USGS Science⌕ Search

SEARCH · USGS Science

Results for “Report”

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.

712 records · Page 39Linked to original sources
USGS2026

Patterns and drivers of cliff erosion in Big Sur, California, USA using repeat photogrammetry, 2017–2023

Seacliff erosion in steep terrain poses major risks to transportation and critical infrastructure. In Big Sur, California, USA, seacliff erosion threatens the sustainability of the central coast stretch of California State Route 1, a transportation corridor that is critical to the region's economy. Published cliff retreat rates for the region range from 1 to 40 cm yr −1 , highlighting that high-resolution, process-based studies could enhance understanding of the causes of spatial and temporal variability. We quantified cliff erosion and investigated its drivers along ∼13 km of the Big Sur coastline at week–month timescales during the late fall to early spring wet seasons between January 2017 and June 2023 by analyzing 3D point clouds developed from aerial imagery using four-dimensional structure-from-motion (4D SfM) photogrammetry techniques. We calculated cliff face retreat rates of 2.23 ± 3.06 cm yr −1 (mean ±1 σ ), an order of magnitude lower than long-term estimated rates for the region (which included large deep-seated landslides), but in line with short-term rates reported across California. Change detection imagery comparison, cliff profiles through time, and statistical analysis reveal a cyclical cliff evolution process in which erosion by wave action at the cliff base destabilizes the cliff and primes it for subsequent failure during precipitation events. Although more erosion by volume could be attributed to precipitation-induced increases in soil moisture (784 m 3 km −1 yr −1 ) compared with erosion attributed to wave power (282 m 3 km −1 yr −1 ), our observations underscore the coupled nature of these processes in driving cliff evolution, consistent with established theory and observations.

California↗
USGS2021

Critical review of mercury methylation and methylmercury demethylation rate constants in aquatic sediments for biogeochemical modeling

Mercury is a toxin that causes neurological impairments in adults, is particularly harmful for fetuses and children, and is deadly in severe cases, making it a worldwide health concern. Methylmercury (MeHg) is the environmentally relevant form of mercury (Hg) because it biomagnifies along the food chain. Methylmercury is mainly produced in aquatic sediments via methylation of inorganic Hg (Hg(II)) and transformed back via demethylation. Because transformation rates determine MeHg concentrations, quantification of methylation and demethylation rates is needed to inform management of MeHg. Published rate constants for Hg(II) methylation ( 𝑘 𝑚 ) and MeHg demethylation ( 𝑘 𝑑 ) vary greatly, stemming partly from differences in experimental methods. We conducted a comprehensive review of rate laws, evaluated published rate constants, and performed biogeochemical simulations to assess variability in reported 𝑘 𝑚 and 𝑘 𝑑 . Based on selected studies employing the same pseudo-first-order rate law and similar experimental methods, we found that 𝑘 𝑚 = 0.04 ± 0.03 d −1 is a reasonable range for wetland sediments. Over a number of environments, maximum 𝑘 𝑑 was smaller at sites without Hg source ( 𝑘 𝑑 = 0.5 d −1 ) than at sites with identified Hg source ( 𝑘 𝑑 = 1.8 d −1 ). Larger variability and higher uncertainty in 𝑘 𝑑 compared to 𝑘 𝑚 highlight the need for more research on MeHg demethylation rates. This critical review: (a) aids the design of future experimental studies of 𝑘 𝑚 and 𝑘 𝑑 ; (b) provides guidance for comparing rate constants from different studies; (c) presents a biogeochemical reaction model to assess rate constants; and (d) informs selection of 𝑘 𝑚 and 𝑘 𝑑 values from the literature for use in model simulations.

Critical Reviews in Environmental Science and Tech↗
USGS2025

Seismic intensities, ground motions and basin amplification from the 2022 Mw 6.4 Ferndale, California intraslab earthquake

On December 20, 2022, a Mw 6.4 earthquake occurred at a depth of 18 km within the subducting Gorda plate in the Mendocino Triple Junction (MTJ), one of the most seismically active regions of the contiguous United States, causing widespread damage to local communities. Here we document the seismic intensities, ground motions, and basin amplification effects recorded by this earthquake across Humboldt County as part of an ongoing scientific effort to understand subduction zone earthquake hazards. Modified Mercalli Intensity (MMI) values from our post-earthquake field survey report shaking intensities as high as VIII (Severe). Strong ground motion data from 54 seismic stations were processed to calculate amplitude and frequency content parameters across Northern California. The maximum calculated geometric mean PGA and PGV are 1053 cm/s 2 and 52 cm/s (respectively), both recorded within the 3 km deep Eel River sedimentary basin. Comparisons with four published Ground Motion Prediction Equations indicate that PGA and PGV measurements align with expected attenuation-distance patterns for intraslab earthquakes of this nature. Within the Eel River Basin, ground motions for frequencies above 1 Hz are amplified, with respect to reference stations, by factors greater than 2. Our findings suggest that peak ground motions were mainly caused by sedimentary basin site-effects within the Eel River basin, although rupture directivity may have also increased ground motion amplitudes locally. While the Ferndale area is primarily impacted by shallow intraslab earthquakes, our results also raise questions about site-response and basin amplification hazards from a potential megathrust earthquake. The high seismicity rates of the southernmost Cascades call for stronger regional preparedness and improved strategies to mitigate the effects of such a large-scale disaster.

California↗
USGS2024

Bioconcentration of per- and polyfluoroalkyl substances and precursors in fathead minnow tissues environmentally exposed to aqueous film-forming foam-contaminated waters

Exposure to per- and polyfluoroalkyl substances (PFAS) has been associated with toxicity in wildlife and negative health effects in humans. Decades of fire training activity at Joint Base Cape Cod (MA, USA) incorporated the use of aqueous film-forming foam (AFFF), which resulted in long-term PFAS contamination of sediments, groundwater, and hydrologically connected surface waters. To explore the bioconcentration potential of PFAS in complex environmental mixtures, a mobile laboratory was established to evaluate the bioconcentration of PFAS from AFFF-impacted groundwater by flow-through design. Fathead minnows ( n = 24) were exposed to PFAS in groundwater over a 21-day period and tissue-specific PFAS burdens in liver, kidney, and gonad were derived at three different time points. The ∑PFAS concentrations in groundwater increased from approximately 10,000 ng/L at day 1 to 36,000 ng/L at day 21. The relative abundance of PFAS in liver, kidney, and gonad shifted temporally from majority perfluoroalkyl sulfonamides (FASAs) to perfluoroalkyl sulfonates (PFSAs). By day 21, mean ∑PFAS concentrations in tissues displayed a predominance in the order of liver > kidney > gonad. Generally, bioconcentration factors (BCFs) for FASAs, perfluoroalkyl carboxylates (PFCAs), and fluorotelomer sulfonates (FTS) increased with degree of fluorinated carbon chain length, but this was not evident for PFSAs. Perfluorooctane sulfonamide (FOSA) displayed the highest mean BCF (8700 L/kg) in day 21 kidney. Suspect screening results revealed the presence of several perfluoroalkyl sulfinate and FASA compounds present in groundwater and in liver for which pseudo-bioconcentration factors are also reported. The bioconcentration observed for precursor compounds and PFSA derivatives detected suggests alternative pathways for terminal PFAS exposure in aquatic wildlife and humans.

Environmental Toxicology and Chemistry↗
USGS2024

On algorithmically determined versus traditional macroseismic intensity assignments

The utility of macroseismic data, defined as the effects of earthquakes on humans and the built environment, has been increasingly recognized following the advent of online systems that now produce unprecedented volumes of macroseismic intensity information. Contributed reports from the U.S. Geological Survey “Did You Feel It?” (DYFI) system ( Wald et al ., 1999 ) are used to generate intensity values with an algorithm based on seminal work by Dengler and Dewey (1998) . The algorithm was developed initially to reproduce intensity values assigned by expert opinion using questionnaire results collected by telephone survey. In this article, I discuss reasons why intensity values from (self‐selected) DYFI responses can differ from values that would be assigned by expert opinion given more complete data from randomly selected participants. For example, with the data used by Dengler and Dewey (1998) , intensities near 4 could be determined from the percentage of people who felt shaking in each town. With less spatially rich data from self‐selected participants, this percentage often cannot be determined reliably. Audible noises are key additional diagnostic criteria for modified Mercalli intensity (MMI) 4, but, although the DYFI system includes a question about noise, following Dengler and Dewey (1998) , the DYFI algorithm does not include a noise indicator. At the upper end of the scale, as defined the DYFI algorithm yields a maximum intensity value of 9.05, nominally corresponding to peak ground acceleration of 75% g . These and other factors can result in DYFI values that are low compared to traditional MMI values assigned using expert opinion, even absent factors that can bias traditional MMI assignments. Modern ground‐motion intensity conversion equations determined using DYFI intensities are expected to be appropriate for DYFI intensities, but the results of this study suggest that biases may be introduced if DYFI and traditional intensities are assumed to be interchangeable.

Seismological Research Letters↗
USGS2024

Static and dynamic strain in the 1886 Charleston, South Carolina, earthquake

During the 1886 Mw 7.3 Charleston, South Carolina, earthquake, three railroads emanating from the city were exposed to severe shaking. Expansion joints in segmented railroad tracks are designed to allow railroad infrastructure to withstand a few parts in 10,000 of thermoelastic strain. We show that, in 1886, transient contractions exceeding this limiting value buckled rails, and transient extensions pulled rails apart. Calculated values for dynamic strain in the meizoseismal region are in reasonable agreement with those anticipated from the relation between strain and moment magnitude proposed by Barbour et al. (2021) and exceed estimated tectonic strain released by the earthquake by an order of magnitude. Almost all of the documented disturbances of railroad lines, including evidence for shortening of the rails, can thus be ascribed to the effects of dynamic strain changes, not static strain. Little or no damage to railroads was reported outside the estimated 10 −4 dynamic strain contour. The correspondence between 10 −3 and 2×10 −4 contours of dynamic strain and Mercalli intensity 9 and 8, anticipated from the dependence of each quantity on peak ground velocity, suggests it may be possible to use railroad damage to quantitatively estimate shaking intensity. At one location, near Rantowles, ≈20 km west of Charleston, a photograph of buckled track taken one day after the earthquake has been cited as evidence for shallow dextral slip and has long focused a search for a causal fault in this region. Photogrammetric analysis reveals that the buckle was caused by transient contraction of <10 cm with no dextral offset. Our results further weaken the evidence for faulting in the swamps and forests south of the Ashley River in 1886, hitherto motivated by the photograph and limited macroseismic evidence for high‐intensity shaking.

South Carolina↗
USGS2026

Monazite and xenotime U-Pb geochronology and thermometry of the Blue Ridge and Inner Piedmont of North Carolina: Implications for the thermal-metamorphic evolution of the southern Appalachian metamorphic “core”

The southern Appalachian orogen preserves a complex distribution of metamorphism and deformation varying in timing, magnitude, and spatial extent. These complexities give rise to disparate interpretations for southern Appalachian tectonic evolution, which complicates the testing and interpretation of tectonic models in this system. New monazite (Mnz) and xenotime (Xtm) laser ablation split stream (LASS) analyses alongside Mnz-Xtm thermometry in the orogenic core in the eastern Blue Ridge (EBR), western Inner Piedmont (WIP), and Cat Square terranes (CST) of North Carolina yield new constraints that define distinct pro- and retrograde metamorphic events. The EBR preserves two prograde thermal events: the Taconic (∼470-440 Ma, >660°C) and Neoacadian (∼380-340 Ma, 600–700°C), separated by a period of cooling (exhumation?) and followed by garnet breakdown from 339 to 329 Ma. Evidence of pervasive Neoacadian ductile deformation in the EBR is largely limited to the Brevard fault zone (BFZ), indicating that a major rheological gradient existed across the BFZ during the Neoacadian and early Alleghanian. Southeast of the BFZ, in the WIP and CST, monazite data define a protracted Neoacadian evolution from early mineral growth at ∼405 Ma at ∼450–600°C to >700°C at ∼360 Ma, followed by early Alleghanian retrograde metamorphism and deformation (<345 Ma, 350–500°C). These constraints, together with previously reported thermobarometric data, define a P-T-t evolution for the WIP and CST consistent with Neoacadian crustal flow, while the coeval presence of a thermal-rheological boundary along the BFZ further supports a model of Neoacadian crustal “escape” flow within the orogen.

North Carolina↗
USGS2025

From hydrated silica to quartz: Potential hydrothermal precipitates found in Jezero crater, Mars

On Earth, silica-rich phases from opal to quartz are important indicators and tracers of geological processes. Hydrated silica, such as opal, is a particularly good matrix for the preservation of molecular and macroscopic biosignatures. Cherts, a type of silica-dominated rocks, provide a unique archive of ancient terrestrial life while quartz is the emblematic mineral of the Earth's continental crust. On Mars, hydrated silica has been detected in several locations based on remote sensing and rover-based studies. In the present article we report on the detection of cobbles made of hydrated silica (opal or chalcedony), as well as well-crystallized quartz. These detections were made with the SuperCam instrument onboard Perseverance (Mars 2020 mission), using a combination of LIBS, infrared and Raman spectroscopy. Quartz-dominated stones are detected unambiguously for the first time on the Martian surface, and based on grain size and crystallinity are proposed to be of hydrothermal origin. Although these rocks were all found as float, we propose that these detections are part of a common hydrothermal system, and represent different depths / temperatures of precipitation. This attests that hydrothermal processes were active in and around Jezero crater, possibly triggered by the Jezero crater-forming impact. These silica-rich rocks, in particular opaline silica, are very promising targets for sampling and return to Earth given their high biosignature preservation potential.

Earth and Planetary Science Letters↗
USGS2026

Rapid seismic and infrasound assessment of large landslides: A case study from Denali National Park and Preserve (Alaska)

Large, rapid landslides are a global hazard that can occur in remote, mountainous areas. Eyewitness reports of landslides and satellite imagery can often be limited or delayed, particularly during inclement weather. However, landslide-generated seismic and infrasound (low-frequency atmospheric sound) waves can be remotely detected in near real-time. This information can significantly expedite characterization and possible landslide response activities. Here, we highlight these capabilities using a > 4 million m³ ice–rock avalanche in Denali National Park and Preserve (Alaska). This event was detected via a landslide-specific seismic location and volume estimation algorithm deployed in Alaska, and — notably — by standard earthquake monitoring systems. Following rapid detection of this event, we combined its seismic and infrasound dataset with optical, synthetic aperture radar, and oblique aerial imagery, multitemporal digital elevation models, and a numerical flow model to reconstruct its failure timeline and dynamics. We apply array processing to infrasound signals traveling > 250 km and find that two precursory events occurred minutes prior to the main failure. We use long-period seismic signals to infer the force exerted by the landslide on the Earth and constrain the rheological parameters of our numerical flow simulation with this result and deposit morphology. The main failure produced a steeply-dipping impulsive initial downward force and reached speeds exceeding 60 m/s. This impulsive force generated relatively strong seismic body waves, which contributed to the earthquake system detection. This large, remote Alaska landslide underscores the key value of seismic and infrasound analysis for rapid landslide assessment and motivates efforts to further operationalize these approaches.

Alaska↗
USGS2024

Precariously balanced rocks in northern New York and Vermont, U.S.A.: Ground-motion constraints and implications for fault sources

Precariously balanced rocks (PBRs) and other fragile geologic features have the potential to constrain the maximum intensity of earthquake ground shaking over millennia. Such constraints may be particularly useful in the eastern United States (U.S.), where few earthquake‐source faults are reliably identified, and moderate earthquakes can be felt at great distances due to low seismic attenuation. We describe five PBRs in northern New York and Vermont—a region of elevated seismic hazard associated with historical seismicity. These boulders appear to be among the most fragile PBRs in the region, based on reports from hobbyists. The PBRs are glacial erratics, best evidenced by glacial striations on bedrock pedestals. The pedestals themselves are locally high knobs, often situated on regionally high topography; this setting limits soil development and indicates that any outwash deposits were likely ephemeral. As a result, PBR ages can be reliably established by the retreat of the last continental ice sheet, ∼15–13 ka. To quantify the fragility of the PBRs, we surveyed them with ground‐based light detection and ranging and calculated geometric parameters from the point clouds, field observations, and seismic responses. Preliminary validation of the 2023 time‐independent U.S. National Seismic Hazard Model (NSHM) shows that the existence of PBRs is generally consistent with the median site‐specific hazard curves. Only the Blue Ridge Road site suggests a modest reduction in hazard. To visualize the ensemble of data, we mapped the minimum permissible distance to potential source faults around each PBR site as a function of source magnitude by using the ground‐motion models from the 2023 NSHM. Viewed in this manner, our data are consistent with potential M ∼6.5 earthquake‐source faults in many parts of the Lake Champlain Valley and northern Adirondack Mountains. Our work illustrates a potential pathway for better constraining earthquake‐source faults in regions of cryptic faults.

Bulletin of the Seismological Society of America↗
USGS2026

The petrogenesis of Þingmúli volcano, East Fjords, Iceland

In this work we revisit Þingmúli volcano (Þ = Th), a classic locality known as an example of a complete tholeiitic differentiation. Þingmúli is a ~ 9.5 Ma extinct central volcano located in the East Fjords of Iceland, in which the whole compositional spectrum from basalt to rhyolites have erupted. These volcanic products have been previously considered as petrogenetically related by an ideal fractionation trend, regardless any temporal relationship or volumetric considerations. Here we report new whole-rock geochemistry, mineral chemistry, isotope analyses, estimation of residence times of the different eruptive deposits, and an update of the original petrogenetic model. Our results highlight that an enriched source, likely spinel lherzolites, generated transitional-alkaline basaltic melts after 15–20% of partial melting at depths of 40–45 km. Many of these basaltic melts erupted at various stages of the volcano's history, while others remained longer in the volcanic plumbing system. These evolved by fractional crystallisation into basaltic andesite magmas with a residence time of ~5 years based on the crystal size distribution of the plagioclase population. Isotopic differences between the basalts/basaltic andesites ( 87 Sr/ 86 Sr ~ 0.7034; 143 Nd/ 144 Nd ~ 0.51315) and the erupted rhyolites ( 87 Sr/ 86 Sr ~ 0.7037; 143 Nd/ 144 Nd ~ 0.51304) indicate that the latter are not petrogenetically related to the former. Therefore, instead of a fractional crystallisation mechanism to generate the rhyolites, we propose the partial melting of ignimbrite layers located beneath the volcano. The broad range of trace element concentrations in andesites and dacites and their different isotopic values compared to the basalts strongly suggest that these magmas have been generated by magma mixing between basaltic and rhyolitic melts, similar to modern day Icelandic volcanoes such as Hekla. These results highlight the need to revisit previously studied Icelandic classic localities and reassess their traditionally proposed petrogenetic models.

Þingmúli volcano, East Fjords↗
USGS2026

A practical decision tool for marine bird mortality assessments

Given the rise in anthropogenic, environmental, and disease events contributing to marine bird mortality, there is a critical need to improve the rigor of mortality assessments. Deficits in data collection and mortality estimation can hinder a manager’s ability to document the scale of events and assess population level impacts. Therefore, to inform decisions required during activities, such as conservation status assessments or harvest management, organizations may choose to incorporate mortality assessments into response plans. Resources, capacity, and assets to assess mortality vary across jurisdictions (federal, state, Indigenous, local, etc.), and clear guidance to support mortality assessments is often unavailable or not clearly addressed. Here, we present a decision support tool to help managers identify and evaluate survey options to assess bird mortality in a diverse array of scenarios. The objective of the decision tool is to improve data collection and availability, which will increase the ability to estimate mortality robustly, given situation-specific attributes and constraints. This decision tool is designed to guide the response when a mortality event is initially encountered and offers suggestions for assessment and reporting procedures in the absence of other guidance or to complement existing protocols. The decision tool is also meant to inform decision making for response determination and resource allocation. The tool facilitates examination of options for further assessment and monitoring, which users determine by examining questions pertaining to species prioritization, determination of mortality minimum spatial extent, and the potential magnitude of impacts on affected species. Finally, identification of appropriate survey methods that address imperfect detection when a complete census is not possible are determined by exploring location, spatial and temporal extent, and the type of species affected. Ultimately, this decision tool aims to facilitate and improve the standardization of mortality assessments, equipping managers with a practical resource to navigate the decision-making process for marine bird mortality estimation.

Ornithological Applications↗
USGS2025

Local, regional, and distal recordings of seismic unrest at Tau Island volcano, American Samoa

A seismic swarm near Taʻū Island, a volcanic island in eastern American Samoa, occurred from July to October 2022. The earliest unrest was noted as felt shaking reports in late July, and instrumentation varied in the beginning of the sequence as the U.S. Geological Survey (USGS) Hawaiian Volcano Observatory responded by installing temporary and then permanent seismometers to monitor the activity. This network variability made it difficult to characterize the earliest seismicity and contextualize the entire sequence to discriminate between an underlying tectonic or volcanic source. Here, we present results analyzing hydroacoustic detections from an International Monitoring System hydrophone array near Wake Island, 4500 km northwest of Taʻū Island volcano. Using least-squares beamforming analysis, we create a catalog of T-wave detections from the direction of Taʻū Island to track the earthquakes, some of which were located by the USGS National EarthquakeInformation Center. Both the rate and hydroacoustic pressures, which we interpret as a proxy for earthquake size, gradually increased from late July to August, peaking on August 19 (rate) and August 24 (size), before decreasing to background in late September. Minutes-long bursts of tremor were also contemporaneously recorded as local network data became avail-able on August 20. Tremor activity continued throughout the rest of August, peaking on August 25, before ending in earlySeptember. These tremor bursts were band-limited to ~ 1–5 Hz and recorded as S waves at a regional station on the island of Upolu in Samoa, 250 km to the west of Taʻū Island. Our results do not constrain the tremor locations, but comparisons of earthquake and tremor reduced displacements recorded locally and regionally suggest a deeper tremor source. We interpret the increase in earthquake size and rate, together with the occurrence, characteristics, and relative depth of the tremor to be the result of magmatic activity beneath Taʻū Island volcano.

American Samoa, Taʻū Island↗
USGS2026

Groundwater quality near an oil field in a stream-dominated recharge setting, California, USA

Alluvial valley aquifers are important sources of water supply in many areas but effects of co-located oil and gas development on these resources have not been widely reported, especially in settings where recharge is dominated by stream infiltration. Interpreting the presence of geochemical indicators in the context of hydrology, geology, and other factors provides a more complete understanding of the relations between groundwater and sources of oil-field fluids and aids in identifying risks associated with oil and gas development. Groundwater and Salinas River water samples were collected in an alluvial valley near the San Ardo Oil Field in Monterey County, California and analyzed for a wide range of dissolved chemical, gas, and isotopic constituents to determine if oil-field fluids (water and gas from oil-producing and non-producing zones) have mixed with fresh groundwater used for supply. Hydraulic gradients, age-dating tracers, and other geochemical indicators show that recharge from the Salinas River has the potential to dilute oil-field fluids that might migrate or seep into the aquifer. Groundwater and Salinas River water collected downgradient of the San Ardo Oil Field showed little or no evidence of mixing with oil-field fluids. Some samples within the oil field contained trace amounts of hydrocarbons or elevated temperatures, indicating that any potential effects from oil-field activities are minor or have been diluted by recharge from the Salinas River. The two samples with the most geochemical evidence of potential mixing with oil-field fluids (SP-18 and GW-17) were collected west of or along the Los Lobos fault, where naturally occurring hydrocarbons are near the land surface. Those samples are also near active or inactive oil-field wells, and so anthropogenic activities and pathways cannot be ruled out as a cause of trace detections of hydrocarbons and elevated temperatures in the aquifer.

California↗
USGS2024

Near-real-time earthquake-induced fatality estimation using crowdsourced data and few-shot large-language models

When a damaging earthquake occurs, immediate information about casualties (e.g., fatalities and injuries) is critical for time-sensitive decision-making by emergency response and aid agencies in the first hours and days. Systems such as the Prompt Assessment of Global Earthquakes for Response (PAGER) by the U.S. Geological Survey (USGS) were developed to provide a forecast of such impacts within about 30 min of any significant earthquake globally. However, existing disaster-induced human loss estimation systems often rely on early casualty reports manually retrieved from global traditional media, which are labor-intensive, time-consuming, and have significant time latencies. Recent approaches use keyword matching and topic modeling to identify human casualty-relevant information from social media but tend to be error-prone when dealing with complex semantics in multi-lingual text data and parsing dynamically changing and conflicting human death and injury numbers shared by various unvetted sources in social media platforms. In this work, we introduce an end-to-end framework to significantly improve the timeliness and accuracy of global earthquake-induced human loss forecasting using multi-lingual, crowdsourced social media. Our framework integrates (i) a hierarchical casualty extraction model built upon large language models, prompt design, and few-shot learning to retrieve quantitative human loss claims from social media, (ii) a physical constraint-aware, dynamic-truth discovery model that discovers the truthful human loss from massive noisy and potentially conflicting human loss claims, and (iii) a Bayesian updating loss projection model that dynamically updates the final loss estimation using discovered truths. We test the framework in real-time on a series of global earthquake events in 2021 and 2022 and show that our framework effectively automates the retrieval of casualty information faster but with comparable accuracy to those now retrieved manually by the USGS. The code associated with this work is made available at: https://github.com/SusuXu-s-Lab/Hierarchical-Earthquake-Casualty-Information-Retrieval

International Journal of Disaster Risk Reduction↗
USGS2025

Regional characterization of coal resources in the U.S. Gulf Coast

There is increasing interest in extracting critical minerals (CM), including rare earth elements (REE), from coals in the United States to address the overreliance on imported REE. The U.S. Gulf Coast and the Williston basins are the two major lignite-bearing basins within the country. Recent REE and CM studies of the lignite in these basins have indicated that the coals may be a viable source material for REE and CM extraction. To evaluate in-place coal as a potential source of REE and CM, the coal resources need to be quantified. This study presents the results of a regional analysis of the U.S. Gulf Coast lignite and bituminous coal resources that might be available as potential sources of REE and CM. The resource analysis used kriging methods to develop isopleth maps of cumulative coal thickness throughout the region using data from 31,181 drill holes and other data points. The estimated total coal resource in the Gulf Coast is about 83 billion metric tons in the upper 90 m (~ 300 ft) of the subsurface. Texas accounted for 40 percent (32 billion metric tons) of the total resource, followed by Mississippi (24 %, 20 billion metric tons), Louisiana (14 %, 12 billion metric tons), Tennessee (10 %, 8.5 billion metric tons), and Arkansas (6 %, 5.1 billion metric tons). The remaining states each accounted for less than 5 percent of the total resource. Georgia had the smallest resource estimated at 7 million metric tons. Here we report the first known state-wide lignite resource estimates for Georgia, Kentucky (820 million metric tons), and Missouri (1,800 million metric tons). A comparison of the results of this study with those of previous Gulf Coast and Williston Basin resource studies is difficult because each study used different data sources, assessment methodologies, overburden depths, and qualifying coal thicknesses. Coal-power electric generation has sharply decreased in past decades and mining of these coals for CM and REE could provide additional co-products such as activated carbon and other uses such as fertilizer (soil enhancer).

EarthArXiv↗
USGS2026

Evaluating machine learning approaches to identify and predict oil and gas produced water lithium concentrations

Recently, the demand for battery-grade lithium has substantially increased, largely due to electrification of the transportation sector. The search for new lithium sources has turned to produced waters (frequently brines), a large-volume wastewater by-product of oil and gas extraction. Geochemical analysis indicates the presence of varying concentrations of lithium from produced water samples collected across the United States and represented in the U.S. Geological Survey’s National Produced Water Geochemical Database, as well as mixtures of Marcellus Shale produced water included in the Pennsylvania Department of Environmental Protection’s Oil and Gas Well Waste Reports. We first examined whether the geochemical signature of the lithium-bearing produced waters is sufficiently distinct so that machine learning (ML) can be used to correctly classify samples to the formation of origin. The produced water sample data used to assess classification accuracy were from the Marcellus Shale, Utica Shale and Point Pleasant Formation (Utica), and Smackover Formation oil and gas wells. Further, we evaluated the potential for ML to accurately classify Marcellus Shale produced water spatially (i.e., northeast versus southwest Pennsylvania). We then investigated whether ML algorithms applied to a suite of geochemical concentration data (i.e. Ba, Br, Cl, K, Mg, Sr) may be used to predict the lithium concentration of an unknown sample. Finally, we applied an estimated economic lithium grade cutoff of 150 milligrams per liter (mg/l) and assessed the utility of ML to predict whether a produced water sample would fall above or below the grade cutoff based on the suite of geochemical parameters. Four machine learning algorithms—Random Forest (RF), Gradient Boosting Trees (GBT), Extreme Boosting (XGBoost), and Deep Neural Networks (DNN) were assessed. This study successfully demonstrates that all four machine learning methods can precisely and accurately estimate lithium concentrations and geologic formation classification. The products of this study contribute to the growing body of knowledge aimed at expanding the lithium resource base within the United States.

Alabama, Arkansas, Florida, Georgia, Louisiana, Mi↗
USGS2026

Uncertainty quantification of geologic energy storage in depleted gas reservoirs using material balance equations embedded in a hierarchical errors-in-variables model

The storage potential for gas in geologic settings, such as depleted hydrocarbon reservoirs and solution-mined salt caverns, is becoming salient to future energy infrastructure planning. Technologies such as carbon capture, utilization, and storage, carbon dioxide-enhanced oil recovery, and natural gas and hydrogen storage help meet growing energy demands, reduce carbon emissions to meet climate goals, and provide energy security amid geopolitical uncertainties. Therefore, estimates of underground gas storage capacity could be useful for efficiently navigating the energy transitions. Material balance is a fundamental method in reservoir engineering for estimating original gas in place and potential storage capacity at the scale necessary for national assessments of subsurface pore space resources. However, the deterministic method cannot accommodate multiple data sources or quantify uncertainty in predictions. In this study, a method that embeds material balance equations within a hierarchical errors-in-variables model is proposed which allows the estimation of the distributions of reservoir properties needed for assessments. Uncertainties associated with these reservoir properties have traditionally been expert-elicited, whereas the uncertainty estimates from the proposed models are data-driven. Capacity and uncertainty estimates can be used in a probabilistic resource assessment, supplementing information traditionally used by assessors or even replacing this expert elicitation step when data are unavailable. Various regression models are compared in a case study of the Michigan Basin, a large contributor to the United States’ current natural gas storage capacity. In particular, errors-in-variables models help ameliorate regression dilution and can quantify uncertainty in predictions of pressure in addition to storage capacity. Overfitting is addressed by quantifying generalization error and model averaging in simple and stratified cross-validation against reported working gas capacity, representing the varying quality and quantity of available data. Incorporating a statistical framework into existing numerical methods in reservoir engineering can improve the quality of estimation, and in particular, this method brings rigor to uncertainty quantification as part of a larger effort by the U.S. Geological Survey to assess domestic energy gas storage resources in depleted hydrocarbon reservoirs.

Mathematical Geosciences↗