Geology Reports⌕ Search

SEARCH · Geology Reports

Results for “Input”

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.

At least 1,387 records · Page 77Linked to original sources
USGS2025

Mapping predicted ecological states at landscape scales using remote sensing data and machine learning

Dryland ecosystems, covering 45% of the Earth's land and supporting over one-third of the global population, face significant threats from land degradation and ecological state change. Managing these ecosystems is complex, and science-based frameworks like Ecological Site Descriptions and state-and-transition models are essential tools for guiding decisions to support ecological health while maintaining stakeholder values such as grazing, wildlife, and recreation. However, alignment of these frameworks with smaller scale soil survey maps limits their applicability to broader ecological processes. Here, we extend these frameworks to larger landscapes with a machine learning approach that integrates large-scale, high-resolution vegetation data with identified ecological states from a data-driven state-and-transition model developed for a landscape-scale Ecological Site Group. A “global” model, which used combined inputs from multiple remotely sensed datasets, outperformed individual dataset models based on evaluation with independent data. Ecological state maps generated through this approach broaden the utility of state-and-transition models across Ecological Site Groups, providing a more spatially robust tool for land management at watershed and larger landscape scales. These methods, and the associated ecological state maps, can help meet critical needs for improved land condition assessments that support development of resource management plans and help identify priority areas for restoration and conservation.

Arizona, Colorado, New Mexico, Utah, Wyoming↗
USGS2026

Integrating climate data and river modeling to reveal Chinook salmon habitat conditions in subarctic river basins

Climatic extremes can impact the productivity of aquatic species, affecting ecosystems and fishery-dependent communities. Advances in climate products, such as gridded datasets and downscaled projections, may be useful for quantifying freshwater habitat conditions and predicting climate change effects on fish. However, limited guidance exists for selecting climate products to develop indicators of freshwater habitat conditions that influence fish population dynamics. Here, we develop an approach for identifying streamflow and stream temperature models to address this need. We evaluated skill in predicted versus observed streamflow and stream temperature, with predictions depending on different models and gridded climate data as inputs. The best performing models were used in a case study exploring habitat conditions influencing Chinook salmon in the Yukon and Kuskokwim River basins, two remote high-latitude watersheds with few in situ habitat observations and recent salmon declines. Three modeled streamflow datasets had variable performance (median Nash–Sutcliffe efficiencies from 0.39 to 0.70). Three gridded temperature products differed in their ability to explain variation in weekly stream temperatures (median r 2 from 0.42 to 0.76). We selected a single gridded air temperature dataset to compare two novel predictive stream temperature models, both of which had good accuracy (root mean squared error [RMSE] of 1.19 and 0.95°C). Stream temperature indicators calculated from modeled daily data, maximum temperatures during adult migration and cumulative temperatures during juvenile rearing, had high spatial correlation across tributaries within the Yukon and Kuskokwim River basins and showed significant warming over the past 40 years. Streamflow indicators calculated from modeled daily data, maximum flow during spawning and median flow during rearing, had few trends and were largely uncorrelated within the Yukon River basin and moderately correlated within the Kuskokwim River basin. Overall, we found that generic measures of model performance varied considerably, and it was important to consider the models best suited to our case study. For both streamflow and stream temperature, multiple high-performing models allowed estimation of ecologically relevant conditions affecting Chinook salmon. The approach we used to estimate local-scale habitat conditions has value to identify synchronous conditions that may influence multiple salmon populations under a changing subarctic climate.

Alaska↗
USGS2017

Host density increases parasite recruitment but decreases host risk in a snail-trematode system

Most species aggregate in local patches. High host density in patches increases contact rate between hosts and parasites, increasing parasite transmission success. At the same time, for environmentally-transmitted parasites, high host density can decrease infection risk to individual hosts, because infective stages are divided among all hosts in a patch, leading to safety in numbers. We tested these predictions using the California horn snail, Cerithideopsis californica (=Cerithidea californica), which is the first intermediate host for at least 19 digenean trematode species in California estuaries. Snails become infected by ingesting trematode eggs or through penetration by free-swimming miracidia that hatch from trematode eggs deposited with final-host (bird or mammal) feces. This complex life cycle decouples infective-stage production from transmission, raising the possibility of an inverse relationship between host density and infection risk. In a field survey, higher snail density was associated with increased trematode (infected snail) density, but decreased trematode prevalence, consistent with either safety in numbers, parasitic castration, or both. To determine the extent to which safety in numbers drove the negative snail density-trematode prevalence association, we manipulated uninfected snail density in 83 cages at eight sites within Carpinteria Salt Marsh (CA, USA). At each site, we quantified snail density and used data on final-host (bird and raccoon) distributions to control for between-site variation in infective-stage supply. After three months, overall trematode infections per cage increased with snail-biomass density. For egg-transmitted trematodes, per-snail infection risk decreased with snail-biomass density in the cage and surrounding area, whereas per-snail infection risk did not decrease for miracidium-transmitted trematodes. Furthermore, both trematode recruitment and infection risk increased with infective-stage input, but this was significant only for miracidium-transmitted species. A model parameterized with our experimental results and snail densities from 524 field transects estimated that safety in numbers, when combined with host aggregation, halved per-capita infection risk in this snail population. We conclude that, depending on transmission mode, host density can enhance parasite recruitment and reduce per-capita infection risk.

Ecology↗
USGS2020

Magnitude and direction of stream–forest community interactions change with timescale

Networks of direct and indirect biotic interactions underpin the complex dynamics and stability of ecological systems, yet experimental and theoretical studies often yield conflicting evidence regarding the direction (positive or negative) or magnitude of these interactions. We revisited pioneering data sets collected at the deciduous forested Horonai Stream and conducted ecosystem‐level syntheses to demonstrate that the direction of direct and indirect interactions can change depending on the timescale of observation. Prior experimental studies showed that terrestrial prey that enter the stream from the adjacent forest caused positive indirect effects on aquatic invertebrates during summer by diverting fish consumption. Seasonal and annual estimates of secondary production and organic matter flows along food web pathways demonstrate that this seasonal input of terrestrial invertebrate prey increases production of certain fish species, reversing the indirect effect on aquatic invertebrates from positive at the seasonal timescale to negative at the annual timescale. Even though terrestrial invertebrate prey contributed 54% of the annual organic matter flux to fishes, primarily during summer, fish still consumed 98% of the aquatic invertebrate annual production, leading to top‐down control that is not revealed in short‐term experiments and demonstrating that aquatic prey may be a limiting resource for fishes. Changes in the direction or magnitude of interactions may be a key factor creating nonlinear or stabilizing feedbacks in complex systems, and these dynamics can be revealed by merging experimental and comparative approaches at different scales.

Ecology↗
USGS2006

Wind erodibility of soils at Fort Irwin, California (Mojave Desert), USA, before and after trampling disturbance: Implications for land management

Recently disturbed and 'control' (i.e. less recently disturbed) soils in the Mojave Desert were compared for their vulnerability to wind erosion, using a wind tunnel, before and after being experimentally trampled. Before trampling, control sites had greater cyanobacterial biomass, soil surface stability, threshold friction velocities (TFV, i.e. the wind speed required to move soil particles), and sediment yield than sites that had been more recently disturbed by military manoeuvres. After trampling, all sites showed a large drop in TFVs and a concomitant increase in sediment yield. Simple correlation analyses showed that the decline in TFVs and the rise in sediment yield were significantly related to cyanobacterial biomass (as indicated by soil chlorophyll a). However, chlorophyll a amounts were very low compared to chlorophyll a amounts found at cooler desert sites, where chlorophyll a is often the most important factor in determining TFV and sediment yield. Multiple regression analyses showed that other factors at Fort Irwin were more important than cyanobacterial biomass in determining the overall site susceptibility to wind erosion. These factors included soil texture (especially the fine, medium and coarse sand fractions), rock cover, and the inherent stability of the soil (as indicated by subsurface soil stability tests). Thus, our results indicate that there is a threshold of biomass below which cyanobacterial crusts are not the dominant factor in soil vulnerability to wind erosion. Most undisturbed soil surfaces in the Mojave Desert region produce very little sediment, but even moderate disturbance increases soil loss from these sites. Because current weathering rates and dust inputs are very low, soil formation rates are low as well. Therefore, soil loss in this region is likely to have long-term effects.

Earth Surface Processes and Landforms↗
USGS2006

The Piedmont landscape of Maryland: a new look at an old problem.

Both equilibrium and episodic erosion features can be recognized in the modern landscape. An equilibrium condition is suggested by adjustment of first and second order streams to rock structure and lithology, entrenchment of some streams against gneiss domes, altitudinal zonation of rock types around gneiss domes, correlation of lithology with overburden thickness on uplands, etc. The long-term episodic character of erosion is suggested by clastic wedges on the adjacent Coastal Plain, an upland of low relief that truncates non-carbonate rocks of different lithologies, isovolumetric chemical weathering of alumino-silicate rocks, clastic deposition in marble valleys, and weathering profile truncation by modern drainage. The upland surface preserved in the eastern Piedmont developed by the Late Cretaceous. In the interval from the Late Cretaceous to the Late Miocene, low input of terrigenous sediments to the Coastal Plain, dominance of marine sedimentation, and spotty evidence of saprolite formation on crystalline rocks, suggest that the Maryland Piedmont was an area of low relief undergoing intense weathering. Incised valleys were formed during a cycle of erosion probably initiated in the Late Miocene and extensive colluvial sediments were deposited on hillslopes by periglacial processes during the Pleistocene.-after Authors

Earth Surface Processes and Landforms↗
USGS2013

Large wood budget and transport dynamics on a large river using radio telemetry

Despite the abundance of large wood (LW) river studies there is still a lack of understanding of LW transport dynamics on large low gradient rivers. This study used 290 radio frequency identification tagged (RFID) LW and 54 metal (aluminum) tagged LW, to quantify the percent of in-channel LW that moves per year and what variables play a role in LW transport dynamics. Aluminum tags were installed and monitored on LW in-transit during the rising limb of a flood, the mean distance traveled by those pieces during the week was 13.3 river kilometers (km) with a maximum distance of 72 km. RFID tagged LW moved a mean of 11.9 km/yr with a maximum observed at 101.1 km/yr. Approximately 41% of LW low on the bank moves per year. The high rate of transport and distance traveled is likely due to the lack of interaction between LW floating in the channel and the channel boundaries, caused primarily by the width of the channel relative to length of the LW. Approximately 80% of the RFID tags moved past a fixed reader during the highest 20% of river stage per year. LW transport and logjam dynamics are complicated at high flows as pieces form temporary jams that continually expand and contract. Unlike most other studies, key members that create a logjam were defined more by stability than jam size or channel/hydrologic conditions. Finally, using an existing geomorphic database for the river, and data from this study, we were able to develop a comprehensive LW budget showing that 5% of the in-channel LW population turns over each year (input from mass wasting and fluvial erosion equals burial, decomposition, and export out of system) and another 16% of the population moving within the system.

North Carolina;Virginia↗
USGS2015

Measurements of the initiation of post-wildfire runoff during rainstorms using in situ overland flow detectors

Overland flow detectors (OFDs) were deployed in 2012 on a hillslope burned by the 2010 Fourmile Canyon fire near Boulder, Colorado, USA. These detectors were simple, electrical resistor-type instruments that output a voltage (0–2·5 V) and were designed to measure and record the time of runoff initiation, a signal proportional to water depth, and the runoff hydrograph during natural convective rainstorms. Initiation of runoff was found to be spatially complex and began at different times in different locations on the hillslope. Runoff started first at upstream detectors 56% of the time, at the mid-stream detectors 6%, and at the downstream detectors 38% of the time. Initiation of post-wildfire runoff depended on the time-to-ponding, travel time between points, and the time to fill surface depression storage. These times ranged from 0·5–54, 0·4–1·1, and 0·2–14 minutes, respectively, indicating the importance of the ponding process in controlling the initiation of runoff at this site. Time-to-ponding was modeled as a function of the rainfall acceleration (i.e. the rate of change of rainfall intensity) and either the cumulative rainfall at the start of runoff or the soil–water deficit. Measurements made by the OFDs provided physical insight into the spatial and temporal initiation of post-wildfire runoff during unsteady flow in response to time varying natural rainfall. They also provided data that can be telemetered and used to determine critical input parameters for hydrologic rainfall–runoff models.

Earth Surface Processes and Landforms↗
USGS2018

A regime shift in sediment export from a coastal watershed during a record wet winter, California: Implications for landscape response to hydroclimatic extremes

Small, steep watersheds are prolific sediment sources from which sediment flux is highly sensitive to climatic changes. Storm intensity and frequency are widely expected to increase during the 21st century, and so assessing the response of small, steep watersheds to extreme rainfall is essential to understanding landscape response to climate change. During record winter rainfall in 2016–2017, the San Lorenzo River, coastal California, had nine flow peaks representing 2–10‐year flood magnitudes. By the third flood, fluvial suspended sediment showed a regime shift to greater and coarser sediment supply, coincident with numerous landslides in the watershed. Even with no singular catastrophic flood, these flows exported more than half as much sediment as had a 100‐year flood 35 years earlier, substantially enlarging the nearshore delta. Annual sediment load in 2017 was an order of magnitude greater than during an average‐rainfall year, and 500‐fold greater than in a recent drought. These anomalous sediment inputs are critical to the coastal littoral system, delivering enough sediment, sometimes over only a few days, to maintain beaches for several years. Future projections of megadroughts punctuated by major atmospheric‐river storm activity suggest that interannual sediment‐yield variations will become more extreme than today in the western USA, with potential consequences for coastal management, ecosystems, and water‐storage capacity. The occurrence of two years with major sediment export over the past 35 years that were not associated with extremes of the El Niño Southern Oscillation or Pacific Decadal Oscillation suggests caution in interpreting climatic signals from marine sedimentary deposits derived from small, steep, coastal watersheds, to avoid misinterpreting the frequencies of those cycles.

California↗
USGS2003

Processes and rates of sediment and wood accumulation in headwater streams of the Oregon Coast Range, USA

Channels that have been scoured to bedrock by debris flows provide unique opportunities to calculate the rate of sediment and wood accumulation in low-order streams, to understand the temporal succession of channel morphology following disturbance, and to make inferences about processes associated with input and transport of sediment. Dendrochronology was used to estimate the time since the previous debris flow and the time since the last stand-replacement fire in unlogged basins in the central Coast Range of Oregon. Debris flow activity increased 42 per cent above the background rate in the decades immediately following the last wildfire. Changes in wood and sediment storage were quantified for 13 streams that ranged from 4 to 144 years since the previous debris flow. The volume of wood and sediment in the channel, and the length of channel with exposed bedrock, were strongly correlated with the time since the previous debris flow. Wood increased the storage capacity of the channel and trapped the majority of the sediment in these steep headwater streams. In the absence of wood, channels that have been scoured to bedrock by a debris flow may lack the capacity to store sediment and could persist in a bedrock state for an extended period of time. With an adequate supply of wood, low-order channels have the potential of storing large volumes of sediment in the interval between debris flows and can function as one of the dominant storage reservoirs for sediment in mountainous terrain.

Oregon↗
USGS2019

Measurement method has a larger impact than spatial scale for plot-scale field-saturated hydraulic conductivity (Kfs) after wildfire and prescribed fire in forests

Abstract Wildfires raise risks of floods, debris flows, major geomorphologic and sedimentologic change, and water quality and quantity shifts. A principal control on the magnitude of these changes is field-saturated hydraulic conductivity (Kfs), which dictates surface runoff generation and is a key input into numerical models. This work synthesizes 73 Kfs datasets from the literature in the first year following fire at the plot scale (≤ 10 m2). A meta-analysis using a random effects analysis showed significant differences between burned and unburned Kfs. The reductions in Kfs after fire, expressed by the ratio of Kfs Burned / Kfs Unburned, were 0.46 (95% confidence interval of 0.31-0.70) combining wildfire and prescribed fire and 0.3 (95% confidence interval of 0.13-0.71) for wildfire. No significant differences for Kfs were observed between wildfire and prescribed fire or moderate and high fire severity. Both Kfs magnitude and variability depended more on measurement method than measurement support area at the plot scale, with methods applying head ≥0.5 cm producing larger estimates of Kfs. It is recommended that post-fire efforts to characterize Kfs for modeling or process-based interpretations use methods that reflect the dominant infiltration processes: tension infiltrometers and simulated rainfall methods when soil matrix flow dominates and ponded head methods when macropore flow is critical.

Earth Surface Processes and Landforms↗
USGS2024

A framework to facilitate development and testing of image-based river velocimetry algorithms

Image-based methods have compelling, demonstrated potential for characterizing flow fields in rivers, but algorithms like particle image velocimetry (PIV) must be further tested and improved to enable more effective use of these techniques. This paper presents a framework designed for this exact purpose: Simulating Hydraulics and Images for Velocimetry Evaluation and Refinement (SHIVER). The approach involves coupling a hydrodynamic model with a synthetic particle generator to advect particles between frames, as dictated by local velocity vectors and thus construct a plausible image sequence specific to the reach of interest. The resulting time series can then be used as input to a velocimetry algorithm to compare image-derived estimates with known (modelled) velocities to perform an exhaustive, spatially distributed accuracy assessment. As an example application of SHIVER, we examined the effects of interrogation area (IA) size, frame rate, flow velocity, and image sequence duration on the performance of a standard PIV algorithm. This analysis indicated that image-derived velocities were generally in close agreement with those from the flow model (root mean square error <10% and mean bias <3%), except when small IAs were coupled with low frame rates. Velocity estimates were most accurate for the lowest modelled discharge ( R 2 =0.97 at baseflow) and became less reliable as the mean flow velocity increased ( R 2 =0.92 for an intermediate discharge and R 2 =0.86 at bankfull). Accuracy was essentially independent of image sequence duration, implying that long occupations might not be necessary. Errors were concentrated along channel margins, where PIV-based velocities tended to be greater than those from the flow model. Small IAs led to underpredictions of velocity, while larger IAs led to overpredictions. SHIVER is highly modular and could be updated to make use of different hydrodynamic models or image simulators. The framework could also facilitate more thorough sensitivity analyses and comparison of various velocimetry algorithms.

Earth Surface Processes and Landforms↗
USGS2024

Debris-flow entrainment modelling under climate change: Considering antecedent moisture conditions along the flow path

Debris-flow volumes can increase along their flow path by entraining sediment stored in the channel bed and banks, thus also increasing hazard potential. Theoretical considerations, laboratory experiments and field investigations all indicate that the saturation conditions of the sediment along the flow path can greatly influence the amount of sediment entrained. However, this process is usually not considered for practical applications. This study aims to close this gap by combining runout and hydrological models into a predictive framework that is calibrated and tested using unique observations of sediment erosion and debris-flow properties available at a Swiss debris-flow observation station (Illgraben). To this end, hourly water input to the erodible channel is predicted using a simple, process-based hydrological model, and the resulting water saturation level in the upper sediment layer of the channel is modelled based on a Hortonian infiltration concept. Debris-flow entrainment is then predicted using the RAMMS debris-flow runout model. We find a strong correlation between the modelled saturation level of the sediment on the flow path and the channel-bed erodibility for single-surge debris-flow events with distinct fronts, indicating that the modelled water content is a good predictor for erosion simulated in RAMMS. Debris-flow properties with more complex flow behaviour (e.g., multiple surges or roll waves) are not as well predicted using this procedure, indicating that more physically complete models are necessary. Finally, we demonstrate how this modelling framework can be used for climate change impact assessment and show that earlier snowmelt may shift the peak of the debris-flow season to earlier in the year. Our novel modelling framework provides a plausible approach to reproduce saturation-dependent entrainment and thus better constrain event volumes for current and future hazard assessment.

Illgraben basin↗
USGS2024

Wildfire, extreme precipitation and debris flows, oh my! Channel response to compounding disturbances in a mountain stream in the Upper Colorado Basin, USA

Compounding changes in climate and human activities stand to increase sediment input to rivers in many landscapes, including via discrete perturbations such as post-fire debris flows. Because sediment supply is a dominant control on river morphology, understanding mountain river responses to sediment regime perturbations is critical to predicting and addressing downstream effects to infrastructure, water security and aquatic habitat. A growing body of literature explores the causes, likelihood, size and composition of post-fire debris flows, but the channel response to these disturbances remains poorly studied. This study used repeat field surveys, time-lapse photographs and pre- and post-disturbance remote sensing datasets to document and analyse space- and time-varying channel response to post-fire debris flows along a steep mountain stream in the Upper Colorado River Basin, USA. Specifically, we evaluated channel morphology and bed composition changes, correlations between channel changes and valley and channel attributes, and the relative importance of spring snowmelt versus summer monsoon events. Several cross-sectional channel change types were observed from lidar a month after post-fire debris-flow events, including channelized and braided incision into deposits, incision into the pre-fire channel bed, bank erosion and no change. Channel changes were most correlated with pre-fire channel width, valley width and unit stream power, and these relationships could be tested in other burned locations to evaluate their transferability. Repeat channel surveys before and after snowmelt indicate rapid recovery and channel narrowing following major sediment disturbances, although sediment deposits remained in the channel margins. Together, these results highlight the importance of field and remote sensing-based channel surveys to improve understanding of, and potential to predict, mountain channel response to compounding climate disturbances.

Earth Surface Processes and Landforms↗
USGS2024

Redistribution of debris-flow sediment following severe wildfire and floods in the Jemez Mountains, New Mexico, USA

Severe fire on steep slopes increases stormwater runoff and the occurrence of runoff-initiated debris flows. Predicting locations of debris flows and their downstream effects on trunk streams requires watershed-scale high-resolution topographic data. Intense precipitation in July and September 2013 following the June 2011 Las Conchas Fire in the Jemez Mountains, New Mexico, led to widespread debris flows in the watershed of Rito de los Frijoles. We differenced lidar Digital Elevation Models (DEMs) collected in 2010 and 2016 to map subwatersheds experiencing debris flows and changes in elevation of the trunk stream. Debris flow occurrence was well predicted by previous assessments of debris-flow hazard; debris flows occurred in 7 of 9 sub-basins where the debris-flow hazard was above 60% for the 25-year rainfall event, and in 0 of 21 basins where debris flow hazard was less than 60%. Debris flows resulted in fan deposition at the confluence with the trunk stream followed by transport during three documented floods. The bed of the 22 km trunk stream increased in elevation by a mean of 0.29 m, but the local change in thalweg elevation was controlled by inputs of water and sediment and longitudinal variation in gradient. Downstream of the mouths of tributaries with debris flows, the thalweg of the trunk stream rose as much as 2 m. Downstream of the mouths of tributaries without debris flows the thalweg of the main stem degraded by as much as 2 m, mobilizing sediment that was then deposited further downstream where the gradient of the trunk stream decreases. In conclusion, the transport of sediment generated by debris flows was predictably related to spatial variation in sediment supply, discharge and gradient.

New Mexico↗
USGS2026

The contribution of a surge event to infilling in a barrier-enclosed estuary: Insights from field observations

Many estuaries worldwide face increasing sediment loading caused by catchment land use change and intensification, creating subsequent adverse effects on estuarine ecosystems. Extreme weather events can disproportionately alter sediment pathways and loading. Although storm-driven sediment exchange has been widely examined at open coasts and inlets, key transport mechanisms within constricted, sheltered estuaries remain understudied. This study presents an observational dataset capturing the impact of a 99th percentile water-level event (based on 20 years of records) on sediment transport pathways in a sheltered, barrier-enclosed estuary. This event, driven by a 3-day storm surge (>0.5 m) combined with a spring tide, was recorded during a 3-week field campaign. Sediment transport pathways and riverine contributions were analysed, and observations revealed substantial changes in suspended sediment concentrations increasing from 18 mg/l to 70 mg/l during the event. The elevated water levels and resulting pressure gradient at the constricted study site entrance caused by the storm surge increased local flood dominance. Combined with higher flow velocities and resuspension, the storm led to a sixfold increase in sediment import at the estuary entrance and a 600-fold increase in sediment flux to the upper estuary. The decoupling of peak suspended sediment concentrations from streamflow indicates that the resuspension of estuarine legacy sediment, rather than catchment inputs, dominated the system's response. These findings challenge assumptions about estuarine sediment budgets and emphasise that incorporating high water-level surge events into models can enhance the prediction of long-term estuarine evolution. Given projected increases in storm frequency under climate change, understanding these episodic but highly consequential sediment pulses can support the assessment of wetland resilience and inform estuarine management strategies.

Whangateau Estuary↗
USGS2026

Deep learning models for mapping surficial geology in selected physiographic regions of New York

Surficial geologic mapping is required for decisions including infrastructure and water-resource development and management. Deep learning, a type of machine learning that uses training data to self-learn and perform tasks, is explored as a tool to help increase efficiency of the labour- and time-intensive mapping process. Deep learning models were trained, and their potential to aid in mapping surficial geology was explored for two physiographic regions in New York: the high-relief Allegheny Plateau and the low-relief Erie-Ontario Lowlands. Key to the development of deep learning models is the availability of highly detailed surficial geologic maps in each of these regions from which the models can learn. Through exploring different groupings of surficial deposits and their corresponding spatial data, deep learning models were iteratively developed to reproduce published mapping for greater than 79% of the training areas, and with similar accuracy in test areas within the same physiographic region. These models were trained using only two inputs: high-resolution lidar data and previously published surficial geologic maps. This straightforward approach is intended to make the methods and models created easily reproducible using widely accessible datasets. This research describes the strengths and limitations of lidar-derived surficial geologic models and how broadly grouping surficial types, characteristics of different physiographic regions, and testing models in physiographic regions outside of their training areas affect model creation and performance. In addition, the models created could be used as a tool for mapping surficial geology in similar physiographic regions, in addition to establishing a framework for creating similar models elsewhere.

New York↗
USGS2026

Ground-motion aleatory-variability models for Puerto Rico and the U.S. Virgin Islands

I develop independent logic trees for aleatory variability for crustal and subduction-zone (interface and intraslab) earthquakes for seismic hazards analyses in Puerto Rico and the U.S. Virgin Islands (PRVI) from existing suites of ground-motion models (GMMs) and from ground-motion datasets, including a regional PRVI dataset. The aleatory variability models are parameterized using a partially nonergodic partitioning of standard deviation that consists of independently developed between-event ( ), site-to-site ( ), and event-corrected single-station ( ) standard deviation components. The effects of nonlinear site response on aleatory variability are incorporated through additional terms that modify the standard deviation components. Because one goal of this work is to develop independent logic trees for aleatory variability that synthesize the aleatory variability models from GMMs, I make use of the functional forms of the input GMMs. The PRVI dataset contains a limited number of stations with high-quality site metadata and does not contain records from earthquakes with magnitudes greater than 6.1, so I choose not to develop the aleatory variability models from the regional dataset alone. Instead, the standard deviation components from regional ground-motion data are evaluated against the components derived from GMMs and from available global datasets, and regionalized standard deviation components are incorporated where there is evidence that regional effects exhibit substantial differences. The resulting logic trees for aleatory variability consist of models of and that are consistent with semiempirical GMMs for active crustal and subduction-zone regimes, and two alternative models of , including one model that exhibits site-to-site variability informed by PRVI data, with values that exceed global models. The aleatory variability models may be considered in future hazards assessments in PRVI to simplify the hazard calculations, to incorporate regional ground-motion variability effects, and to enable direct logic-tree weighs of aleatory variability.

Puerto Rico, U.S. Virgin Islands↗