USGS Science⌕ Search

SEARCH · USGS Science

Results for “Ecology of Freshwater Fish”

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.

1,661 records · Page 6Linked to original sources
USGS2021

New York and Landsat

From the iconic skyline of New York City to the forested landscapes of the Adirondack Mountains and the countryside of the Allegheny Plateau, the State of New York is overflowing with diversity and life. Bordered by the Atlantic Ocean on the east and two of the Great Lakes to the north and west, New York has more than 7,600 lakes, ponds, and reservoirs and more than 70,000 miles of rivers and streams. New York’s stewardship of its freshwater resources is fundamental to the health and well-being of all who work at, reside in, and visit the State’s landmarks and places. Harmful algal blooms in the State’s waterbodies are a growing concern and threaten the health of the region and its inhabitants. Images and data from Landsat satellites continue to provide critical information to scientists, public health officials, and resource managers who are studying the effects and risks of the problem.

New York↗
USGS2026

Incorporating location uncertainty improves inference with stop-level North American Breeding Bird Survey data

Ecological models should account for uncertainty to be most effective and useful. Yet, uncertainty from model covariates—unlike that from other sources, such as sampling error or process variability—is seldom explicitly incorporated. This can cause underestimates of uncertainty to cascade through model parameter estimates, predictions, and downstream uses. Burner et al. proposed a method for quantifying uncertainty in covariates and incorporating it into models using informative Bayesian priors. This method was applied to stop-level Breeding Bird Survey (BBS) analyses, where land cover uncertainty at each stop arises from substantial stop location uncertainty. A limited validation of model-estimated land cover, using stops with known locations, indicated the method’s potential effectiveness, but it was not rigorously evaluated. We conduct a robust simulation-based test, generating stop locations, extracting land cover, and simulating bird communities across 210 BBS routes in the upper Midwest. We compare 3 models: a “known” model with true land cover, a “naive” model assuming consistent 800-m stop spacing, and a “full” model using informative priors to estimate land cover. Species parameter estimates and predicted prevalence patterns across gradients in land cover from the full model approached those of the known model and were substantially closer to the true values used in simulations relative to those from the naive model. Naive model parameters were more biased relative to the other models, and credible intervals of predicted species prevalence rarely included the true simulated values. The full model also produced land cover covariate estimates closer to true simulation values relative to the mean informative priors. Our results show that, for the BBS, informative priors enable more accurate stop-level analyses despite location uncertainty. In contrast, naive models that ignore this uncertainty yield poor inferences. More broadly, we demonstrate empirically the utility of informative priors to account for covariate uncertainty in ecological models.

Michigan, Minnesota, Wisconson↗
USGS2026

Restoring dryland water cycles for precipitation feedback and climate stability; a review

Drylands across the globe are experiencing intensifying water scarcity, land degradation, and hydroclimatic extremes. This review integrates evidence from multidecadal field studies, hydrologic monitoring, geomorphic and ecological assessments, remote sensing, and land–atmosphere science to evaluate how restoration influences key components of the terrestrial water cycle. Low-tech natural infrastructure in dryland streams (NIDS)—including check dams, leaky weirs, one-rock dams, and gabions—has emerged as a promising but under-synthesized nature-based solution for restoring hydrologic function in these environments. We describe the mechanisms through which these interventions modify runoff detention, infiltration, sediment and alluvial storage, shallow-groundwater recharge, vegetation recovery, and surface-energy partitioning, and we summarize outcomes across diverse dryland settings. Findings consistently show increased water residence time, enhanced soil-moisture storage, expanded riparian vegetation, extended flow duration, and shifts toward greater latent-heat flux—producing localized cooling and strengthened ecohydrological feedbacks. Building on these localized effects, we articulate a hypothesis that links the spatial extent of restoration, the density of NIDS per unit drainage area, and the magnitude of the latent-to-sensible-heat contrast generated by wetter post-rainfall conditions. Specifically, we hypothesize that when NIDS are implemented at densities permitted by topography and across areas large enough to maintain elevated soil moisture after storm events, the resulting increases in latent heat flux, surface cooling, and boundary-layer moistening may enhance moisture convergence and boundary-layer development, potentially increasing the likelihood or stability of convective precipitation, analogous to how reductions in these processes have contributed to regional drought intensification. These land–atmosphere feedbacks remain untested at scale but represent an important research Frontier. By integrating hydrologic, geomorphic, ecological, and atmospheric perspectives, this review provides a comprehensive framework for considering how low-tech, landscape-scale interventions can strengthen watershed resilience and contribute to climate-relevant nature-based solutions.

Arizona↗
USGS2026

Nonstationary demographic state-space models using unreplicated counts for species undergoing environmental stressors

A fundamental task in ecological statistics is to estimate abundance and growth rate distributions from wildlife monitoring data to inform conservation management. Modeling time series of wildlife populations presents a number of challenges from both statistical and ecological perspectives, including discreteness; lack of replication; nonstationarity; and observation, demographic, and other phenomenological processes. Nonstationary dynamics are often exhibited by populations undergoing environmental stressors. Models must account for these characteristics to produce reliable estimates of abundance and trends, yet estimation can be challenging with unreplicated data. We propose nonstationary demographic state-space models using unreplicated counts for populations undergoing environmental stressors. A reduced growth rate model matches the complexity of the unreplicated count data, and a fecundity bound on growth rate distributions allows the separation of processes affecting growth rates like environmental stressors from those affecting abundance external to growth rates like migration. NDSSMs allow for the embedding of nonstationary model components, and we explore the use of changepoints, volatility clustering, and migration processes. We apply the proposed nonstationary models in case studies of herons affected by predator/competitor reestablishment and three bat species affected by a fungal pathogen causing white-nose syndrome. Nonstationary models outperform stationary models and generalized linear mixed effects models according to model scoring and visual inspection of predictions, and provide estimates more consistent with published values. Incorporating migration improves model fit universally, even with approximate one-way immigration, most likely because populations are extirpated, recolonized, and increase multiple-fold over the upper bound set by species fecundity. In addition, estimates of the timing and severity of the environmental stressor differed for models with migration. Including nonstationary and demographic components in a fecundity-bounded growth rate model improves inference and benefits interpretability of hyperparameters. In turn, this adjusts uncertainties in predictions of abundance and growth rates over time, providing the ingredients needed for informed conservation analysis and for directing future monitoring of at-risk species.

Journal of Agricultural, Biological and Environmen↗
USGS2026

The role of groundwater in contributing to surface water salinization in the Upper Colorado River Basin

Freshwater salinization impacts the availability of water for human use and ecosystem needs worldwide. It has been estimated that total dissolved solids (TDS) in the Colorado River Basin cause $350 million/year in damages and substantial resources are devoted to reducing TDS loading to streams. This study describes the development and application of coupled watershed models that enable TDS source tracking through the subsurface and across the landscape at a seasonal timestep for 35 years in the Upper Colorado River Basin. Results indicate that, on average, 75% of TDS loading to streams originates as baseflow, and 50% of loading is lagged in delivery by longer than one season. Snowmelt was identified as a dominant process controlling the transport of lagged TDS to streams. This approach informs when and where TDS mitigation efforts may be effective in a watershed that serves as a critical water supply for the southwestern United States.

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

Late Miocene Colorado River arrival in the Bidahochi basin supports spillover origin of Grand Canyon

The timing and mechanism of the integration of the Colorado River and incision of the Grand Canyon remain among geology’s enduring controversies. A key question is the configuration of the upper Colorado River watershed between 11 and 6 million years ago. In this study, we present new evidence from zircon uranium-lead geochronology for the arrival of distinctive Colorado–Green River sediment in the Bidahochi basin by 6.6 million years ago derived from the Browns Park Formation. This is coeval with an order-of-magnitude increase in depositional rate, an increase in carbonate strontium isotope ( 87 Sr/ 86 Sr) ratios, the appearance of large fish species characteristic of fast-flowing waters, and other sedimentological changes. This evidence is consistent with the Colorado River supplying water and sediment to the Bidahochi basin before spillover integration of the river through the Grand Canyon.

Arizona, Nevada, Utah↗
USGS2024

Metal mobilization from thawing permafrost to aquatic ecosystems is driving rusting of Arctic streams

Climate change in the Arctic is altering watershed hydrologic processes and biogeochemistry. Here, we present an emergent threat to Arctic watersheds based on observations from 75 streams in Alaska’s Brooks Range that recently turned orange, reflecting increased loading of iron and toxic metals. Using remote sensing, we constrain the timing of stream discoloration to the last 10 years, a period of rapid warming and snowfall, suggesting impairment is likely due to permafrost thaw. Thawing permafrost can foster chemical weathering of minerals, microbial reduction of soil iron, and groundwater transport of metals to streams. Compared to clear reference streams, orange streams have lower pH, higher turbidity, and higher sulfate, iron, and trace metal concentrations, supporting sulfide mineral weathering as a primary mobilization process. Stream discoloration was associated with dramatic declines in macroinvertebrate diversity and fish abundance. These findings have considerable implications for drinking water supplies and subsistence fisheries in rural Alaska.

Alaska↗
USGS2026

Groundwater spatial variability within an atoll island: Assessing shallow aquifer heterogeneity with geophysical and physicochemical measurements

This study examines the spatial variability of shallow groundwater on Dhigelaabadhoo Island using electromagnetic induction surveys, groundwater monitoring, and sediment analyses. The research reveals how variations in island morphology—such as differences in elevation, reef flat width, and sediment composition—affect the spatial distribution of groundwater lenses and the overall aquifer dynamics. Saltwater intrusion is especially pronounced in low elevated areas, with narrow reef flat plate and areas where higher hydraulic conductivity—driven by the presence of coarser sediments—is observed, whereas regions characterized by finer sediments, higher elevation, and wider reef flat plates tend to support more symmetric and less saline groundwater lenses. The geophysical investigations reveal that tidal oscillations alter groundwater movement by markedly changing water levels and conductivity, thereby underscoring the critical need to account for temporal dynamics in atoll coastal aquifer systems and the importance of integrating tidal dynamics into the aquifer zone. The findings highlight the significant role of intrinsic morphological and external hydrodynamic factors in shaping groundwater distribution on atoll islands, offering critical insights for sustainable freshwater resource management.

Dhigelaabadhoo Island↗
USGS2024

Simulation of groundwater flow in the Long Island, New York regional aquifer system for pumping and recharge conditions from 1900 to 2019

The U.S. Geological Survey has developed a transient, groundwater-flow model that simulates hydrologic conditions in the Long Island aquifer system as part of an ongoing (since 2016) multiyear, cooperative investigation with the New York State Department of Environmental Conservation. The goals of this investigation are to assist stakeholders and resource managers to evaluate the response of the hydrologic system to changes in future hydraulic stresses. Responses in the hydrologic system include changes in water levels in the hydrogeologic units; discharge to streams, coastal waters, and subsurface infrastructure; and the extent of saline groundwater in the aquifers. Hydraulic stresses include future water-supply management and changes in land use and infrastructure. The numerical model synthesizes a diverse set of physiographic, geologic, climatic, land-use, and historical population, water use, and infrastructure data to physically represent the Long Island aquifer system from land surface to bedrock and to simulate annual hydrologic conditions between 1900 and 2019. A three-dimensional hydrogeologic framework was developed from existing and recently collected borehole geologic and geophysical data collected as part of a companion drilling program. Water-transmitting properties of the principal aquifer sediments were defined in three dimensions from new and existing lithologic logs. The distribution of recharge from precipitation was estimated from landscape characteristics and climate data. Anthropogenic recharge from wastewater, leaky infrastructure, and storm runoff were estimated from population, infrastructure, and pumping data. Water-use data, including well locations, depths, and pumping rates, were obtained from historical sources and records and used to estimate pumping stresses continuously in time and space, at an annual average time scale. The data were incorporated into a three-dimensional numerical model using the U.S. Geological Survey finite difference modeling code MODFLOW 6; the model encompassed all of Long Island and surrounding surface waters and simulated historical hydrologic conditions from 1900 to 2019. The calibration process involved trial and error adjustments using prior knowledge to improve general fit to observations followed by an inverse calibration to update and optimize input parameters, using an iterative ensemble smoother algorithm implemented in PEST++ version 5.0. This resulted in a model that generally was in good agreement with observed, dynamically varying hydrologic conditions from 1900 to 2019. The calibrated model was used to develop two base-case models for scenario testing of future, hypothetical conditions where one represented average-annual conditions, and one represented average-seasonal conditions from 2010 to 2019. The model representing average-annual conditions was modified further to represent an alternate sea-level position of 6 feet above the North American Vertical Datum of 1988, and the model representing average-seasonal conditions was modified to represent the average seasonal effects of a 5-year drought imposed upon current hydrologic conditions. Recharge is the sole source of water to the aquifer system; groundwater discharges to coastal water and streams and is withdrawn by pumped wells. Model-estimated annual recharge ranged from about 11 inches in 1965 to 41 inches in 1983. On average, from 2010 to 2019, about 23 percent of water was pumped from wells, and about 47 and 27 percent discharged to coastal waters and streams, respectively; the remaining 4 percent was water that moved into storage in the aquifer matrix. Water levels on Long Island vary naturally during time in response to changes in recharge; the amount of variation is largest in the interior of the island, in areas with highest water table altitudes near groundwater divides and lowest near streams and the coastal waters. The total range of water table altitudes on Long Island between 1900 and 2019 ranged from near 0 to more than 70 feet in western parts of Long Island. The largest range in altitudes is in New York City and is associated with areas of large historical withdrawals between the 1920s and the late 1980s. Water table altitudes generally varied by less than 10 feet in eastern Suffolk County, where the aquifer is under more natural conditions. Saltwater intrusion is of great concern on Long Island, particularly in western Long Island where both the unconfined and confined parts of the aquifer system have been intruded in response to large-scale groundwater withdrawals; however, the volume of freshwater in the islandwide aquifer system only has changed by about 5 percent between 1900 and 2019. The decadal change in the freshwater volume was largest during the early and mid-20th century, corresponding to the largest historical pumping, but that volume change did not exceed 1 percent. The negligible change in freshwater volume suggests that saltwater intrusion as of 2019 was limited at an islandwide scale but continues to occur in local areas of Queens and Nassau Counties, adversely affecting current water supplies and limiting future water supplies for affected communities. The regional groundwater model developed for this investigation is a tool that can be used to help determine the viability of current and future water supplies at a regional scale and can be used to support development of additional models at finer scale to support more focused assessments of groundwater sustainability.

New York↗
USGS2025

Tracing mercury from land to river: Global sources, retention, and implications for sustainability

Mercury (Hg) pollution in river systems is a global sustainability challenge. Yet the transport, transformation, and retention of Hg within global rivers remain poorly quantified, particularly in regions with sparse observations such as Southeast Asia and Africa, hindering effective pollution mitigation and reinforcing geographic inequities in scientific knowledge and environmental governance. Here, we present the first global, high-resolution simulation of riverine Hg dynamics using a process-based model that traces Hg from land-based sources through river networks to the ocean. Under a realistic scenario, we estimate that ~1,900 megagrams per year (Mg/yr) of Hg enters global rivers, including 1,500 Mg/yr from human-induced sources and 400 Mg/yr from soil erosion. Nearly half of this flux (~1,000 Mg/yr) is retained in reservoirs and dams, which act as major sinks. While such retention limits downstream delivery to the oceans, it also heightens in-reservoir Hg methylation risks. By bridging the gap between Hg releases and observed riverine exports, our framework offers a scalable tool for data-limited regions, promotes data access, and supports global freshwater and pollution-management strategies.

EarthArXiv↗
USGS2026

Tracing mercury from land to river: Global sources, retention, and implications for sustainability

Mercury (Hg) pollution in river systems is a global sustainability challenge. Yet the transport, transformation, and retention of Hg within global rivers remain poorly quantified, particularly in regions with sparse observations such as Southeast Asia and Africa, hindering effective pollution mitigation and reinforcing geographic inequities in scientific knowledge and environmental governance. Here, we present the first global, high-resolution simulation of riverine Hg dynamics using a process-based model that traces Hg from land-based sources through river networks to the ocean. Under a realistic scenario, we estimate that ∼1900 megagrams per year (Mg/yr) of Hg enters global rivers, including 1500 Mg/yr from human-induced sources and 400 Mg/yr from soil erosion. Nearly half of this flux (∼1000 Mg/yr) is retained in reservoirs and dams, which act as major sinks. While such retention limits downstream delivery to the oceans, it also heightens in-reservoir Hg methylation risks. By bridging the gap between Hg releases and observed riverine exports, our framework offers a scalable tool for data-limited regions, promotes data access, and supports global freshwater and pollution-management strategies.

Environmental Science & Technology↗
USGS2025

Constraining landslide frequency across the United States to inform county-level risk reduction

Informative landslide hazard estimates are needed to support landslide mitigation strategies to reduce landslide risk across the United States. Whereas existing national-scale landslide susceptibility products assess where landslides are likely to occur, they do not address how often , which is a critical element of landslide hazard and risk assessments. In particular, the U.S. Federal Emergency Management Agency's National Risk Index (NRI) requires landslide frequency estimates to inform expected annual loss estimates. We present county-level landslide frequency (landslides per area per year) estimates for the 50 US states. We applied Bayesian negative binomial regression to estimate both the expected (average) reported landslide frequency and full distribution of annual landslide counts for each county. We compared a suite of models that used combinations of landslide-susceptible area, probability of potentially triggering earthquakes, frequency of potentially triggering precipitation, and ecological region as predictors. We trained our models with landslide inventory data from counties with the most comprehensive records available nationwide and used zero-inflated negative binomial distributions as an incompleteness model to correct for temporal reporting gaps. We selected a preferred frequency model to inform the NRI based on information criteria and physically plausible parameter estimates. The model showed that average annual reported landslide frequencies vary by 5 orders of magnitude across US counties, ranging from 0.002 (0.00015–0.05) landslides 1000 km −2 yr −1 in Kusilvak Census Area, Alaska, to 29 (19–46) landslides 1000 km −2 yr −1 in Lake County, California, reflecting the country's strong variations in landslide susceptibility, earthquake probability, and other factors for which ecological region serves as a proxy. Counties with estimated frequencies in the top 20 % of all counties are predominately along the West Coast of the continental United States, in mountainous regions of the Pacific Northwest and Intermountain West, in locally steep or earthquake-prone regions of the Midwest and Southeast, along the Appalachians, in southern and southeastern Alaska, and on some Hawaiian islands. By examining the number of landslides predicted in 99th percentile years for each county, we identified that 26 % of US counties likely have potential for widespread landsliding with more than 10 landslides 1000 km −2 yr −1 , even when such large events have not been reported in the training data for that county. Overall, our results better represent the range of possible landslide frequencies and spatial variations than previous national-scale estimates reported in the NRI, and our approach can inform other risk-reduction and loss-mitigation efforts across the United States and globally.

Natural Hazards and Earth System Sciences↗
USGS2022

South Carolina and Landsat

South Carolina, the eighth State admitted to the union, transcends its size with its deep, rich history; striking beauty; vast natural resources; and extensive cultural diversity. Home to part of the Blue Ridge Mountains of the Central Appalachians, the Upstate is graced with more than 100 waterfalls, while the Lowcountry borders the Atlantic Ocean with 187 miles of coastline and 35 barrier islands. Forests cover two-thirds of the State, and forestry and agriculture together, as agribusiness, make up South Carolina’s leading industry. Two historic crops—cotton and tobacco—still rank in the top 10 commodities, though corn and soybeans now rank higher. Poultry, cattle, peanuts, and flowers also make the list. South Carolina’s population totals more than five million. Other residents include a variety of wildlife, bird, reptile, and fish species, including Ursus americanus (black bears), Alligator mississippiensis (American alligators), and Tursiops truncatus (bottlenose dolphins). More than 100 tree species also reside in South Carolina, which pays homage to one with its “The Palmetto State” nickname. South Carolina’s subtropical climate, long coastline, and lower elevations make it highly susceptible to tornado and hurricane activity and coastal flooding. Projected sea-level rise is a growing concern. A view from space can help monitor and manage natural resources on the land and in rivers, marshes, and the coast. Landsat reveals not just what an area looks like now, but also insights from decades ago.

South Carolina↗
USGS2026

Per- and polyfluoroalkyl substances in waters associated with oil and gas development in the Denver Basin

Use of per- and polyfluoroalkyl substances (PFAS) in the petroleum industry could be a cause for concern due to the large volumes of produced water (PW) generated during oil and gas extraction, the reuse of these wastes in water-stressed regions, and adverse health outcomes related to PFAS exposures. However, PW PFAS characterization is nearly absent in the literature, and hydraulic fracturing (HF) chemical disclosures often omit the identities of additives as proprietary. Here we evaluate PFAS in PW samples from three petroleum wells in the Denver Basin during their first year of production. Total concentrations of targeted PFAS (Σ 40 PFAS) were < 35 ng/L in PW samples, with short-chain PFAS like perfluorobutanoic acid persisting throughout the sampled duration. Analysis of freshwater inputs for hydraulic fracturing (Σ 40 PFAS ~ 113 ng/L) and mixed fracture fluid (Σ 40 PFAS ~ 69 ng/L) indicated much of the targeted PFAS content was derived from the input water, and not from HF additives, however samples subjected to oxidation indicated the presence of PFAS precursors that would not be detected by targeted analysis. This study highlights that while PFAS content is low in the studied PWs, the potential for redistribution of PFAS in the environment may be a consideration for reuse applications.

Colorado, Nebraska, Wyoming↗
USGS2025

Observing northern high-latitude river systems to understand changes in a warming Arctic

Purpose of Review Streams and rivers are undergoing rapid change as the Arctic warms and thaws. We review recent observations in Arctic stream systems to identify ubiquitous changes and the most useful tools for observing change and exploring the underlying processes. Recent Findings Recent literature indicates increasingly significant trends in river hydrology and chemistry due to persistent warming in the Arctic and longer observational records for analysis. However, regional differences in the magnitude and direction of these trends persist. We also observe thresholds in ground thaw and surface–groundwater interactions that can impact river hydrology and chemistry. Summary Warming and thaw are occurring rapidly at high latitudes, resulting in increasing, yet variable responses in stream systems across regions and scales. These differences highlight the need for long-term records and an interdisciplinary approach to explain trends and predict future states. Stream systems respond to multiple landscape changes related to hydrology (changing precipitation and subsurface flow), geology (ground thaw dynamics), and ecology (vegetation change).

Current Climate Change Reports↗
USGS2026

Assessment of groundwater quantity and quality contributions to Lake Huron

Lake Huron, one of the five Great Lakes, borders the United States and Canada, with Michigan as the only U.S. State on its shoreline. Like other freshwater lakes, it faces water-quality challenges from nutrients and chemicals applied across its drainage basin. Although past studies focused on surface-water sources, groundwater contributions remain less understood. To address this gap, the U.S. Geological Survey, as part of the Cooperative Science and Monitoring Initiative, classified drainage basins to Lake Huron into eight hydrogeologic zones based on bedrock rock type and glacial sediment transmissivity. Utilizing existing data and empirical field data, we quantified groundwater discharge and identified areas of concern for loading of chloride and nitrate to Lake Huron. Groundwater contributions, including indirect and shoreline discharge, ranged from 5.8 to 11.5 inches annually, totaling 1.9 cubic miles and 0.09 cubic mile, respectively. Hydrogeologic zones with higher glacial sediment transmissivity yielded greater indirect groundwater discharge. Chloride levels above the U.S. Environmental Protection Agency’s 250-mg/L recommendation were mainly in the Saginaw lowlands, whereas nitrate above the 10-mg/L standard was rare—found in only 11 wells. Together, the analysis of where groundwater discharge is occurring in the Lake Huron Basin and the identification of areas with potential groundwater-quality concerns can help prioritize areas that are critical to protecting the long-term health of Lake Huron.

Michigan↗
USGS2026

Generating geochemical and mineralogy distributions of soil in the conterminous United States using Bayesian hierarchical spatial models

Characterizing geochemical and mineralogical soil distributions across large spatial extents is essential for understanding mineral resources, ecosystem processes, and environmental risks. Rasters of soil geochemical distributions for the conterminous United States, however, are limited. We present a Bayesian modeling workflow and tool for generating predictive geochemical and mineralogy distribution maps for the conterminous United States using integrated nested Laplace approximation (INLA) with the stochastic partial differential equation approach. By modeling soil geostatistical data with environmental covariates (soil properties, topography, climate, and land cover), we generate predictive distributions of soil geochemistry that can be mapped or extracted for further analyses. As an example, we model the spatial distribution of trace elements in soil relevant to vertebrate health (cobalt, copper, iron, manganese, selenium, and zinc) and provide a workflow that can be used to generate and visualize predictive distributions of 39 other major and trace elements and 21 minerals of the soil survey, supporting a variety of ecological, environmental, and agricultural applications.

MethodsX↗
USGS2025

Evaluation of daily stream temperature predictions (1979-2021) across the contiguous United States using a spatiotemporal aware machine learning algorithm

Stream temperature controls a variety of physical and biological processes that affect ecosystems, human health, and economic activities. We used 42 years (1979–2021) of data to predict daily summary statistics of stream temperature across >50,000 stream reaches in the contiguous United States using a recurrent graph convolution network. We comprehensively documented the performance – both across all reaches and by stream type (e.g., reservoir or groundwater influence) – as a baseline for future improvement. The model showed reach-level RMSE of <2 °C with 90 % prediction intervals that contain 90.7 % of observations. We also assessed how the model captured variability in ecologically relevant metrics (e.g., R 2 for annual 7-day maximum = 0.76; R 2 for days exceeding 25 °C = 0.75). This model does not outperform state-of-the-art machine learning efforts (e.g., RMSE ≤1.5 °C) due to a limited input set but does provide the most spatially complete modeling to date to support water availability assessments.

contiguous United States↗