Geology Reports⌕ Search

SEARCH · Geology Reports

Results for “Journal of Volcanology and Geothermal Research”

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.

1,690 records · Page 17Linked to original sources
USGS2025

Subduction zone earthquake catalog separation tool: Implementation in the USGS 2025 Puerto Rico and U.S. Virgin Islands National Seismic Hazard Model

The U.S. Geological Survey (USGS) periodically releases updates to National Seismic Hazard Model (NSHM) for the United States and its territories leveraging current scientific knowledge and methodologies to guide public policy, building codes, and risk assessments regarding potential ground shaking due to earthquakes that may result in infrastructure damage. In subduction zones, there is a need to separate the earthquake catalog into tectonic regimes to create specific seismicity models for which the most appropriate ground‐motion models are then applied. Here, we describe newly developed methods and software, called CatSep, that classifies subduction zone events into three primary tectonic regimes: crustal, interface, and intraslab. This method incorporates information about the location of the earthquake relative to the subducting slab, the depth of the Mohorovičić discontinuity, and the earthquake’s moment tensor. Applying this method is a first step in the NSHM workflow for regions covering U.S. subduction zones. Results using this subduction zone earthquake catalog separation tool for the 2025 Puerto Rico and U.S. Virgin Islands NSHM earthquake catalog are presented and analyzed.

Puerto Rico, U.S. Virgin Islands↗
USGS2025

High-pass corner frequency selection and review tool for use in ground-motion processing

Raw seismological waveform data contain noise from the instrument’s surroundings and the instrument itself that can dominate recordings at low and high frequencies. To use these data in ground‐motion modeling, the effects of noise on the signals must be reduced and the signals’ usable frequency range identified. We present automated procedures to efficiently reduce low‐frequency noise that are implemented in the software package gmprocess. These procedures check for, and as needed remove, low‐frequency artifacts in the displacement record using polynomial fits, which can be used in combination with existing signal‐to‐noise ratio (SNR)‐based corner‐frequency selection procedures. The automated selections are then efficiently verified and refined using a graphical user interface (GUI) that plots relevant ground‐motion time series and spectra and tracks modifications to signal processing parameters. We demonstrate these procedures using recordings from the 2020 M 5.1 Sparta, North Carolina, and the 2013 M 4.7 southern Ontario earthquakes. Data processed with the SNR‐only and polynomial criteria for these events contain displacement artifacts in 37% and 23% of processed traces, respectively. Records with remaining artifacts are corrected manually using the GUI. These processing steps illustrate the workflow for efficient data processing with quality control.

Seismological Research Letters↗
USGS2025

Haunted Summerville: Ghostly lights or earthquake lights?

Among the colorful local lore in the Charleston, South Carolina, area, are a number of ghost stories, shared not only over campfires but also in published books. Among the most well-known of the stories is the tale of the Summerville Light. Local lore holds that a strange light sometimes seen in a remote area is a lantern carried by the ghost of a local woman who once waited hours for her husband, who turned out to have been decapitated earlier that day in a train accident (DePoppe, 2023). Extant sources suggest the ghost stories began to circulate in the 1950s to 1960s. So pervasive was the lore that (Old) Sheep Island Road became known among local residents as Light Road, with a local stretch of road known today as Old Light Road. Reviewing the location where the lights appear as well as the nature of accounts, I suggest that many if not all of the anecdotal observations can be most readily attributed to natural phenomena, including earthquake lights from earthquakes that were too small to be felt. Accounts of lights near Summerville cluster in proximity to the generally accepted epicenter of the 1886 Charleston, South Carolina, earthquake, where foreshocks to the 1886 mainshock were apparently concentrated, and within a few kilometers of three M3.5 – 4.4 earthquakes in 1959 and 1960.

South Carolina↗
USGS2025

Pre-eruptive characteristics of “suspect” silicic magmas in Carlin-type Au-forming systems

World-class Carlin-type Au deposits hosted in sedimentary rock were formed when profuse Eocene silicic magmatism swept across northern Nevada in response to arc migration. Carlin-type Au deposits formed along with porphyry/skarn Cu-Mo-W-Au deposits, epithermal Ag-Au deposits, and distal disseminated Ag-Au deposits. But unlike these other Au-bearing deposits that have clear associations with igneous intrusions, Carlin-type ore deposits appear to have formed distant from concealed plutons, and their origin remains controversial. Despite decades of abundant geophysical, geochronological, and geochemical studies suggesting the involvement of magmas, concrete evidence for magmatic involvement is still lacking. Consequently, the involvement of contemporaneous igneous systems remains inferred based on age, proximity, and variable isotopic, geochemical, and geophysical clues. A recent synthesis of deposit models postulates that Carlin-type Au deposits are intrusion-related, but that the causative magmas reside deeper (∼6–12 km) than in typical porphyry and peripheral systems (∼3–5 km), meaning that Carlin-type deposits are perhaps more distal expressions of igneous intrusions. We investigate a collection of “suspect” magmatic systems over a ∼7 m.y. timespan (∼41–34 Ma) that are contemporaneous with and near known Carlin-type ore deposits. We report results of a multifaceted array of in situ geochemical analyses (FTIR, EMP, SHRIMP-RG, LA-ICP-MS) of quartz-hosted melt inclusions, biotite, and quartz to better characterize the pre-eruptive characteristics of these magmas. We also report results of thermobarometry and thermodynamic phase equilibria modeling to help place constraints on magmatic reservoir depths and processes. Rather than a single “flavor” of silicic magma, we observe a surprisingly broad compositional spectrum of rhyolites, with one end of the spectrum exhibiting more arc-like (I-type) characteristics and the other end displaying more post-subduction, thick-crust extensional (A-type) characteristics. This broad compositional spectrum suggests a more complex picture of silicic crustal magmatism operating over a narrow span of time during slab rollback. Despite this spectrum, magmatic systems in this study are consistently ferroan and generally peraluminous, which we interpret as an expression of the relatively elevated geotherm at the time and incorporation of variable amounts of highly peraluminous metasedimentary crustal components. The silicic magma spectrum encompasses a range of mineralization associations, including subduction-related Cu-Mo-W-Au-Ag and post-subduction, thick-crust extensional rare-metal Mo-Sn-W-F-Be-Ag-Au, consistent with the prolific and diverse array of ore deposits that formed during this time. Carlin-type Au deposition appears to be associated with nearly the entire magmatic spectrum. This apparent indifference to silicic magma “flavor” would seem to imply that if magmas are involved in Carlin-type Au deposit genesis, they perhaps do not need to be compositionally specialized and/or possibly are only relevant as heat sources driving circulation to remobilize and redistribute metals.

Nevada↗
USGS2025

Over, under, and through: Hydrologic connectivity and the future of coastal landscape salinization

Seawater intrusion (SWI) affects coastal landscapes worldwide. Here we describe the hydrologic pathways through which SWI occurs - over land via storm surge or tidal flooding, under land via groundwater transport, and through watersheds via natural and artificial surface water channels—and how human modifications to those pathways alter patterns of SWI. We present an approach to advance understanding of spatiotemporal patterns of salinization that integrates these hydrologic pathways, their interactions, and how humans modify them. We use examples across the East Coast of the United States that exemplify mechanisms of salinization that have been reported around the planet to illustrate how hydrologic connectivity and human modifications alter patterns of SWI. Finally, we suggest a path for advancing SWI science that includes (a) deploying standardized and well-distributed sensor networks at local to global scales that intentionally track SWI fronts, (b) employing remote sensing and geospatial imaging techniques targeted at integrating above and belowground patterns of SWI, and (c) continuing to develop data analysis and model-data fusion techniques to measure the extent, understand the effects, and predict the future of coastal salinization.

Water Resources Research↗
USGS2025

Late Pleistocene kinematics of the Great Southern Puerto Rico Fault Zone, Puerto Rico

Several onshore faults in southern Puerto Rico have recently been recognized as Quaternary active. However, the kinematics of these faults, particularly any lateral component, remain largely unconstrained. It is difficult to characterize low strain‐rate faults, partially due to extensive erosional and anthropogenic landscape modification, steep relief, and frequent landsliding, limiting the preservation of geomorphic features that could serve as recorders of fault motion. Here, we constrain the kinematics along sections of the Great Southern Puerto Rico Fault Zone (GSPRFZ) on the southern coastal plain of Puerto Rico. We integrate ∼1‐m‐resolution light detection and ranging (lidar)‐derived topography, historical air photos, and field mapping to identify a series of ∼50–1200‐m‐long fault scarps and lineaments that trend northwest–southeast and extend for ≥25 km across the southern coastal plain. Fault scarps are primarily south facing, cut across topography, and displace Quaternary deposits and landforms. We document multiple offset geomorphic markers, including channel thalwegs and interfluves formed in deposits previously mapped as Quaternary piedmont alluvial plain. We observe both vertical (south‐side‐down) and right‐lateral meter‐scale displacements, which indicate that the GSPRFZ accommodates right‐lateral oblique motion in the late Pleistocene, consistent with northeast motion of the Puerto Rico and the Virgin Islands microplate away from the Hispaniola block.

Puerto Rico↗
USGS2024

On the provenance of field reports of the 1886 Charleston, South Carolina, earthquake: A seismo-historical whodunnit

Much of what is known about the effects of the 1886 Charleston, South Carolina, earthquake throughout the epicentral region can be attributed to meticulous field investigations by an individual with training in geology and engineering, Earle Sloan ( Clendenin, 1926 ). In a recent study, Bilham and Hough (2024) undertook a detailed analysis of the effects of the earthquake on railroads in the Charleston region, drawing heavily from Sloan’s reports. This exercise identified several inconsistencies in Sloan’s field reports, including understandable measurement imprecision, inferred data entry mistakes, and transcription errors. The study also begged the question, where was Sloan at the time of the mainshock and over the following week? And to what extent did he draw from secondhand information in compiling his reports? On this question Sloan’s reports were sometimes enigmatic, lending themselves to misinterpretation in contemporaneous as well as modern interpretations. Beyond the details that were germane for, and briefly summarized by, the studies of Bilham and Hough (2023 , 2024) , in this report we don our historical seismologist caps to chronicle Sloan’s activities following the earthquake. We summarize our inferences here for the benefit of future scholars who might attempt to retrace either Sloan’s footsteps or our own. This study also serves to highlight Sloan’s singular contributions to earthquake science, which were never published separately.

South Carolina↗
USGS2026

Landslide volume estimation from seismic waveform features using a global catalog of seismogenic events

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.

Seismological Research Letters↗
USGS2026

Decoding the oxidative digestion mechanism for polystyrene nanoplastic detection in the Great Lakes using a customizable Raman spectral processing algorithm

Despite the concerns surging around nanoplastics (NPs) regarding their prevalence and bioavailability in freshwater systems, robust detection of NPs in complex environmental matrices is hindered by the lack of standardized sample pretreatment and a mechanistic understanding of oxidative digestion. Here, we systematically investigate the interaction between hydrogen peroxide (H 2 O 2 ) and polystyrene (PS) NPs during digestion in deionized (DI) water and four environmental matrices from in and around the Great Lakes Basin. To facilitate high-throughput analysis, we develop Pre_peak, a customizable Raman spectral processing algorithm that achieves >99% accuracy for both NP identification and interference rejection, allowing reliable NP quantification via pixel counting and systematic decoding of the oxidative digestion mechanisms. In DI water, varying H 2 O 2 doses from 0 to 30% has negligible effects on the recovery and Raman signal intensity of PS NPs over 24 hours of digestion. However, morphological changes and aggregation of PS NPs are observed when the H 2 O 2 dose exceeds 20%. Prolonged digestion further leads to progressive NP loss. In natural waters, the optimal dosage and digestion duration depend on matrix characteristics, including dissolved organic matter (DOM) and ion composition. This study provides mechanistic insights into NP–oxidant interactions and underscores the need for matrix-tailored digestion protocols to advance standardized NP detection in freshwater environments.

Great Lakes↗
USGS2024

A conceptual framework to assess post-wildfire water quality: State of the science and knowledge gaps

Wildfire substantially alters aquatic ecosystems by inducing moderate to catastrophic physical and chemical changes. However, the relations of environmental and watershed variables that drive those effects are complex. We present a Driver-Factor-Stressor-Effect (DFSE) conceptual framework to assess the current state of the science related to post-wildfire water-quality. We reviewed 64 peer-reviewed papers using the DFSE framework to identify drivers, factors, stressors, and effects associated with each study. A total of five drivers were identified and ranked according to their frequency of occurrence in the literature: atmospheric processes > fire characteristics > ecologic processes and characteristics > land surface characteristics > soil characteristics. Commonly reported stressors include increased nutrients, runoff, and sediment transport. Furthermore, although several different factors have been used at least once to explain water-quality effects, relatively few factors outside of precipitation and fire characteristics are frequently studied. We identified several gaps indicating the need for long-term monitoring, multi-factor studies, consideration of organic contaminants, consideration of groundwater, and inclusion of soil characteristics. This assessment expands on other reviews and meta-analyses by exploring causal linkages between influential variables and overall effects in post-wildfire watersheds. Information gathered from our assessment and the framework itself can be used to inform future monitoring plans and as a guide for modeling efforts focused on better understanding specific processes or to mitigate potential risks of post-wildfire water quality.

Water Resources Research↗
USGS2026

Transit time modeling framework for predicting freshwater salinization in urban catchments

The salinity of inland freshwaters is rising globally, particularly in urban watersheds where winter road deicers are widely applied. Attributing stream salinity dynamics to specific sources and transport pathways remains challenging due to episodic salt inputs, engineered drainage, and strong coupling between hydrology and subsurface storage. We present a modeling framework that couples climate-driven deicer build-up and wash-off with transient transit time distribution theory to simulate salt transport through drainage, interflow, and groundwater pathways. Applied to an urban watershed in Northern Virginia (USA), the model reproduces ten years of high-frequency stream salinity measurements across daily-to-decadal timescales. The calibrated model implies an average deicer application of 206 tonnes Cl −1 yr −1 , or roughly one 20 kg bag of rock salt person −1 yr −1 when normalized by the 20,000 people living in the watershed. In winter months, higher infiltration routes a large fraction of snowmelt and deicers into shallow subsurface pathways, enhancing vadose-zone and interflow contributions to stream salinity. Limited subsurface storage capacity and seasonal hydrologic turnover flush excess chloride from the vadose zone and groundwater during subsequent summer storms. By linking climate-driven deicer inputs, hydrologic connectivity, and stream water age, the framework provides a transferable basis for diagnosing and managing freshwater salinization in urban watersheds.

Virginia↗
USGS2025

Long‐period ground motions from dynamic rupture simulations of large earthquakes on the creeping Hayward–Calaveras–Rodgers Creek fault system

he Hayward, Calaveras, and Rodgers Creek faults in the San Francisco Bay region of California have a high probability of producing a large earthquake in the next decades. Although these faults creep, the creep is insufficient to keep up with their relatively rapid slip rates on their deepest sections, so they have been storing tectonic strain since their last large earthquakes, with the Hayward’s and Rodgers Creek’s more than 150 yr ago. We do not know what the next large Hayward–Calaveras–Rodgers Creek earthquakes will look like or how strongly they will shake the San Francisco Bay region. Harris et al. (2021) used the 3D dynamic (spontaneous) rupture method to simulate large earthquakes on these creeping faults. In this article, we examine the resulting simulated long‐period ( T > 1 s) ground shaking from 0 to 50 km distance, for earthquakes nucleating on the Hayward fault and earthquakes nucleating on the Rodgers Creek fault. We compare these simulated long‐period ground motions with the Boore et al. (2014) well‐established empirically based ground‐motion model suitable for the slowest material velocity in our 3D velocity structure. We find that the simulated long‐period ground motions from the creeping‐fault earthquake scenarios produce a reasonable agreement with the empirical expectations if frictional cohesion is included only where it is appropriate.

California↗
USGS2024

Examining 22 years of ambient seismic wavefield at Mount St. Helens

An increase in seismic activity precedes most volcanic eruptions. Whereas event-based forecasting approaches have been successful, some eruptions remain unanticipated, resulting in casualties and damage. Our study leverages the recent advancements in ambient field seismology. We explore features extracted from continuous ambient fields using traditional methods, for example, peak ground velocity, peak ground acceleration, root mean square, root median square, real-time seismic amplitude measurement, and novel methods (displacement seismic amplitude ratio and spectral width). In addition, we explore unsupervised learning of higher order wavelet features using scattering networks. We find that combining all the methods was necessary to disentangle the effects of seismic sources from structural changes at Mount St. Helens. Although the ambient wavefield-based approach does not yield additional or more significant precursory signals than event-based methods at Mount St. Helens, our study demonstrates that the ambient wavefield provides supplementary information, mainly about structural changes and complements traditional methods. The ambient seismic wavefield offers additional insights into long-lasting processes. We find enhanced wave attenuation correlating with geochemical measurements. We interpret this as ongoing structural changes, such as dome growth or the evolution of the volcanic conduit system. On annual and decadal timescales, we interpret seasonal seismic attenuation in the shallow subsurface as groundwater fluctuations, corroborated by observations at the nearby Spirit Lake level. This multimethod approach at Mount St. Helens sheds light on a volcanic system’s underlying dynamics and structure.

Washington↗
USGS2026

Conventional hydrothermal power-producing systems of the Great Basin, USA

As part of the update to the electric-grade conventional hydrothermal assessment of the Great Basin, USA, Monte Carlo analyses of identified resources within explored regions will be performed to make estimates of discovered resources and associated uncertainty. Analyses use conditional statistics where estimates are conditioned upon a hydrothermal favorability map, allowing for the likelihood that more resources exist in regions of higher hydrothermal favorability. For these analyses, a dataset of identified hydrothermal systems is compiled, and the new compilation is described herein. Recognizing that a single hydrothermal system may be developed with multiple power plants, and that the hydrothermal upflow zone may be several kilometers across with many measurements characterizing a single hydrothermal system, a procedure was developed and employed to create clusters of points (power plants, measurements, etc.) that are associated with a single system, and a new central point was defined as the best estimator of the center of the hydrothermal system. Hydrothermal systems were uniquely identified by grouping electric-grade hydrothermal measurements and operating power plants within a distance of 10 km. Groups that are >10 km apart are assumed to be different electric-grade hydrothermal systems. While 10 km was used as the threshold, most systems were significantly further apart, and most points within groups were typically within 5 km of each other. A well measurement was considered an electric-grade measurement of a hydrothermal system if it had two properties: a measured temperature of >85 °C and evidence of hydrothermal convection. Other points that were added to the dataset are locations of operating powerplants or locations that have been classified as an electric-grade hydrothermal resource by either the U.S Geological Survey (USGS) or the Great Basin Center for Geothermal Energy. After all points are assigned to systems, new points were computed with the goal of identifying the center of the throat of the hydrothermal upflow zone. If operating powerplants exist for a system, then the arithmetic average of all power plant locations is used. Otherwise, if USGS made an estimate, that location is used. In the absence of both powerplants or USGS estimates, the arithmetic average of all electric-grade measurement locations is used. An example is shown of how these newly compiled locations might be ranked for uncertainty analyses, where higher confidence is assumed if measured temperature is higher and there are many supporting measurements indicating an electric-grade resource. In summary, 28 systems have operating power plants, an additional 78 systems are known identified electric-grade hydrothermal resources, and 99 new systems were identified as probable electric-grade systems with varying levels of confidence. These 205 locations are shown as a function of a recent hydrothermal favorability map, conceptually illustrating the conditional statistics that can be used to make estimates of the undiscovered resources of the Great Basin. An accompanying data release provides summaries of developed capacity by system and USGS estimates of likely total capacity and associated uncertainty.

Arizona, California, Idaho, Nevada, Oregon, Utah↗
USGS2025

Towards mobile wind measurements using joust configured ultrasonic anemometer for applications in gas flux quantification

Small uncrewed aerial systems (sUASs) can be used to quantify emissions of greenhouse and other gases, providing flexibility in quantifying these emissions from a multitude of sources, including oil and gas infrastructure, volcano plumes, wildfire emissions, and natural sources. However, sUAS-based emission estimates are sensitive to the accuracy of wind speed and direction measurements. In this study, we examined how filtering and correcting sUAS-based wind measurements affects data accuracy by comparing data from a miniature ultrasonic anemometer mounted on a sUAS in a joust configuration to highly accurate wind data taken from a nearby eddy covariance flux tower (aka the Tower). These corrections had a small effect on wind speed error, but reduced wind direction errors from 50° to >120° to 20–30°. A concurrent experiment examining the amount of error due to the sUAS and the Tower not being co-located showed that the impact of this separation was 0.16–0.21 ms − 1 "> ms − 1 , a small influence on wind speed errors. Lower wind speed errors were correlated with lower turbulence intensity and higher relative wind speeds. There were also some loose trends in diminished wind direction errors at higher relative wind speeds. Therefore, to improve the quality of sUAS-based wind measurements, our study suggested that flight planning consider optimizing conditions that can lower turbulence intensity and maximize relative wind speeds as well as include post-flight corrections.

Alaska↗
USGS2024

Reference 1D seismic velocity models for volcano monitoring and imaging: Methods, models, and applications

Seismic velocity models of the crust are an integral part of earthquake monitoring systems at volcanoes. 1D models that vary only in depth are typically used for real‐time hypocenter determination and serve as critical reference models for detailed 3D imaging studies and geomechanical modeling. Such models are usually computed using seismic tomographic methods that rely on P ‐ and S ‐wave arrival‐time picks from numerous earthquakes recorded at receivers around the volcano. Traditional linearized tomographic methods that jointly invert for source locations, velocity structure, and station corrections depend critically on having reasonable starting values for the unknown parameters, are susceptible to local misfit minima and divergence, and often do not provide adequate uncertainty information. These issues are often exacerbated by sparse seismic networks, inadequate distributions of seismicity, and/or poor data quality common at volcanoes. In contrast, modern probabilistic global search methods avoid these issues only at the cost of increased computation time. In this article, we review both approaches and present example applications and comparisons at several volcanoes in the United States, including Mount Hood (Oregon), Mount St. Helens (Washington), the Island of Hawai’i, and Mount Cleveland (Alaska). We provide guidance on the proper usage of these methods as relevant to challenges specific to volcano monitoring and imaging. Finally, we survey‐published 1D P ‐wave velocity models from around the world and use them to derive a generic stratovolcano velocity model, which serves as a useful reference model for comparison and when local velocity information is sparse.

Seismological Research Letters↗
USGS2025

What 25+ years of "Did You Feel It" intensities tell us about shaking in California

“When will the Big One happen?” is a question that people often have for earthquake scientists. But while waiting for the “Big One” to occur, people will usually experience frightening or damaging shaking from multiple relatively smaller‐magnitude earthquakes. Given this context, it raises the question: “Where does most of the damage come from?” Could smaller, yet more frequent, earthquakes account for the majority of reported impactful shaking? To explore this question, we consider reports of earthquake damage and felt shaking experiences from a catalog of community‐collected intensity values from the U.S. Geological Survey’s “Did You Feel It?” system. Comparing these intensities to expectations from a ground‐motion model, we find that earthquakes of magnitudes smaller than expected are responsible for most reported intensities of community decimal intensities (CDI) 4.5 and above (moderate and higher shaking intensity levels). (Here “expected value” is meant in its mathematical sense of the mean or equivalently median shaking intensity.) We also present a regional analysis of observed earthquake shaking for specific areas, detailing the maximum intensity experienced within a predetermined area. We identify several instances of M < 4.5 events that generated maximum intensities of CDI > 5 in regions in California surrounding the cities of Eureka, San Francisco, Los Angeles, and San Diego. Our results motivate the need to include smaller‐magnitude earthquakes in communications about earthquake hazard and risk reduction.

California↗
USGS2026

The digital archivist: Automating legacy macroseismic data processing using large language models

Macroseismic data are a key resource to investigate shaking and damage from preinstrumental and early instrumental eras. However, data are often stored as inconsistently formatted reports describing observed shaking and damage, making manually parsing and interpreting accounts labor‐intensive. We introduce a novel workflow using Google’s Gemini 2.5 Pro large language model (LLM) to automate the extraction and structuring of macroseismic observations from summary reports. We apply this workflow to the 22 March 1957 M 5.3 Daly City, California, earthquake as a case study. We used Gemini to extract addresses, originally assigned modified Mercalli intensity values, and descriptions from each report. To address coordinate precision limits, addresses were geocoded via Google’s Geocoding application programming interface. This workflow yielded over 2300 geocoded intensity reports for the Daly City earthquake. We use the geocoded accounts, with the original report intensity assignments, to develop a shaking intensity map that in some respects rivals modern Did You Feel It? Maps. We also extract and present data for the 9 February 1971 M L 6.7 Sylmar, California, earthquake. Our results demonstrate the potential of LLMs for reliably extracting and analyzing large, unstructured macroseismic datasets. LLMs offer a scalable solution for rapidly digitizing macroseismic archives, enabling their broader use to constrain ground‐motion models in modern seismic hazard analysis and to improve our understanding of site effects in urban areas. The concepts explored here may also be applied to the handling of other legacy seismological and earth science data.

Seismological Research Letters↗