Adrian Rodriguez-Marek

During three decades of experience in successfully implementing the SSHAC guidelines for seismic hazard analyses, numerous refinements have been made to the procedures, especially to clarify the process for the lower study levels and simultaneously enhance the relevance of Level 1, 2, and 3 SSHAC studies. Lessons recently learned from practical applications, which have yet to be incorporated into formal guidelines, are discussed, particularly with regard to the selection of project participants and to the structuring of their interactions during the course of the PSHA project. The most important insights, however, are related to technical issues, through which ever more efficient and transparent procedures for representing the center, body, and range of model inputs have evolved. The extensive experience and technical achievements gained from conducting many SSHAC Level 3 studies provide the foundational footing for the improvements to the SSHAC process and advancements in underlying PSHA data, models, and methods proposed in the paper. These technical advances and practical experience gained have resulted in mature and widely accepted frameworks for seismic source characterization, ground‐motion modeling, and site response analysis, enabling recent studies to focus on project‐specific data, interpretations, and uncertainty characterization rather than on developing new methodological approaches. This paper summarizes the technical frameworks that have emerged from recent SSHAC studies and presents additional procedural recommendations that will allow future studies to focus their resources on site‐specific data gathering and characterization rather than on redevelopment of the underlying hazard methodologies. These proposed recommendations are intended to initiate discussions with a view to inform potential future improvements to regulatory guidance for siting critical facilities that will be needed in order to fully realize the benefits of the lessons learned from past SSHAC studies, especially as they can be adapted to applications for siting of the next generation of advanced reactors.

Expanding road and railway networks in developing countries is essential for economic growth and reducing regional disparities. However, such projects often face significant geotechnical stability and risk management challenges. This study presents a three-phase optimization framework consisting of: (1) deterministic, probabilistic, and risk analyses; (2) Random Variable (RV)-based optimization; and (3) Random Field (RF)-based refinement. Soil variability is modeled using both RV and RF approaches. Two Artificial Neural Network (ANN) surrogate models are employed to improve computational efficiency by predicting failure probabilities and associated costs. Two application cases are presented to demonstrate the integrated framework for optimizing earth slopes in transportation infrastructure using Deterministic Design Optimization (DDO), Reliability-Based Design Optimization (RBDO), and Risk Optimization (RO). The application cases further demonstrate the model’s potential for early-stage roadway design under limited geotechnical data. Overall, the framework supports more reliable, cost-effective, and sustainable infrastructure development.

We investigate the use of microtremor horizontal-to-vertical spectral ratio (mHVSR) measurements to predict the time-averaged shear-wave velocity in the upper 30 m (VS30) using data-driven machine learning models. We develop a global dataset comprising 2861 three-component ambient noise recordings made at 536 sites. We consider two different types of predictive models: low-dimensional models, which use mHVSR features such as the fundamental site frequency (f0,HVSR) and amplitude (A0,HVSR), and high-dimensional models, which use the entire lognormal median mHVSR curve to predict VS30. In addition, we integrate two topographic features: binned elevation, as a proxy for geologic composition, and topographic position index, as a proxy for topography features. The low-dimensional models are shown to reasonably approximate VS30 when using mHVSR and the topographic features (coefficient of determination, R2, up to 0.69 on the testing set). The high-dimensional models outperform the low-dimensional models (R2 of 0.82 on the testing set) regardless of whether the lognormal median mHVSR curve is supplemented by the topographic features. In addition, we show that our top-performing model outperforms a recently developed model to predict VS30 from remote sensing data at sites from our testing set. Finally, we apply our models to predict VS30 from mHVSR computed at 1855 permanent broadband stations. We compare our data-driven VS30 predictions at a subset of 156 stations with measured VS30 values and find a reduction in the recommended model’s predictive power (R2 of 0.18). The reduction in performance at permanent broadband stations is shown to be related in part to differences in instrument type and data distribution between the temporary and permanent broadband datasets. Therefore, the proposed models are expected to perform best when mHVSR is measured using temporary deployments of broadband sensors. All the data-driven models developed in this study have been made publicly available.

An indispensable component of any site‐specific probabilistic seismic hazard analysis (PSHA) is incorporating local site effects through site response analyses, rather than relying on generic amplification factors in ground‐motion models (GMMs). Traditionally, the seismic hazard is calculated at a buried rock horizon and then convolved with amplification factors for the overlying layers, which requires the definition and characterization of the buried rock profile (without which neither the model for rock motions nor the site response analyses can be developed). An alternative approach is to retain the host reference rock profile in the GMM for the hazard calculations and then perform the host‐to‐target site response adjustment for the full profiles in a single step. This approach is illustrated by its application to a site‐specific Senior Seismic Hazard Analysis Committee Level 3 PSHA for Idaho National Laboratory. The site adjustment factors are calculated using a logic‐tree approach that captures the full epistemic uncertainty in these factors. This is not necessarily achieved when the logic tree only captures uncertainty in the near‐surface shear‐wave velocity profile. The application illustrates the use of borehole and noninvasive shear‐wave velocity measurements, in conjunction with local recordings of weak earthquake motions, to constrain the logic‐tree branches and their associated weights.

Proper incorporation of shallow site effects in hybrid broadband ground-motion simulation is essential to improving predictions at soil sites. This paper validates and compares four methods to account for these effects in the predominantly linear regime, using 1446 ground motions from 213 small-magnitude earthquakes ( 3 . 5 ≤ M W ≤ 5 . 0 ) recorded at 38 sites in New Zealand representative of a wide range of soil conditions. These methods require V S 30 or a shear-wave velocity profile, with the second approach relying on either the square-root-impedance method or the theoretical one-dimensional (SH1D) transfer function. Incorporating shallow site effects with any of the considered methods significantly improves accuracy and precision for several intensity measures, including pseudo-spectral acceleration at short vibration periods. Although the reduction in prediction variability was comparable between methods, the SH1D-based approach showed particular benefits in the case of (1) stiff sites for which the simulation overestimates attenuation effects; (2) relatively stiff sites with high-frequency resonances; (3) relatively soft sites with strong resonances in the intermediate frequency range; and (4) sites with a near-surface predominant frequency lower than the frequency range considered for the application of the V S 30 -based adjustment. This study shows that the potential prediction improvements obtained by using more advanced site-adjustment methods that incorporate additional site-characterization data are highly dependent on the site characteristics and the quality of the ground-motion simulation in a given region. These findings can help to inform site response method selection in forward applications.

Over a thousand moment magnitude (Mw) 3 or greater earthquakes and nine Mw≥4.8 earthquakes in Oklahoma were attributed to induced seismicity between 2009 and 2021. The 2016 Mw 5.8 Pawnee, Oklahoma, earthquake was the first induced seismic event worldwide, that the authors are aware of, in which liquefaction was documented, raising concerns regarding liquefaction during future induced earthquakes. Because of the potential differences in ground-motion characteristics and regional geology, the applicability of commonly used “simplified” stress-based liquefaction triggering models developed for tectonic earthquakes in active shallow-crustal tectonic regimes (e.g., California) is uncertain for assessing liquefaction due to induced earthquakes in Oklahoma, Texas, and Kansas (OTK). To address these issues, a new liquefaction triggering model specific to fracking and deep wastewater injection-induced seismicity in OTK is developed. The new model incorporates regional stress reduction factor (rd) and magnitude-scaling factor (MSF) relationships that are specific to induced seismicity and can be used within the context of the stress-based “simplified” liquefaction triggering model. A comparison of the factors of safety against liquefaction triggering (FSliq) computed using the proposed model versus existing models for tectonic events shows that existing tectonic models overpredict the liquefaction potential for induced earthquakes.

With the increasing recognition of the importance of nonergodic components in seismic hazard analysis, it is becoming increasingly important to obtain site-specific amplification functions (AFs). The most common approach to obtain site-specific AFs is to perform site response analyses backed by detailed site characterization with measures of uncertainty. An alternative approach would be to utilize ground motions recorded at the site, if any, to constrain site effects. But this approach is only viable if ground-motion recordings are available, which they generally are not. However, in a seismically active region (such as California), the rate of occurrence of small-magnitude earthquakes (i.e., M<4.5) is sufficient for weak ground motions to be recorded using a temporarily deployed station, especially for the design of critical structures. This then raises the question: can weak ground motions resulting from small-magnitude events recorded using temporary seismometers be used to constrain site effects? This study explores the possibility of employing a Bayesian inversion approach for constraining site terms using weak ground motions. Toward this end, ground-motion recordings associated with two dam sites in northern California were used in a Bayesian inversion scheme to constrain source, path, and site terms with measures of uncertainty. The empirical site transfer functions (TFs) at both sites were compared with analytical site TFs computed using 1D site response analyses. At both sites, the empirical site TFs were observed to be similar to the analytical site TFs. The standard error of the TF estimate suggests that the results are stable when sufficient recordings are available. This work serves to highlight that Bayesian inversions can be used to recover site terms and outlines a procedure for how the results should be interpreted post hoc.

Shallow site effects are indirectly considered in conventional hybrid broadband ground‐motion simulations, and their proper incorporation may be key to improving ground‐motion predictions at soil sites. This article presents and examines five methods to adjust hybrid simulations to account for these effects. These methods use different approaches to modeling site response and require different amounts of site‐characterization data: Methods 1 and 2 use only proxy parameters (e.g. , ) to describe the site conditions, with Method 1 relying solely on proposed site response scaling factors in existing semi‐empirical ground‐motion models, and Method 2 incorporating a host‐to‐target velocity‐profile adjustment; Methods 3 and 4 use a shear‐wave velocity profile along with two different frequency‐domain approaches to predict the linear site response, coupled with the nonlinear component of Method 1; and Method 5 uses 1D time‐domain nonlinear site‐response analysis and generally requires additional data to constrain nonlinear constitutive‐model input parameters. The five methods are applied to four sites subjected to two levels of ground‐motion intensity to illustrate the challenges involved in their implementation and to compare and contrast the resulting adjustments in terms of site amplification. The suitability and performance of each method will depend on multiple factors, including the availability of site‐characterization data and complexity of the site, and hence, their relative advantages and disadvantages are discussed considering a broad range of scenarios.