Geo-Risk 2023: Hazards and Climate Change

The Bureau of Overseas Buildings Operations (OBO) serves as the manager of the US Department of State’s properties abroad with a portfolio of over $82 billion in assets, spread over 290 missions across the globe. Its core mission is to provide safe, secure, functional, and resilient facilities that represent the US government to the host nation and support the Department’s achievement of US foreign policy objectives. Pluvial, fluvial, and coastal flooding has the potential to negatively affect many OBO overseas operations. Projected climate change is anticipated to increase exposure to flooding and other natural hazards at many Department of State missions, thereby escalating the need to modernize current risk management techniques. OBO’s Climate Security & Resilience Program is tasked with this and, as one of its responsibilities, works to guide site acquisitions away from flood exposure and facilitate flood risk adaptation in circumstances where flood exposure is unavoidable. For new sites where OBO plans property acquisition and development, maps from publicly available data sources indicating the extent of potential flood exposure and vulnerability are often unreliable or nonexistent, thereby necessitating the need for commissioned studies. In some urban areas, developing in floodplains is inevitable and requires data-informed adaptation measures to reduce long-term impacts in accordance with OBO design guidelines. Recent flood studies provided essential data on flood exposure that informed site acquisitions and allowed early planning for necessary adaptation measures. The completed flood studies included projected sea level conditions to ensure the flood risk is minimized throughout the project lifetime. New major projects in Bangkok, Thailand, and Hanoi, Vietnam, informed a variety of measures to prevent unavoidable flood exposure while also facing geotechnical challenges including excessive consolidation settlement and city-wide ground subsidence. These case studies highlight the importance of obtaining accurate projected flood exposure mapping at the onset of project planning, and early coordination of flood risk management strategies with geotechnical conditions.

Rain-induced landslides are the top one natural hazard in Hong Kong. A single severe rainstorm could trigger thousands of natural terrain landslides in a short time, which highlights the importance of prompt landslide forecasting. However, natural terrain landslides are commonly identified through post-event geological surveys or interpretation from remote sensing images, leaving the accurate failure time unknown. To address this challenge in storm-based landslide forecasting, a probabilistic “at least one failure” modeling strategy is proposed to link the storm-based landslide probability and year-based landslide observation in a unified machine learning framework. The performance of the machine learning model is validated against historical landslide records in 1984–2017 and compared with conventional statistical rainfall-landslide correlations. Results indicate that the machine learning model enables storm-based landslide forecasting with only the information on the year of landslides and outperforms the statistical correlations significantly, which can serve as a powerful decision-making tool for landslide early warning and risk management.

This study aims to develop a quantitative framework to optimize embedded footing performance subjected to extreme historical climate events with respect to the uncertainties associated with site-specific soil and climatic parameters. The proposed method was applied to a semi-arid climate site in Riverside, California, as a case study. By performing the optimization procedure, it is found that the proposed climate-adaptive design method provides a better optimal design for the embedded footing in comparison with the conventional method. The optimization result indicates that the footing width significantly drops from 4.72 m to 3.79 m for the design at the Riverside site, compared to the conventional method. Regarding the total construction cost, the optimal design from the proposed method for the Riverside site shows 19% lower cost, compared to the one from the conventional method.

Yarlung Tsangpo on the Tibetan Plateau is the highest large river in the world. Valley glaciers in the Yarlung Tsangpo Grand Canyon pose high risk to the infrastructure in the region. In particular, the Zelunglung glaciers at the entrance of the canyon are experiencing rapid movements in the warming climate. This study presents a comprehensive assessment of glacier detachment risks based on remote sensing and physically based numerical analysis. Long-term meteorological records and debris-ice rheological modeling are used to reveal the volume and thickness of the instable source mass. An erosion–deposition debris flow analysis simulation program is employed to analyze multi-scenario hazard dynamic processes to predict the run-out distance, velocity, and deposition distributions. The deposition distributions along the river course in different glacier detachment scenarios are quantified to assess the size of possible barrier dam, which lays a good foundation for a quantitative risk assessment.

Predicting landslide dam formation is essential in mitigating landslide risks in alpine valley regions. This study assesses the landslide damming probability with the consideration of landslide characteristics, valley topography, and hydrological factors using machine learning techniques. A landslide inventory is collected, including both damming landslides and non-damming landslides in the 2008 Wenchuan earthquake region and the Bailong River basin. Three machine learning algorithms are compared, including logistic regression, random forest, and support vector machine. Results show that machine learning techniques can well predict the landslide damming probability. The random forest model achieves the best prediction performance, followed by logistic regression and support vector machine. Among six learning features, landslide area, upstream watershed area, and valley floor width are the three most important variables for landslide dam formation. An illustration example of the Tangjiashan landslide dam is used to demonstrate how the developed model can be integrated to predict landslide dam formation.

Under the changing climate, extremely massive landslides, destructive avalanches, glacial debris flows, and dam-breaching outburst floods occur frequently on the Tibetan Plateau, posing severe threats to valley-crossing bridges of the railways and highways. Previous studies examined bridge-landslide interactions corresponding to separate geohazard scenarios. However, in reality, hazard type can transform from one to another. Considering generic physical mechanisms of bridge-landslide interactions under different geohazard types or evolutions of hazard chains, a unified vulnerability assessment framework is established in this study. The resultant impact forces in various geohazards are categorized into three components: boulder impact, coarse particle impact, and slurry impact. Four failure modes of the bridge are revealed under different combinations of impact forces. The integrated vulnerability of a bridge is calculated as the union failure probability of different failure modes. As an example, the proposed methodology framework is applied to assess the vulnerability of the Zhibai Bridge, which was damaged by a high mobility debris flow on September 10, 2020.

Since the Wenchuan earthquake in 2008, debris flows erupted in many gullies along the Minjiang River when subjected to heavy rainfall, leading to partial or complete river blockage and serious damage to infrastructure at gully outlets. Climate change drives the occurrence of extreme rainfall more frequently; thus it is necessary to evaluate the scale of debris flows in the Wenchuan area under extreme rainfall in the future. This study aims to evaluate the possible river blockage caused by debris flows under different rainfall scenarios. Taking Gaojia Gully as an example, a physically based model was used to simulate the whole process of debris flow including initiation, movement, and deposition under extreme rainfall. The deposition volume and the distribution of deposition depth in the outlet area are used to characterize the debris flows. Based on the deposition volume and the geometry of river reach, the risk of river blockage can be assessed. The outcome of this study provides practical support for the management of risks of potential barrier lakes and dam breaching floods.

Due to a combination of mountainous terrain, Mediterranean climate, and frequent wildfire event, post-fire hazards are inevitable in Southern California. Quantifying the post-fire consequences of a burn event provides the foundation for risk analysis and hazard mitigation planning. This study aims to simulate the hazards after a fire in Montecito using a triggering-propagation model. The model considers varying surface material conditions and erosion effects, which can simulate the initiation and propagation processes of debris flows. The scheme and the parameters are evaluated by the debris-flow event in Montecito following the Thomas Fire. The simulated debris-flows depths basically coincide with the observed ones, which indicate that the model is reliable to be used to predict the post-fire debris flows. Bed erosion mainly occurs on the surface bed of burned areas, where the erosion depths are much larger than those on unburned areas. Wildfires increase the magnitude of erosion. The bed-erosion materials travelling down from hillsides enlarge the debris-flow intensity significantly. Surface-erosion runoff processes are by far the most predominant contributors to debris flows. Though the materials provided by slope failures are much smaller than that from erodible surface beds, the slope failure process still augments erosion entrainment. The entrainment from surface beds significantly increases when there is a sharp increase of slope failures. Once the materials are entrained by surface runoff, the flows gain mass quickly and descend along steep slopes and channels.

Subseasonal to seasonal (S2S) precipitation forecasts encompassing lead times ranging from two weeks to three months can bridge the gap between weather and seasonal forecasts. For practical applications, calibration is a necessary step to improve predictive performances of raw forecasts from S2S models. This paper illustrates the calibration of the S2S precipitation forecasts by a case study of the European Centre for Medium-Range Weather Forecasts. The Bernoulli-Gamma-Gaussian model and quantile mapping are used to calibrate raw S2S forecasts. The results of three catchments in China show that raw forecasts exhibit reasonable correlations with observed precipitation but suffer from considerable biases and unreliable spreads, leading to negative skill. The Bernoulli-Gamma-Gaussian model overall outperforms the quantile mapping in correcting biases and improving reliability and skill. For S2S precipitation, the Bernoulli-Gamma-Gaussian model can serve as an effective tool to generate skillful ensemble time series forecasts for the early warning of rainfall-induced geohazards.

The Thornthwaite Moisture Index (TMI) is commonly used by geotechnical engineers to model soil moisture-climate interaction within the vadose zone. The long-term average (20+ years) TMI has correlated well with unsaturated soil moisture flow parameters but struggles to characterize sporadic and extreme climate events crucial to the performance of civil infrastructure. Short-term TMI, or precipitation data directly, better captures seasonal climate variability and extreme events, although the nonstationary nature of these parameters can hinder their usefulness in civil design. This study presents a Bayesian forecast framework for estimating the seasonal TMI using time-series analysis techniques with a component-wise transitional Markov Chain Monte Carlo (MCMC) approach. The proposed framework is presented using an example study site in San Antonio, Texas, from using 30 years of prior climate date to forecast 10 years of the seasonal TMI from June 2012 to June 2022. Validation and performance studies of the example forecast demonstrate the effectiveness of the proposed framework to capturing seasonal variability and extreme climate events.

With the continuous development of the global urbanization process, the underground space transportation networks of cities in various countries or regions have shown different scales. The underground transportation network connects all corners of the cities, which greatly improves the efficiency of commuting within the cities. However, under the background of gradually intensifying climate change, frequent occurrence of geological disasters, frequent terrorist attacks, and old and degraded engineering facilities, the urban metro network is vulnerable to a certain extent, and the failure of any link in the network may cause significant reduction in global network efficiency. Using the complex network theory, this paper takes three cities, New York City in the United States, Shanghai and Jinan in China as examples, representing well-developed, developed, and developing cities respectively, and compares and analyzes the vulnerability of metro networks of different scales to network efficiency in the face of emergencies. Then, considering the ridership and the length of the tunnel, taking extreme rainfall as an example, this paper analyzes the changes in network efficiency after the metro networks in Shanghai and Jinan were flooded. Finally, taking Jinan metro as an example, this paper analyzes the impact of entrance-exit lines being flooded on the entire metro network and points out that the entrance-exit lines are the key points of flood control that cannot be ignored in the metro network.

Climate change has induced more frequent geo-hazards in the mountainous area, especially in the Tibetan Plateau. A basin named Zelunglung located in southeast Tibet near Namcha Barwa Peak exhibited the footprint of climate change in terms of a recent large, long traveling glacial debris flow initiated from a debris-ice mixture slope. A slope failure occurred on the moraine deposit on September 10, 2020, which transformed into a glacial debris flow, marching downwards along the channel for about 10 km. The debris flow entrained moraines along the channel, which enlarged the magnitude of the debris flow in return. The debris flow rushed out the gully, damaged a steel bridge, and formed a deposition fan wider than 200 m. The solid deposition material of the debris flow changed the morphology of the watercourse and partially blocked the Yarlung Tsangpo River. The meteorological recordings showed that there was little precipitation before and on the day of the event, which brings out the question of how this debris flow traveled so far without precipitation water supply. To study this glacial debris flow’s transport mechanism and high mobility, multi-phase numerical experiments were carried out based on an advanced open-source simulation tool, r.avaflow. The benchmark test was determined with an optimized parameter set to back-calculate the debris flow. Different initiation conditions including the water content and released volume were applied to better understand the mechanism of the debris flow. The result shows that even a small change in initial water content can lead to different flow behavior, which means the initial conditions are vital for the accuracy of the back-calculation of the glacial debris flow. Besides, the entrainment of saturated glacial till that remains in the channel is critical for the high mobility of the glacial debris flow, which prolonged the travel distance to up to 10 km.

The present study investigates the potential utility of a capillary barrier system that limits water infiltration into the underlying soil. This was achieved by developing a cylindrical apparatus capable of monitoring matric suction and water content at different depths in real-time when subjected to hydraulic loading. The climate adaptive barrier layer (CABL) was prepared by using an amended soil containing a 5% by dry mass of waste produced from the water treatment plant. The result of monitoring without a CABL indicated the quick reduction in suction and quick increment in the volumetric water content under artificially induced rainfall. The top layer of the soil has shown the variation of suction up to 2,000 kPa under continuous air-drying (four months) where the suction was higher during the day compared to the night. Such variation was not observed under the application of the CABL, thereby limiting the changes in the water content and soil suction of the underlying soil, showing the potential applicability of the CABL to limit the impact of water content fluctuation.

On the Tibetan Plateau, landslide risk is increasing under changing climate. The deposits of moraine and glacier mixture (MGM) provide an abundant source for glacial debris flow hazard chain, posing great threats to society and the fragile ecological environment of the surrounding areas. On September 10, 2020, an MGM landslide occurred in Zelunglung basin, located on the right bank of Yarlung Tsangpo River (YTR) in Milin County, Nyingchi, Tibet, and triggered a glacial debris flow. The glacial debris flow traveled along the channel for about 10 km before rushing into YTR, and the debris flow deposits after the event narrowed the river by about 100 m, damaging a steel bridge at the basin outlet and forming a potential threat to residents. To better assess and manage such risks, it is of great importance and urgency to understand failure mechanisms of MGM landslides. In this study, through post-event investigations with remote sensing interpretation, meteorological observations, and thermal-hydraulic-mechanical (THM) coupling numerical model with ice-water phase change, the impact of freeze-thaw cycle due to temperature change on the stability of the MGM deposits is analyzed. The results show that freeze-thaw cycle and associated phase change plays an important role in the evolution process of the stability of the MGM deposits. The phase change affects the seepage field and pore pressure, and further modifies the effective stress and mechanical field. It indicates that the stability of MGM deposits decreases with the increasing number of freeze-thaw cycles, but failure is not likely triggered directly and solely by freeze-thaw cycles. This study would be a typical benchmark to understand the initiation mechanism of MGM deposit failure that might generate large glacial debris flow hazard chain under changing climate on the Tibetan Plateau.

Desiccation cracking in clayey soils can cause significant issues to slopes and earthen structures. Environmental and climatic variables such as temperature and relative humidity are among the key factors controlling the formation and propagation of soil desiccation cracks. Spatial patterns of soil desiccation cracking inherently pose a non-deterministic and complex problem. The main objective of this study is to investigate the statistical distribution of geometrical properties of soil desiccation cracks at different levels of relative humidity (RH). For this purpose, we employed an image processing technique to probabilistically analyze the spatial patterns in an extensive set of soil cracking images reported in the literature showing the dynamics of surface cracks in a fat clay tested at six different RH levels ranging from 15% to 93%. Specifically, we determined the probability density function for several geometrical properties of cracking, including the average length and width of the crack, the crack intersection angle, and the crack area. The results show that RH level strongly influences the geometrical attributes of cracks. Elevated levels of the applied RH are found to increase the most probable value (MPV), representing the maximum value of distribution probability for each geometrical property. This observation implies that elevated RH increases the likelihood of forming longer and wider desiccation cracks. In all cases examined, the crack intersection angles are primarily distributed from 80 to 100 degrees with an MPV of about 90 degrees.

In 2019, we conducted a study in Tri-Valley, California, to identify potential failure zones (PFZ) along 12.4 mi of unlined flood control channels located within Zone 7 Water Agency’s jurisdiction. PFZs were calculated using a multi-linear regression with a suite of geospatial variables from publicly available data and proprietary model results. The statistics suggested that there are two controlling variables that lead to historical failure of the channel embankments: Clear Lake clay soil type and historical willow thicket ecology. The channels are underlain by these two variables in 38.6% of the length of channel, but account for 62.7% of the failures in the study area. Since completion of the study, 60% of the newly reported failures in the study area fall within the boundaries of the defined PFZs and are underlain by Clear Lake clays and historical willow thickets. This approach to identify PFZs is broadly applicable to asset management through identification of maintenance priorities and planning for future likely repairs.

Climate change has major implications on the durability and performance of a pavement. The smoothness of pavement in freeze climate zone is affected by seasonal frost heave and thaw weakening, which are subjected to the effects of climate change. This paper proposed a new empirical International Roughness Index (IRI) model, which captures seasonal environmental influence on road sections. The model was nationally calibrated for flexible pavements deal with seasonal freezing/thawing via data from the Long-Term Pavement Performance (LTTP). The calibrated new model is suitable for analyzing the climatic impacts of frost action to flexible pavement. The future climate conditions are predicted based on different climate scenarios. The new IRI model is combined with the climate forecast to analyze the effects of climate change on the pavement. Case studies for three sites at different climatic regions were conducted, from which the regional climate adaptation is suggested. The analyses provide guidance on planning and adaptation for flexible pavement with respect to climate change.

This paper presents the probabilistic seismic hazard assessment (PSHA) of Red Sea state in the eastern region of Sudan, located adjacent to the Red Sea gulf and in the most active earthquake zone in Sudan. This region has significant economic importance, not just for the country, but for the entire East African region. As such, appropriate consideration for the seismic risk on structures is required. Data on seismic events and sources were collected, and an updated catalog was compiled from different agencies. A tectonic map of the region was prepared, and based on the spatial distribution of earthquakes, we divided the area into three seismic source zones. As regional ground motion prediction equations (GMPEs) are unavailable, not unlike in many parts of Africa, two GMPEs suitable for our study area were used with a logic tree framework. Hazard maps for variation of the PGA and Sa at 0.1 s and 1 s are shown. The PGA for a 50-year return period considering 10% and 2% probabilities of exceedance varies from 0.286 g to 0.4 g, respectively, and the highest hazard was seen near the border with Eritrea. The results will be invaluable in risk mitigation and future regional infrastructure development planning.

Ground motion prediction models (GMPM) are essential for probabilistic seismic hazard analyses. We developed a GMPM for spectral accelerations up to a period of 10 s for the central zone of Chile, specifically for the region from Los Vilos to San Antonio (latitudes −32° to −33.77°), the busiest industrial area in the country. For the development of the model, seismic data published by the National Seismological Center (CSN) were added to a published database for the Chilean strong ground motions. New measurements of the average shear-wave velocity for the upper 30 m (Vs30) made at stations in the segment are included. Orientation-independent intensities are computed, reducing epistemic uncertainty in the model development phase. The developed model is a better fit to the seismic intensities for central Chile than global and national models, with lower total standard deviation at the expense of a constraint area of applicability. The importance of studying the seismic segmentation in the Chilean subduction zone is emphasized when observing different behaviors depending on the latitude or the segment along the margin. We developed similar GMPMs for the area north of −32° and south of −33.7° using the original database. For the same subduction zone, there is a heterogeneity in the source characteristics, which generates different maximum earthquake sizes, slab slopes, and wave paths, all of which cause different seismic intensities. This observation should lead, in the future, to the development of zone-dependent seismic design codes, which means an advance in seismic hazard preparedness.

The Atlantic and Gulf Coastal Plains (CPs) in the eastern, southeastern, and southern US consist of deep unconsolidated sediments overlying a high-velocity bedrock, which can significantly affect the amplitude and duration of earthquake ground motions in the region. Evaluating seismic hazards in the US CPs requires an in-depth understanding of how both geology and sediment thickness influence seismic wave propagation in the region. Existing site amplification and seismic hazard models for the Atlantic and Gulf CPs are limited in their characterization and modeling of these factors. Our project seeks to develop a characterization of site response in the CPs and incorporate it into probabilistic seismic hazard estimations for the region. In this paper, we describe our preliminary one-dimensional site response analysis (SRA) simulation results, which include using a new sediment thickness model, randomization of shear wave velocity (V_S) profiles, and a reference velocity at the base of the soil columns of interest equal to 3,000 m/s. Our preliminary site response model is compared to other available amplification models by focusing on their implied characterization of amplification factors as a function of sediment thickness and ground shaking intensity. We find that at low frequencies (i.e., 0.5 Hz, or 2 s period) our site amplifications are in good agreement with V_S30-based (the average shear-wave velocity in the top 30 m) NGA-East models for thinner sediment thicknesses, but this is not the case for larger sediment thicknesses (>100 m). Also at low frequencies, our model is in good agreement with existing amplification models in the region that directly account for sediment thickness. At higher frequencies (i.e., 10 Hz, or 0.1 s period) our preliminary model provides a similar trend to existing sediment-thickness-dependent models, but it does not show the same decrease in amplification for larger sediment thicknesses. Additional comparisons at specific locations of interest, such as Charleston, South Carolina, and Memphis, Tennessee, reveal that our model provides an upper bound of amplifications at low frequencies for younger and thicker deposits, and a lower bound at high frequencies. The latter trend may be explained by the assumed seismic attenuation quality factor (Q) properties in our model.

Earthquake source-to-site distance affects the attenuation of seismic waves in their trajectory, and thus, is a key parameter in the estimation of ground motion (GM). The distance between major slip zones on the fault and the surface represents one of the most common physical concepts that is not usually incorporated into GMPE path term due to the difficulty of identifying these zones. Recent studies have shown that there is a correlation between areas of high interseismic locking and high coseismic slip, information that can be used to infer distances to asperities. Here, we propose a source-to-site distance metric (R_asp) that takes into account an a priori size of potential asperities based on locking models to estimate the slip distribution in the rupture plane. We propose a simple modification to an existing GMPE, which needs an interseismic locking model as input. The metric we propose is aimed at large megathrust earthquakes and was tested in Japan subduction zone. This metric can be used for new ground motion prediction equations (GMPE) but also to improve the results given by existing GMPEs. We find that R_asp produces an improvement in the prediction of ground motion intensities for periods under 1 s. Our results show a reduction in standard deviation and mean of intra-residuals, which could be relevant to deterministic and probabilistic seismic hazard applications.

The current state of the practice in probabilistic seismic hazard analyses (PSHA) relies on ergodic ground motion models (GMMs), which assume that the ground motion variability observed in a global database is the same as the variability at a single site-source combination. However, as ground motion databases have grown, it has become clear that there are significant systematic differences in ground motion for a combination of sites and seismic sources in a particular region. These differences do not support the ergodic assumption, prompting the transition from ergodic to non-ergodic GMMs. In this study, we use the Ridgecrest ground motion database to evaluate the performance of ergodic and non-ergodic GMMs. The large database in Ridgecrest allows for quantifying repeatable effects within a small region, which has seldom been attempted in previous efforts. The spatial variations of the repeatable site, source, and path effects in the non-ergodic GMMs are modeled using Gaussian processes and the cell-specific attenuation approach. Our results show that the non-ergodic GMMs can capture the spatial variation of repeatable effects; they have a lower aleatory variability and a better performance for the considered earthquake scenarios in the Ridgecrest area; and the repeatable effects quantified in small-magnitude earthquakes improve the estimation of IMs at larger magnitudes. These observations suggest that non-ergodic GMMs can have a significant impact on the results of probabilistic seismic hazard analyses.

In seismic hazard analysis (SHA) usually, the prediction of seismic intensity measures (IM) at a specific site is obtained using ground motion prediction equations (GMPEs). GMPEs are empirical models based on observation of past earthquakes that return the distribution, median, and standard deviation of the predicted IM values in the frequency band of engineering interest as a function of magnitude, source-site distance, and site conditions. Large earthquakes (LE) are infrequent, so recordings are scarce for events in the near-source region (e.g., 30–70 km). The physics-based numerical simulation (PBS) for LE has become necessary for a better understanding of the ground-motion characteristics and as a complement to recorded data. The work shows progress through the PBS using both deterministic and stochastic methods for the problem recognized in LE along the north Chilean megathrust. In the kinematic approach, we used the finite elements method (FEM) to evaluate the low-frequency range. For the high-frequency range, the stochastic method is used; finally, both responses are included in single broadband ground-motion time histories for a number of selected stations; in a future work, the results will be validated with event registers available and a bias model.

The current state of the practice in probabilistic seismic hazard analysis (PSHA) employs ergodic ground motion models (GMMs) to describe the probabilistic distribution of an intensity measure, which assumes that the ground motion variability observed in a global database is the same as the variability in ground motion at a single site-source combination. However, the fast-growing empirical ground motion databases indicate significant regional differences in ground motions due to repeatable and systematic source, path, and site effects. These systematic effects, which are spatially correlated, are not consistent with the ergodic assumption, promoting the transition to non-ergodic GMMs for PSHA. In this study, we use Gaussian processes to model the spatial correlation structures of systematic source effects for the Ridgecrest area, which provides insights into non-ergodic PSHA. We consider two different models for the spatial correlation of the systematic and repeatable source effects for the peak ground acceleration: an isotropic stationary model and an anisotropic non-stationary model. In the isotropic stationary model, the spatial correlation of the source effects depends only on the separating distances of two seismic sources, which is commonly used in current non-ergodic GMMs. The new anisotropic non-stationary model proposed in this study considers the fault geometries of the earthquakes and the separating distance of the seismic sources. Our results show that the spatial correlation model provides information on the spatial variation of the repeatable source effects induced by complex physical processes and reduces the associated aleatory variability. We also find that the spatial correlation of systematic source effects in the Ridgecrest area is best characterized by the anisotropic non-stationary model, which ensures better extrapolation of the source effects to regions without significant data. The developed spatial correlation model provides insights in the context of nonergodic-based PSHA considering multiple seismic sources. In addition, the proposed model improves the epistemic uncertainty treatment of alternative spatial correlation structures for repeatable source effects.

Pacific Gas & Electric (PG&E) has guided the development of a framework to evaluate permanent ground deformation (PGD) from seismically induced landslides within the San Francisco Bay Area (SFBA) for system-wide risk evaluation. The framework was developed through a combination of geologic maps with limited strength information, a landslide inventory database, and a regional groundwater model that reflects both wet and dry hydrological conditions. The model incorporates a novel application of the multimodal landslide framework, where failures due to shallow disrupted landslides, deep coherent landslides, and rockfalls are evaluated over multiple length scales using a convolutional window. We assess the pseudostatic yield coefficient for each of the various landslide modes with dry and wet seasonal groundwater conditions, landslide geometry, and material properties. We then estimate PGD and landslide run-out considering the aleatory uncertainties in material properties and develop a simple tool to evaluate the probability of landslide deformation and run-out impact for a range of ground motion intensities.

In this paper, the spatial distribution of the liquefaction probability inside the studied dam was calculated from CPTUs including conversion errors and the spatial variability of the soil parameters. In addition, based on the seismic hazard at the studied site and the fragility, the spatial average of the liquefaction probability over the next 50 years was evaluated. Although the CPTU is often preferable for identifying the liquefaction probability, the CPTU is still not widely used in Japan. Therefore, the focus will be placed here on the SPT N-value and the fines content, F_c, because the majority of the accumulated site characterization data in Japan consists of SPT N-values and F_c. Utilizing the conditional simulation, the spatial distribution of the expected values for F_c and N_SPT inside the dam was calculated with high spatial resolution. Based on the results of the simulation the scale of the weak areas and the spatial continuity of the soil profile of the dam were clarified.

The Liquefaction Experiments and Analysis Project (LEAP) is an international effort to produce high-quality experimental data and then use it for a validation exercise of existing computational models and simulation procedures for soil liquefaction analysis. Nine centrifuge teams from across the world performed 24 centrifuge tests on a physical model of a sloping sand deposit. A novel discrepancy metric was used to quantify variabilities in phase, amplitude of dominant frequency components. The effect of variabilities in the achieved input motions on the recorded soil responses were quantified using a sensitivity analysis. The analysis revealed that the effects of variabilities in phase on the recorded soil responses were minor. The variabilities in soil responses were mainly driven by variabilities in higher order frequency components. The discrepancy metric and sensitivity analysis are combined to obtain a performance score for LEAP 2017 experiments.

Liquefaction-induced settlement is one of the main geotechnical issues that may cause failure in the equipment (e.g., circuit breakers, transformers, liquid reactors, etc.) at the substations within a powerline system after earthquakes, and subsequently cause the lack of power supply in the area of service and the economic loss to the clients and the owner of the powerline system. In this paper, a simplified approach is presented to develop the hazard maps that quantify the risk of exceeding the intolerable liquefaction-induced settlement at 174 substations within a power transmission and distribution system in Portland, Oregon, under a magnitude (Mw) 9.0 Cascadia Subduction Zone (CSZ) earthquake scenario. The substations were categorized based on shear wave velocity measurements and geologic information. Probabilistic and deterministic seismic hazard analyses were performed to estimate the ground shaking intensities and the corresponding annual frequency of occurrence at each substation. Based on that, probabilistic liquefaction assessment was conducted to evaluate the liquefaction potential and estimate the liquefaction-induced settlement and the probability of exceeding the threshold value of the liquefaction-induced settlement. The results will be used in the structural, substation fragility analysis and risk assessment as part of the strategic asset management (SAM) study to prioritize the seismic risk reduction projects that will also provide optimum return on investment from the maintenance and capital spending perspectives.

The 2020 Liquefaction Experiments and Analysis Projects (LEAP) included 23 centrifuge tests of a saturated Ottawa sand deposit supported by a cantilever (rigid) sheet-pile quay wall. In this paper, a Gaussian process (GP) regression approach is used to (1) assess the sensitivity of the system response to variability in soil and input motion conditions; and (2) qualify the statistical characteristics of the experimental variability. The conducted analyses showed that the LEAP 2020 experimental results are consistent and shed light on a number of salient characteristics of the soil-structure interaction and lateral spreading during liquefaction.

Pacific Gas & Electric (PG&E) has significant assets in the San Francisco Bay Area (SFBA), with broad areas adjacent to the Bay that are susceptible to liquefaction and permanent ground deformation (PGD) during an earthquake. To facilitate system-wide risk evaluations, we have developed a liquefaction model to predict liquefaction ground surface manifestation, settlement, and lateral spreading deformation within a roughly 9,000 km² study area. We extend site-specific liquefaction analyses to the entire project area through a hybrid probabilistic finite difference groundwater model in conjunction with deposit scale statistics derived from over 4,000 geotechnical borings and 17,000 well records. The resulting liquefaction tool may be used to estimate the probability of liquefaction triggering and PGD based on input peak ground acceleration (PGA) and earthquake magnitude. Comparisons to historic liquefaction observations indicate that the model reasonably matches qualitative observations from previous earthquakes and improves upon existing models by (1) incorporating a larger subsurface database, (2) implementing a spatially continuous groundwater model, (3) updated liquefaction triggering models, and (4) providing estimates of PGD.

Liquefaction-induced reconsolidation settlements occur as earthquake-generated excess pore pressures dissipate and can lead to significant damage to overlying infrastructure as well as buried structures. Design of earthquake resilient infrastructure in areas with high liquefaction susceptibility requires a methodology to predict these settlements for a variety of soil types and boundary conditions. The state-of-practice empirical models that are commonly used to compute settlements exhibit several limitations that might affect the accuracy of predicted settlement, including an inability to include the effects of partial drainage, non-liquefiable crust, thin layers, and soil fabric. Numerical models can overcome some of these limitations but require proper calibration and validation. Centrifuge and shaking table experiments can be used for this validation, but multiple challenges arise when comparing numerical simulations and experimental results for liquefaction-induced settlements. This paper describes the need for soil-specific calibration of reconsolidation behavior which is one of the key challenges in accurate prediction of settlements. Multiple experiments considering free-field conditions are simulated using the numerical platform FLAC and the constitutive model PM4Sand. The bias in predicted settlement using the default PM4Sand reconsolidation formulation is assessed. A new, tentative relationship for the PM4Sand reconsolidation parameter, f_sed,min, is proposed by incorporating the mean grain diameter, D₅₀. Settlement results from simulations using the default and modified reconsolidation schemes are compared to the experiments to assess the level of agreement between them.

Histories of earthquakes in Nepal have shown substantial liquefaction in several regions of the nation, including the capital, Kathmandu Valley. Despite being a liquefaction sensitive zone, Kathmandu Valley soil in Nepal has little studies on liquefaction potential. As a heavily populated metropolitan area, the evaluation of the valley’s liquefaction potential is vital for assuring the safety of engineered structures. We utilised the geotechnical database collected from 400 locations to conduct quantitative analyses and modelling to evaluate the valley soil’s susceptibility of liquefaction in two scenario earthquakes. Some of the most widely used and existent methods based on deterministic and probabilistic approaches are used for the assessments, and the results are shown as liquefaction hazard maps indicating liquefaction potential index (LPI) and probability of liquefaction (P_L). Most of the valley’s areas are moderately to extremely prone to liquefaction, and the valley’s central and southern regions are more susceptible to liquefaction and at greater risk of liquefaction damage than the northern sections. Results of the evaluation are corroborated by the field manifestations of the 2015 Gorkha Earthquake.

As part of its dam safety program, Grant County Public Utility District No. 2 (GCPUD) is undertaking a risk-informed evaluation of the Left Embankment at Wanapum Dam. The Federal Energy Regulatory Commission (FERC)−regulated facility was constructed in the 1960s, and concerns were raised regarding its liquefaction potential and seismic vulnerability. In what is believed to be a first of its kind application to an engineering problem, a seismic fragility analysis of the Left Embankment is being conducted as a Senior Seismic Hazard Analysis Committee (SSHAC) Level 3 (SL3) study. The SSHAC process was developed specifically to address complex technical problems where experts are needed to evaluate sources of epistemic uncertainty in data, models, and methods, and develop the inputs to analyses that are to be performed. This approach is considered more thorough, and includes improved safeguards to reduce bias, than those that are typically applied in dam safety risk evaluations. The SSHAC process is a structured approach to the planning, organizing, executing, and documenting expert elicitation studies. It includes an explicit evaluation of uncertainties and their effect on the characterization and subsequent analyses; in this case, the estimate of the seismic fragility of an embankment dam and overall seismic risk.