the state, the slope failures were found to be concentrated in areas with low salt concentrations. This may be attributed to the lower shear strengths of the soils with lower pore fluid salinity than that of soils with greater concentrations of salinity in the pore fluid.

One of the major causes of slope failure is elevated groundwater; lowering groundwater levels and consequently, maintaining low pore water pressures is a key element toward landslide mitigation. Horizontal drains are a common method toward dewatering and are often installed in specified patterns, lengths, and locations to be most effective based on site conditions and operator experience. As the design of horizontal drains is often observation-based practice, we developed a three-dimensional solution for transient groundwater drawdown stemming from horizontal drain installation in an unconfined aquifer by using a custom-made finite difference solver. First, the solution was verified against experiment results. Then, a sensitivity analysis was performed to evaluate the effect of drain spacing and hydraulic conductivity on drawdown and factor of safety in a simplified hillslope. The results indicated overall effectiveness of horizontal drains and improvement in slope stability; however, changes in drawdown and consequently factor of safety varied with drain spacing, site hydrological conditions, and time after installation.

Landslides triggered by rainfall are becoming major threat in residual soil slopes worldwide. Several rainfall triggered shallow landslides were reported in residual soil slopes of Idukki in Kerala, India, during monsoon. Existing slope susceptibility maps and rainfall thresholds are generally based on statistical methods, which seek empirical relations between the geomorphological, topographical, geological, and environmental factors with the past landslides, without taking into consideration the geotechnical properties of the soil. Both saturated and unsaturated soil properties are important in analyzing the stability of residual soil slopes subjected to rainfall infiltration. In the present study, a landslide susceptibility model was developed for a region in Idukki by using two physical-process based models, TRIGRS (transient rainfall infiltration and grid-based regional slope-stability) and Scoops3D, which have the potential to incorporate both saturated and unsaturated soil properties in the regional modelling analysis. Transient rainfall infiltration analysis was performed using TRIGRS and the resulting pore water pressure profiles imported in Scoops3D for stability analysis. The developed model was validated with past rainfall event of August 2018, which was responsible for many landslides at Idukki, India. Predicted and the observed landslide points were compared against each other in both temporal and spatial perspectives using LR_classindex. The results show that the models have a good capability of landslide prediction, and therefore, these can be used for the development of both slope susceptibility map and in designing of warning system against the rainfall induced landslides.

Due to the complex nature of landslides, statistically based landslide susceptibility mapping has been widely used to evaluate slope failure risk for landslide-prone areas. This study evaluates the capability of different machine learning (ML) methods for landslide susceptibility mapping (LSM) in mountainous regions in the Colorado Front Range. A well-documented and georeferenced landslide inventory for the Colorado Front Range area was used to construct the database for developing and testing ML models. Nine landslide contributing factors were collected for the present study based on the availability of geophysical data and the type of landslides that occurred in the study area. These landslide causative factors represent hillslope geometries, surface hydrology, and soil conditions. Five commonly used ML models were evaluated: logistic regression (LR), support vector machine (SVM), decision tree (DT), random forest (RF), and gradient boosting machine (GBM). The cross-validation technique was used to evaluate the model performance. All the trained models reflected the relationship between landslides and their causative factors in the study area based on cross-validation results. It was found that the performance varied among the ML models; the RF model exhibited the worst performance due to possible overfitting, and the RF and the GBM models achieved the highest performance. The trained models were subsequently used to predict the landslide susceptibility for the entire study area and generate a landslide susceptibility map. The landslide susceptibility map can provide situational awareness of potential landslide hazards within the Colorado Front Range area and provide guidelines for future decision-making.

Seismically induced rock slope failures have caused tens of thousands of deaths and economic losses in the billions of dollars over the last century. They are among the most common, dangerous, and least understood seismic hazards. Recently, progress has been made toward understanding the fundamental mechanisms of seismic rock slope failures using the bonded particle model (BPM). Although powerful, BPM is computationally intensive and difficult to scale for regional or probabilistic analyses. In contrast to BPM, sliding block methods are quick to run and widely used in seismic soil slope analysis. However, sliding block methods do not explicitly capture rock slope behavior during earthquakes. This study presents a framework for correlating coupled sliding block displacement to realistic modes and volumes of failure in 2D homogeneous rock slopes due to strong ground motion. The framework captures three distinct behaviors for homogeneous rock slopes in response to strong ground motion: (1) damage through the initiation of new fractures but without any failure volume leading to a weakened post-seismic condition; (2) the development of shallow, highly disrupted rock falls; and (3) the development of deep-seated rock slumps.

Geotechnical engineers should be considered key members of any project design team. Regardless of the selection criteria, the design team should add value and aim to minimize risk and unnecessary expenses during the project pursuit. The uncertainty of ground conditions and soil performance can often lead to increased risk and unnecessary added expense. Geotechnical engineers can bring value to the design-build team by identifying critical geotechnical components, communicating risks, and providing solutions to improve design efficiency, which should reduce construction costs. Early engagement and proactive communication are important to sharing information and the identification of risk to the design build team during the pursuit. The Interstate 270 (I-270) North Design Build Project is an example of how engaged geotechnical engineers can help develop a winning proposal and increase the chances of team success. This paper makes a case for early geotechnical engineering project involvement, the need for adequate exploration and testing data, and for continued geotechnical engineering involvement during the design and construction phases.

Landslides are the most disastrous natural catastrophe, which causes immense damage to infrastructure and loss of life globally. The properties of soil slope are spatially variable due to various loading and deposition conditions. Hence, slope reliability analysis should take into account the spatial variation of slope material. Previous researchers utilize the random field theory to consider spatial variability in slope reliability analysis. However, this method requires extensive computational resources and time. To address this issue, the present study proposes a methodology based on convolution neural network (CNN) and Monte Carlo simulation. CNN algorithm is utilized as a surrogate model to replicate the random field model of slope as CNN algorithms efficiently learn the spatial variation of the random field. The CNN model is further used to conduct reliability analysis using Monte Carlo simulation. An example application of the proposed method is performed for spatially variable soil slope to validate the proposed method. The example results suggest that the proposed method provides practical values of the probability of slope failure.

The 2017 Mexico earthquake had a high impact on the banks of the old lake of Mexico Valley, in which the spectral acceleration was amplified by site effects. The case study explains the retrofitting of a pharmaceutical building complex in Xochimilco, Mexico City, developed between 2018 and 2019. The buildings have a problem of excessive deformability with relevant non-structural damages during seismic events. The plant is located in an abrupt variation zone on the banks of the old lake, where geotechnical and a geophysical survey identified a soil fracture due to regional subsidence, producing differential settlements, accelerated by the seismic activity. The retrofitting actions involved structural and geotechnical works, including a cut-off wall in order to reduce the flow of underground water, and soil improvement with rootpiles and micropiles. Deformability problems were solved by increasing the structure stiffness, and re-foundation with micropiles. Custom bentonite-cement mix was designed to guarantee the permeability values.

Extreme hydroclimatic events such as heavy rainfall and drought have occurred frequently in recent years, and their impacts on geotechnical systems must be understood for developing climate-adaptive design procedures. A coupled flow-deformation finite element method that incorporates the variation in stiffness properties and failure due to the change in the degree of saturation is best suited for accurately predicting the impacts. PLAXIS, a finite element software widely used by practicing engineers and researchers, models water flow accurately, but the constitutive model does not consider the change in stiffness and failure criterion due to water flow. In this study, the Mohr-Coulomb (MC) model was first modified by updating the modulus and the yield criterion based on the degree of saturation and/matric suction and then used within PLAXIS to analyze selected geotechnical problems under extreme hydroclimatic events. The modified MC (MMC) model for unsaturated soil was validated using the experimental results for different soils found in the literature. The implementation of the MMC model in PLAXIS 2D through a user-defined soil model (UDSM) was also verified. Finally, the new constitutive model and the existing flow model were used for analyzing selected geotechnical systems in a coupled manner to understand and quantify the impacts of extreme hydroclimatic events on them. The results indicated that extreme hydroclimatic events significantly impact the strength and deformation behavior of deep foundations and the stability of earth slopes.

In this study, the effect of extreme hydroclimatic events, including heavy rainfall and drought, on the behavior of drilled shaft was predicted using a numerical method and will be validated using the proposed experimental method. For numerical analysis, a new modified Mohr-Coulomb constitutive model, which updates the yield criterion and modulus with changes in the matric suction and/or degree of saturation, was used in the coupled geotechnical-hydrological finite element code PLAXIS 2D. First, a full-scale model with a 1.0 m diameter and 12 m long drilled shaft was modeled separately in the sand and silty sand, then subjected to a constant mechanical load and varying extreme hydrological loads. Then, the analysis was repeated for a small-scale model with a 76 mm diameter and 1.83 m long drilled shaft and compared with the results of the full-scale model. The results indicated significant changes in the bearing capacity of the full-scale drilled shaft in both soils due to hydrological loads. The materials and methods adopted for the experimental method to validate the small-scale model are discussed.

Climate change has led to more severe and frequent droughts, creating more fuel to sparkle more wildfires worldwide. Wildfires are now occurring more frequently, at higher severity, and higher elevations. Adverse impacts of wildfires accumulate by post-fire geohazards, such as shallow landslides that may occur due to fire-induced changes in soil and land cover. This study aims to develop an analytical framework to evaluate the stability of wildfire-stricken slopes subjected to heavy rainfalls. An analytical framework for transient infinite slope stability analysis of variably saturated soils is developed by employing a closed-form solution of the Richards equation to capture water flow in slopes while considering the transpiration effect. The proposed model captures the effects of initial soil conditions, time-varying rainfall intensity, and wildfire-induced changes in soil mechanical and hydraulic properties, root reinforcement, and transpiration rate. The model is used in a set of parametric studies to examine the effect of soil type, burn severity, slope angle, and soil thickness. It is demonstrated that fire could decrease the factor of safety to rainfall-triggered landslides by up to 20%. The proposed model uses measurable landscape characteristics and hydro-mechanical properties of soils and can be used in developing landslide hazard maps for wildfire-prone areas.

As wildfires become increasingly more prevalent in today’s world, the need for restoration and remediation efforts grows respectively. A fire event can cause hydrophobicity within the top few inches of the soil’s surface, and this hydrophobic layer when combined with the removal of exterior stabilizers such as root systems has the potential to increase runoff by more than ten times the average rate. To rapidly repair these burned environments and preserve any surrounding systems, restoration may be needed. Remedial studies of hydrophobic soils occur in the field or the laboratory with soil samples taken from the field after real fire events. This can make the research site-specific until the overall quantities of research and analyzed data reach a broad enough range to develop empirical correlations. This research outlines a standardized procedure for inducing hydrophobic properties of soil. By standardizing the process to produce hydrophobic specimens, the collection of nonuniform samples from the field can be avoided for fundamental research—before applied field research is pursued—and the degree of soil hydrophobicity can be adjusted to evaluate a range of potential field conditions. The procedure is critical for the development of a range of consistent degrees of hydrophobicity in soil that makes laboratory testing. Once the procedure for inducement of hydrophobicity is standardized and the repeatability is established, further outside research can be conducted in a more controlled, repeatable, and methodical manner. To standardize and develop hydrophobic properties that most closely resemble field conditions, specimens of Ottawa sand with varying hydrophobic strengths are prepared by analyzing and applying wood smoke condensates. Then infiltration tests are performed on these specimens to correlate the impact of hydrophobicity on infiltration through various altered Ottawa sand specimens.

Movement of wildfire ash into soil can alter soil hydrologic behavior and consequently influences the stability of burned hillslopes. There are no standard methods to quantify the concentration of ash in soil. This paper evaluates water vapor sorption as an index test for ash quantification. Soil and ash samples collected after the 2021 Green Ridge Fire, WA, were dry-mixed in mass-controlled ratios and the water vapor isotherms were measured. Ash gave smaller specific surface area than soil, but the maximum adsorbed water content of ash was greater than that of soil because of relative amounts of adsorbed and capillary water. The best-fit of various isotherm models was evaluated to represent the measured data. The results suggested that the Guggenheim-Anderson-de Boer (GAB) model well represents (R²~0.99) the isotherms of ash-soil mixtures and the GAB parameter (k_b) can be used to quantify ash in soil. However, the nonlinear increase in k_b with ash content suggested a stepwise calibration that changes at 25% ash content.