SEARCH · Geology Reports
Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Humpback whales produce complex, structured songs that vary among individuals, populations, and through time. This study presents a novel method of studying changes in humpback whale song elements through time while accounting for detection probability.
Geochronological and thermochronological data are essential for constraining the timing and rates of geological processes, supporting geologic mapping, natural hazard assessment, and resource exploration. The U.S. Geological Survey (USGS) Geochron Database is a centralized, relational database that integrates USGS and State geological survey data in accordance with the National Geologic Mapping Act of 1992 and its reauthorizations. This report documents the structure and development of the database, including its standardized schema, controlled vocabularies, and protocols for compiling legacy and newly published data. The database is designed to adhere to FAIR (Findable, Accessible, Interoperable, and Reusable) data principles and currently (2026) includes data from a wide range of geochronological and thermochronological methods. Public access is provided through the USGS Geochron Database Explorer, a browser-based geographic information system (GIS) interface, and through versioned data releases available on ScienceBase, which provide sample-level summary data and detailed analytical information. The USGS Geochron Database is structured to be extensible, supporting the inclusion of additional methods and data types as they are compiled. This format ensures long-term utility and aligns with the Mapping Act’s directive to create a national archive of geochronological information that adds interpretive value to geologic map data. Its design reflects a commitment to data transparency, scientific reproducibility, and national geoscience priorities.
In-situ chemical oxidation (ISCO) is a common method to remediate chlorinated ethene contaminants in groundwater. Monitoring the effectiveness of ISCO can be hindered because of insufficient observations to assess oxidant delivery. Advantageously, potassium permanganate, one type of oxidant, provides the opportunity to use its strong electrical signal as a surrogate to track oxidant delivery using time-series borehole geophysical methods, like electromagnetic (EM) induction logging. Here we report a passive ISCO (P-ISCO) experiment, using potassium permanganate cylinders emplaced in boreholes, at a chlorinated ethene contamination site, Naval Air Station Pensacola, Florida. The contaminants are found primarily at the base of a shallow sandy aquifer in contact with an underlying silty-clay confining bed. We used results of the time-series borehole logging collected between 2017 and 2022 in 4 monitoring wells to track oxidant delivery. The EM-induction logs from the monitoring wells showed an increase in EM response primarily along the contact, likely from pooling of the oxidant, during P-ISCO treatment in 2021. Interestingly, concurrent natural gamma-ray (NGR) logging showed a decrease in NGR response at 3 of the 4 wells possibly from the formation of manganese precipitates coating sediments. The coupling of time-series logging and well-chemistry data allowed for an improved assessment of passive ISCO treatment effectiveness.
The glacial deposits underlying southeastern Michigan, northeastern Indiana, and northwestern Ohio are a substantial source of water to communities, agriculture, and industry in the region. Previous efforts to characterize aquifer materials in the area cited a need for additional information about the underlying hydrogeologic characteristics and related groundwater availability as well as improved mapping of the extent and properties of the glacial deposits. Recent U.S. Geological Survey multi-State compilations of water-well drilling records have greatly increased access to high-resolution geologic data, particularly in glacial depositional environments. This study by the U.S. Geological Survey, in cooperation with the Ohio Environmental Protection Agency, uses processed data from the State-managed collections of well records to characterize the glacial deposits in the study area using two methods. The first method creates two-dimensional maps of basic hydrogeologic information commonly required for assessments of groundwater availability, including (1) total thickness of glacial deposits, (2) total thickness of coarse-grained deposits, (3) specific-capacity-based transmissivity and hydraulic conductivity, and (4) texture-based estimated equivalent horizontal and vertical hydraulic conductivity and transmissivity. The second method builds a hydrogeologic framework of the complex glacial aquifer through construction of a volumetric geologic model by using three-dimensional kriging. Results of the volumetric model indicate that aquifer materials are primarily concentrated in the western parts of the study area near the Indiana-Ohio border. Coarse-grained sediments are also present as surficial deposits in the north of the study area where intermixing glacial advances created complex distributions of unconsolidated deposits. Two-dimensional maps of hydrogeologic properties support the volumetric model, showing thicknesses of coarse-grained deposits that reach up to 250 feet in the western sections of the study area and progressively thin to near absence in the east. Visualization of the aquifer materials with a volumetric model generally shows a highly discontinuous distribution of coarse- and fine-grained materials, with no clearly defined boundaries to delineate the extent of the aquifer. Comparisons of cross sections derived from the volumetric model with existing published maps support previous near-surface hydrogeologic interpretations while filling gaps where data are sparse, particularly in deeper parts of the aquifer. Both the two-dimensional maps and the volumetric model provide data that can directly inform assessments of groundwater availability, in addition to having future applications to studies of groundwater flow and transport.
Directivity, or the focusing of energy along the direction of an earthquake rupture, is a common property of earthquakes of all sizes and can cause increased hazard due to azimuthally dependent ground‐motion amplification. For small earthquakes, the effects of directivity are generally less pronounced due to reduced rupture size, yet the directivity in small events can bias source property estimates and provide important insights into general regional faulting patterns. However, due to observational limitations, directivity is usually only measured and modeled for large events. As such, many studies of small earthquakes either ignore directivity altogether or assume a constant rupture direction for all events in a cluster. In our study, we apply a refined directivity fitting method constrained with two separate methods of source deconvolution to the dataset of aftershocks of the 2019 Ridgecrest earthquakes, which contain a large number of well‐recorded small‐to‐mid sized earthquakes occurring in close proximity to each other. The revealed directivity of 100+ small (M 2.4–5) earthquakes is highly heterogeneous and primarily oblique to and away from the main fault strike, suggesting a complex postseismic stress redistribution. In addition, the energy focusing effect of directivity appears to bias the selection of high‐quality data from stations in the direction of rupture, leading to average stress‐drop increases of 50% if directivity is not accounted for.
Seismic waves generated by large, rapid landslides encode information about the source and can be analyzed rapidly following an event. Even remote landslides can pose hazards to downstream communities, so rapid detection and characterization using existing seismic monitoring networks could be beneficial. In this study, we expand on past regionally limited work by presenting a globally applicable method for estimating landslide volume from seismic features that could be integrated into future landslide seismic monitoring frameworks. We train the model using multivariable linear regression and five seismic features derived from recordings of 129 landslide events of a range of styles and locations with independently estimated volumes. We present two preferred models, one that combines long‐period (LP) and high‐frequency (HF) features and one for use on smaller landslides without observed long‐period signals. We find that our best‐performing model, applicable to landslides larger than 100,000 m 3 with signals containing observable long‐period energy, requires only two features: LP (20–100 s) absolute maximum amplitude and HF (1–5 Hz) rise time (time between the signal start and the maximum envelope amplitude) and has an R 2 score of 0.79. This model predicted volumes within one order of magnitude for 55 out of 58 events. We find that this combination mitigates the trade‐off between mass and acceleration and other variations of landslide style that limit methods based on amplitude alone.
This report presents results of a water-quality study for the Pine Ridge Indian Reservation, South Dakota. The study was a cooperative effort between the U.S. Geological Survey and the Water Resources Department of the Oglala Sioux Tribe. Discharge and water-quality data were collected during 1992-97 for 14 contact springs located in the northwestern part of the Reservation. Data were collected to evaluate potential alternative sources of water supply for the village of Red Shirt, which currently obtains water of marginal quality from a well completed in the Inyan Kara aquifer. During 1995-97, water-quality data also were collected for 44 public-supply wells that serve about one-half of the Reservation's population. Quality-assurance sampling was used to evaluate the precision and accuracy of environmental samples. Ten of the springs sampled contact the White River Group, and four contact the Pierre Shale. Springs contacting the White River Group range from calcium bicarbonate to sodium bicarbonate water types. Two springs contacting the Pierre Shale have water types similar to this; however, sulfate is the dominant anion for the other two springs. In general, springs contacting the White River Group are shown to have better potential as alternative sources of water supply for the village of Red Shirt than springs contacting the Pierre Shale. Nine of the springs with better water quality were sampled repeatedly; however, only minor variability in water quality was identified. Six of these nine springs, of which five contact the White River Group, probably have the best potential for use as water supplies. Discharge from any of these six springs probably would provide adequate water supply for Red Shirt during most periods, based on a limited number of discharge measurements collected. Concentrations of lead exceeded the U.S. Environmental Protection Agency (USEPA) action level of 15 µg/L for three of these six springs. Five of these six springs also had arsenic concentrations that exceeded 10 µg/L, which could be problematic if the current maximum contaminant level (MCL) is lowered. Blending of water from one or more springs with water from the existing Inyan Kara well may be an option to address concerns regarding both quantity and quality of existing and potential sources. All nine springs that were sampled for indicator bacteria had positive detections on one or more occasions during presumptive tests. Although USEPA standards for bacteria apply only to public-water supplies, local residents using spring water for domestic purposes need to be aware of the potential health risks associated with consuming untreated water. One spring contacting the White River Group and two springs contacting the Pierre Shale exceeded 15 pCi/L for gross alpha; these values do not necessarily constitute exceedances of the MCL, which excludes radioactivity contributed by uranium and radon. Additional sampling using different analysis techniques would be needed to conclusively determine if any samples exceeded this MCL. Nine springs were sampled for selected pesticides and tritium. The pesticides atrazine, carbaryl, and 2,4-D were not detected in any of the samples. The nine springs were analyzed for tritium in order to generally assess the age of the water and to determine if concentrations exceeded the MCL established for gross beta-particle activity. Tritium results indicated two springs are composed primarily of water recharged prior to atmospheric testing of nuclear bombs and two other springs have a relatively large percentage of test-era water. The remaining five springs had tritium values that indicated some percentage of test-era water; however, additional sampling would be needed to determine whether water is predominantly pre- or post-bomb age. Of the 44 public-supply wells sampled, 42 are completed in the Arikaree aquifer, one is completed in an alluvial aquifer, and one is completed in the Inyan Kara aquifer. Water from the alluvial well is a sodium bicarbonate water type, water from Arikaree aquifer ranges from calcium bicarbonate to sodium bicarbonate types, and water from the Inyan Kara well is a calcium sulfate bicarbonate type. Of the 44 wells sampled, 28 (64 percent) tested positive for indicator bacteria in presumptive tests. Because these were single samples that generally were collected upstream from chemical treatment feeders, positive detections do not necessarily constitute exceedances of drinking-water standards. A single sample from an Arikaree well exceeded the MCL for arsenic of 50 µg/L. Arsenic exceeded 10 µg/L for six additional Arikaree wells and for the alluvial well and the Inyan Kara well, which could be problematic if the current MCL is lowered. The alluvial well also exceeded the secondary maximum contaminant level (SMCL) for dissolved solids, which is non-enforceable, and the action level for lead. The Inyan Kara well exceeded the SMCL's for iron and for manganese and the MCL of 5 pCi/L for radium-226 and 228 combined. Several Arikaree wells exceeded SMCL's for either pH, sulfate, dissolved solids, iron, or manganese. One Arikaree well exceeded the MCL of 4.0 mg/L for fluoride and another exceeded the MCL of 10 mg/L for nitrite plus nitrate. Ten Arikaree wells equalled or exceeded 15 pCi/L for gross alpha; however, these values do not necessarily constitute exceedances of the MCL, which excludes radioactivity contributed by uranium and radon. Additional sampling using different analysis techniques would be needed to conclusively determine if any samples exceeded this MCL. Eight wells, all from the Arikaree aquifer, equalled or exceeded the proposed MCL of 20 µg/L for uranium and 33 wells (75 percent) equalled or exceeded one-half of the proposed MCL. Although this standard has only been proposed, additional information regarding the extent of elevated uranium concentrations in the Arikaree aquifer, and the geochemical processes involved, may be beneficial. It was determined from analyses of uranium isotope data for five wells that the source of elevated uranium concentrations is naturally occurring, rather than anthropogenic.
Semantic mapping in 3D is fundamental to a wide range of geoscientific studies and applications, including geomorphology, hazard assessment, and environmental monitoring. However, automatically segmenting geological features from large-scale photogrammetric datasets remains a significant challenge. We present a methodology to address this gap. Using overlapping images collected over environments of interest, Structure-from-Motion (SfM) produces georeferenced point clouds and estimates camera poses. Existing large vision models, such as Segment Anything Model, segment objects in the images, generating pixel-segmentation associations. To produce pixel-point associations, we project the points back onto the camera image planes. As objects are independently segmented across multiple images with different perspectives, we develop a segmentation mosaicking algorithm to build probabilistic point-segmentation associations that combines the pixel-segmentation associations and pixel-point associations. Our methodology is validated using both synthetic data generated by Kubric and real-world UAV-SfM data. The implementation is designed to be compatible with existing SfM software, including Agisoft and OpenDroneMap, for photogrammetry mapping in geoscience studies. As a case study, we apply our method to the semantic mapping of precariously balanced rocks (PBRs), which provide upper-bound constraints on historical ground motion shaking intensity. To support object-level identification of PBRs, we additionally integrated Grounding DINO, enabling text-prompted segmentation of features of interest within UAV imagery. This case study demonstrates the effectiveness of our method in generating a 3D semantic map of PBRs, enabling spatial distribution of PBR fragility for earthquake hazard analysis.
Omori’s law states that the rate of aftershocks decays as a function of inverse time. There are multiple physical explanations that we reduce into a nonlinear mixed effects relation of three terms: (1) a Rate/State expression that can account for static/dynamic and viscoelastic triggering caused directly by the mainshock, (2) a fluid diffusion triggering term, and (3) a randomized secondary triggering (cascade) term. We fit free physical-model parameters to an observed aftershock sequence through two nonlinear regression methods to find the relative contributions of physics-based models in an observed aftershock sequence. Results from both methods show that Rate/State models overpredict aftershock rates by ∼0–30%. Secondary aftershocks cause a net negative contribution (seismicity rate reduction that corrects overprediction by other terms) ranging between ∼0 and 30%. All regression solutions yield negative secondary triggering contributions without being guided to do so. A physical explanation for this is that aftershock occurrence relieves stress from the crust, ultimately causing the sequence to extinguish itself. Fluid diffusion triggering contributions range from ∼0 to 20%. Diffusion processes are observed to be shorter in time than the full duration of an aftershock sequence and they are also spatially limited, diminishing their influence. Our results apply to an aftershock decay curve from the 2016 Central Apennines earthquake sequence, meaning that our specific results may not be general. Our primary conclusion is that any one physical model cannot alone fit the observed sequence as well as the combination of three we investigated.
The ability to accurately map erosion, flooding, and habitat loss in coastal environments is crucial for formulating national strategies aimed at preventing and mitigating the impacts of natural disasters. A fundamental component of this process is the implementation of coastal morphodynamics monitoring through Structure-from-Motion (SfM) techniques, utilizing high-resolution 2D/3D data obtained from aerial photogrammetry. To assess morphodynamic changes over a three-year period (2022 – 2024), several SfM-based photogrammetric studies were conducted, each year, in the Romanian sector of the Danube-Black Sea coastal zone, specifically at the mouth of one of the Danube River distributaries (Sf Gheorghe branch) into the Black Sea, and along the left bank, near Sf Gheorghe locality, located within the Danube Delta Biosphere Reserve (DDBR). The essential equipment for aerial photogrammetry comprises Unmanned Aerial Vehicles (UAVs) and Global Navigation Satellite Systems (GNSS). In this study, the UAV used was a DJI Mavic 3T (Enterprise/Thermal) drone, complemented by two Trimble R12i and R4 GNSS systems, as well as approximately 10 Ground Control Points (GCPs). Data acquisition and processing were carried out using specialized photogrammetric software (Agisoft Metashape) along with various GIS tools (e.g., Blue Marble Geographics Global Mapper and ESRI ArcMap). The photogrammetric products generated for the study, as detailed in this paper, include Digital Elevation Models (DEMs), Digital Terrain Models (DTMs), orthomosaics (orthophotos), and others. At Sfântu Gheorghe beach, a comparison between 2023 and 2024 photogrammetric surveys revealed that the left bank of the Sf. Gheorghe Arm, at the river mouth into the Black Sea, suffered from a twist (erosion) of up to 64 metres. Additionally, on the selected perimetre (total area of 31,910 square meters ) from the beach and dune zone of Sf. Gheorghe, an area of up to 16,202 square meters was eroded between 2023 and 2024. This contrasts with the period between 2022 and 2023, during which deposition predominated. Erosion at the Danube mouths and the adjacent Black Sea coastline is driven by a complex interaction of natural and anthropogenic factors. Natural processes, including subsidence, sea-level rise, and episodic extreme storm events, contribute significantly to coastal dynamics. Meanwhile, human-induced factors, such as upstream hydrotechnical works that limits sediment transport, cutting of navigation canals, as well as the exacerbating effects of climate change, further accelerate erosion. The recent Structure-from-Motion (SfM) surveys provide essential quantitative data, enabling a detailed analysis of both short-term and long-term morphodynamic changes influenced by seasonal variations and extreme hydrometeorological events in this highly dynamic coastal system.
Correlations between grade and tonnage exist in mineral resource data compiled from published reports, but they are not always addressed during quantitative assessment of undiscovered mineral resources. Failure to account for correlated grade and tonnage distributions can result in geologically unrealistic assessment results. Current software tools simulate univariate ore tonnage and multivariate resource grades of undiscovered deposits independently. As a result, analysts are forced to rely on ad-hoc solutions to minimize the correlation issues by: 1) creating subsets of data with restricted criteria; 2) truncating grade and tonnage distributions; and 3) testing model robustness using exploratory data analysis. While these methods represent pragmatic solutions, the statistical solutions presented here provide additional options to address real correlations in grade and tonnage data used for mineral resource assessments. We present a modified version of the MapMark4 package in R that introduces two alternatives for modeling grade and tonnage distributions, consisting of a multivariate solution that accounts for correlations between ore tonnage and metal grades and an empirical solution that utilizes simple random sampling with replacement to reproduce coupled grades and tonnages from the input data. We present simulations for contained ore and metal for three case studies representing tungsten skarn, komatiite-hosted nickel, and sediment-hosted carbonate amagmatic zinc-lead (Mississippi Valley-type) deposits. Employing the methods presented here yields quantitative mineral resource assessment results that more closely reflect the empirical distributions of grades and tonnages observed in nature and expands the applicability of these tools for ongoing critical mineral resource assessments.
We analyze aftershocks of the 2019 M 7.1 Ridgecrest mainshock and isolated M 5–6 mainshocks in southern California to test predictions made by the rate‐and‐state friction model of Dieterich (1994) . Rate‐and‐state friction predicts that the seismicity rate after a stress step follows Omori decay, where the Omori c ‐value, which is the saturation in aftershock rate observed at small times, is larger for smaller stress steps. Put in the context of an aftershock sequence, this predicts that the Omori c ‐value will be systematically larger at greater distances from the mainshock. To our knowledge, this predicted effect has not been observed. In part this may be because the Omori c ‐value is difficult to measure because it often reflects short‐term catalog incompleteness rather than a true saturation in aftershock rate. We explore the dependence of the Omori c ‐value on the distance to the mainshock by applying the “ a ‐positive” method ( van der Elst and Page, 2023 ). This method is insensitive to short‐term aftershock incompleteness and allows resolution of the true aftershock rate deep into the mainshock coda. For aftershocks of the Ridgecrest mainshock and stacked M 5–6 mainshocks, we observe systematic differences in early aftershock rates, relative to mainshock distance, consistent with the predictions of rate‐and‐state friction. Furthermore, for the larger Ridgecrest dataset, we observe that aftershocks nearer to the mainshock start earlier, and we resolve a flattening of the Omori curve consistent with a larger Omori c ‐value for the farthest aftershocks, as predicted by Dieterich (1994) .
Earthquake emergency response and loss modeling require accurate models of ground‐motion intensity metrics (IMs), such as those produced by ShakeMap. In the critical minutes and hours after an earthquake, however, knowledge of the earthquake source geometry limits the accuracy of these models, especially in areas where seismic instrumentation is insufficient to constrain the shaking intensity pattern. Previous attempts to overcome this limitation have centered on computing the mean distance metrics from a suite of assumed source geometries. That approach is inadequate with modern ground‐motion models (GMMs) because they typically include terms for the hanging wall effect, depth to top of rupture, and other nonlinear terms. Such approaches also propagate the distance uncertainty through the GMMs and may not properly allocate the additional uncertainty into the between‐ and within‐event components. Here we present Finite‐Fault SIMulation with Mixed Effects Regression (FFSIMMER), a method to address these limitations by computing IM distributions from a suite of rupture realizations that fully accounts for all geometry‐related GMM terms and allows for the separation of the between‐ and within‐event components of the total uncertainty. The accuracy of this method is limited by the available constraints on the source parameters. To address this, FFSIMMER can incorporate constraints based on knowledge of the source, such as the approximate strike and dip based on knowledge of the plate tectonics and focal mechanism of the earthquake. We demonstrate the improvement in accuracy that can be achieved with the 2025 M 7.7 Mandalay, Burma (Myanmar), earthquake by incorporating constraints on the range of plausible source parameters.
The US National Seismic Hazard Model (NSHM) was updated in 2023 for all 50 states using new science on seismicity, fault ruptures, ground motions, and probabilistic techniques to produce a standard of practice for public policy and other engineering applications (defined for return periods greater than ∼475 or less than ∼10,000 years). Changes in 2023 time-independent seismic hazard (both increases and decreases compared to previous NSHMs) are substantial because the new model considers more data and updated earthquake rupture forecasts and ground-motion components. In developing the 2023 model, we tried to apply best available or applicable science based on advice of co-authors, more than 50 reviewers, and hundreds of hazard scientists and end-users, who attended public workshops and provided technical inputs. The hazard assessment incorporates new catalogs, declustering algorithms, gridded seismicity models, magnitude-scaling equations, fault-based structural and deformation models, multi-fault earthquake rupture forecast models, semi-empirical and simulation-based ground-motion models, and site amplification models conditioned on shear-wave velocities of the upper 30 m of soil and deeper sedimentary basin structures. Seismic hazard calculations yield hazard curves at hundreds of thousands of sites, ground-motion maps, uniform-hazard response spectra, and disaggregations developed for pseudo-spectral accelerations at 21 oscillator periods and two peak parameters, Modified Mercalli Intensity, and 8 site classes required by building codes and other public policy applications. Tests show the new model is consistent with past ShakeMap intensity observations. Sensitivity and uncertainty assessments ensure resulting ground motions are compatible with known hazard information and highlight the range and causes of variability in ground motions. We produce several impact products including building seismic design criteria, intensity maps, planning scenarios, and engineering risk assessments showing the potential physical and social impacts. These applications provide a basis for assessing, planning, and mitigating the effects of future earthquakes.
We present the 2023 U.S. National Seismic Hazard Model (NSHM) for all 50 states that applies new smoothed seismicity, fault rupture, and ground motion models. New data and methods are introduced in the 2023 earthquake rupture forecast that include: new earthquake catalogs - excluding induced earthquakes, alternative declustering methods, spatially smoothed seismicity distributions, full-catalog scaled rates to account for aftershocks, updated CEUS-WUS attenuation boundary, new magnitude-scaling equations, new geodetic and geologic deformation models, and alternative fault system solutions accounting for a more complete representation of epistemic uncertainty potential for earthquake generation in Alaska, Hawaii, and the conterminous U.S. Improved ground motion models consider new Next Generation Attenuation NGASubduction, modified NGA-East, and adjustments to account for regional biases in ground shaking observations. Semi-empirical and 3D simulations of ground motion are applied to account for shaking at 21 oscillator periods, 2 peak motions, and 8 site conditions. Site effects models are constructed for western U.S. basins (Seattle, Portland/Tualatin, San Francisco, Central Valley of California, Los Angeles, and Salt Lake City) and for sites with deep sedimentary wedges found across the central and eastern U.S. Gulf Coast and Atlantic coastal plain regions. These models result in substantial changes compared to the older NSHMs and are differentiated for the earthquake rupture forecast and ground motion model changes to display sensitivities and impacts.
We present a method to update the geospatial liquefaction model used by the U.S. Geological Survey’s near‐real‐time ground failure product with subsurface geotechnical data. The geospatial model estimates liquefaction probability from peak ground velocity (via ShakeMap) and geospatial susceptibility proxies. In many regions, additional information relevant to constraining liquefaction likelihood is also available, including surface geology maps and subsurface geotechnical measurements. There is currently no mechanism to use these data in the ground failure product liquefaction model, even though these data could provide more precise constraints on spatial variations in the lithologic character of the soil (surface geology) and direct measurements of the subsurface mechanical properties that affect liquefaction occurrence and severity (geotechnical measurements). In this study, we develop a method to integrate these data with the geospatial model and assess how these data can improve regional‐scale predictions. We develop a Bayesian updating framework and apply it to the 1989 magnitude 6.9 Loma Prieta, California, earthquake, for which mapped observations are available to evaluate performance. We constrain the Bayesian framework with 373 Northern California cone penetration tests and liquefaction susceptibility classes based on the mapped surface geology. This Bayesian model incorporates geotechnical information into the geospatial model and more accurately predicts liquefaction occurrences than the geospatial model, while sacrificing less accuracy in terms of predicting the absence of liquefaction than the geotechnical model. In future applications, this approach could be adapted to update other geospatial models using locally available subsurface data.
Because of similarities between locating an earthquake with seismic stations and locating a Global Positioning System (GPS) receiver from satellites, the Bancroft algorithm developed for GPS processing can be used to locate earthquakes. Such an approach to earthquake location differs from the conventional method of choosing an initial or trial solution and then iteratively improving the solution until convergence. The Bancroft algorithm has the advantage of being a direct, noniterative solution but with the disadvantage of only being able to accommodate a homogeneous velocity model. An additional limitation of the standard Bancroft algorithm is that it considers arrival times in a medium with a single propagation velocity. This poses no problem for GPS processing because electromagnetic waves travel at the speed of light; however, for seismic waves it means the algorithm can be applied to collections of either P ‐ or S ‐wave arrival times. Here, I show how the Bancroft algorithm can be generalized to handle both P ‐ and S ‐wave arrival‐time measurements simultaneously. I also show how to accommodate depth‐varying P ‐ and S ‐wave velocity models. I apply the generalized Bancroft algorithm to microearthquakes beneath Tanaga Volcano in Alaska and compare standard locations from the widely used HYPOINVERSE location code to Bancroft locations and to the output of HYPOINVERSE when setting the trial location to the Bancroft location. I find the Bancroft locations outperform the results from the other methods for shallow earthquakes near sea level, where a quantity known as the geometric dilution of precision is large and linearized approaches such as HYPOINVERSE are expected to struggle.
Model criticism, discrimination, and selection methods often rely on calibrated model outputs. Because calibration can be computationally expensive, model criticism can first be undertaken by assessing model outputs obtained from limited prior parameter ensembles. However, such prior-based methods are often heuristic and do not formalize the notion of balancing model consistency with data and model complexity (i.e., model adequacy). We present a new framework to discriminate among candidate models prior to calibration that formalizes prior-to-calibration model adequacy into a metric to implicitly balance prior model output data coverage with model complexity represented by prior output (co)variance. The prior model adequacy metric “Mahalanobis distance deviation” quantifies the deviation of (a) the set of squared Mahalanobis distances of data from a prior model output distribution from (b) the set of squared Mahalanobis distances of data from their own distribution. A new data worth metric “discernment value” is also presented which quantifies the value of data for screening less-adequate models prior to calibration. Discernment value is calculated from the change in variance of a weighted average of prior model outputs from all candidate models due to less-adequate model outputs receiving lower weight. The framework is demonstrated using a one-dimensional groundwater flow model with eight possible configurations. A synthetic data network is used to test the framework. Results show the framework identifies the candidate models most similar to the true model used to create the synthetic data. Discernment values show variation in the value of different data types and locations for screening less-adequate models.