Geology topics

Janet R. Barclay

Publications and source records attributed to Janet R. Barclay.

17 recordsLinked to original sources

USGS2026

Big data for small streams: Leveraging a coarse national model to enhance fine-resolution stream temperature predictions

Predictive stream temperature models are commonly used to address spatial and temporal gaps in observational data because stream temperature information is essential for environmental management decisions. National models capture broad patterns driven by climate, topography, and seasonality, but often lack the spatial resolution to simulate fine-scale variation in groundwater discharge, riparian shading, or tributary inflows. In contrast, watershed models may simulate finer-resolution patterns but miss regional drivers or be limited by scarcity of observations. To generate watershed-extent, fine-resolution stream temperature predictions, we developed 18 multiscale machine-learning models that leverage (a) a broad-extent, coarse-resolution stream temperature model and (b) finer-resolution observations and predictor data. The coarse-resolution model simulates daily stream temperature across the conterminous United States on the National Hydrologic Geospatial Fabric (13.7 km median reach length; 865,364 km total). The fine-resolution models simulate daily stream temperature on the National Hydrography Dataset Plus version 2 (1.5 km median; 5,192,808 km total). We compared multiscale models to seven purely broad- (national) or watershed-extent models, testing them in eight watersheds (436–985 km 2 ) with differing hydrologic settings and under realistic data availability scenarios. Especially when observations within a watershed were scarce, multiscale models generated more accurate predictions than national or watershed-specific models. A recurrent graph convolution neural network pretrained on coarse predictions and fine-tuned on watershed-specific observations was most accurate across watersheds and data scenarios. These results highlight that leveraging both coarse, national results and watershed-specific data can increase model accuracies, particularly when observations within a watershed are spatially or temporally limited.

Water Resources Research↗

USGS2026

A spatial analysis of the groundwater emergence flood hazard in Long Island, New York, and coastal areas surrounding Long Island Sound in New York, Connecticut, and Rhode Island

Long Island, New York, and nearby coastal areas surrounding Long Island Sound are densely populated and, like other coastal areas, are susceptible to flooding from several potential sources, including stormwater from precipitation events, tidal flooding and storm surge, and groundwater emergence flooding. Groundwater emergence fooding refers to the intersection of a rising water table with land surface or critical infrastructure. Many studies of flood drivers either neglect or only briefly discuss how shallow groundwater conditions may contribute to or exacerbate flood conditions. As a part of a comprehensive study of compound flood hazards in the coastal areas of Long Island and Long Island Sound, a spatial analysis was completed by the U.S. Geological Survey, in cooperation with the Environmental Protection Agency’s Long Island Sound Partnership, formerly called the Long Island Sound Study, using available regional datasets to characterize the potential hazard for groundwater emergence flooding. The approximately 3,100 square mile study area was subdivided into 11,407 grid cells that were 900 meters by 900 meters (approximately 3,000 feet by 3,000 feet), for the purpose of integrating the spatial datasets to calculate and map the groundwater emergence flood hazard. The depth to the water table, hydrologic soil groups, and National Land Cover Database were harmonized to the common grid. A groundwater emergence flood hazard score was calculated for each grid cell for current average conditions following a set of rules accounting for the depth to the water table and the percentage of area within each cell with slow infiltrating soils. A scenario with a higher sea-level position was also calculated for the Long Island part of the study area. The calculated groundwater emergence flood hazard score was reviewed in conjunction with the National Land Cover Database to identify developed areas and associated infrastructure that may be at risk of potential groundwater emergence flooding. Study results indicate that the groundwater emergence flood hazard is highest in coastal areas and near surface water where the water table is close to ground surface. Inland areas away from surface waterbodies are not likely to be exposed to groundwater emergence flooding. For Long Island, in the scenario with higher sea-level position, a greater groundwater emergence flood hazard is calculated for some locations closer to the coast and where land is submerged. Away from the coast and surface-water drainage, the groundwater emergence flood hazard is similar between the current average sea-level condition and the higher sea-level position scenario.

Connecticut, New York, Rhode Island↗

USGS2026

Effects of groundwater withdrawals for water bottling and municipal use, Wards Brook Valley, Maine and New Hampshire

Hydrologic models for the Wards Brook valley near Fryeburg, Maine were developed for historical (2016 – 2021) and hypothetical future conditions (2046 – 2065 and 2080 – 2099) to understand the effects of groundwater withdrawals for bottled water and municipal use on hydrologic conditions (stream base flows and groundwater levels). Analyses showed that the simulated base flows in Wards Brook were reduced because of pumping for both municipal water supplies and for water bottling, and about half of the total pumping impact on the base flows in Wards Brook was from the bottled water extraction. Simulated flows were greater than the minimum recommended streamflow of 2,180 cubic meters per day (400 gallons per minute) throughout the historical period. Simulated groundwater levels at two of three nearby ponds (Round Pond and Davis Pond) were minimally affected by pumping conditions, and effects were primarily from the municipal well closest to the ponds. Several estimates of future projected recharge were used to understand the potential effects of groundwater withdrawals on hydrologic conditions under multiple hypothetical climate conditions. Annual projected recharge rates in the mid- and late-21st century from two climate scenarios (stabilized greenhouse-gas emissions and high greenhouse-gas emissions) were similar to rates for 2016 – 2021. However, monthly recharge patterns for the future periods shifted toward more recharge in the winter months (December, January, and February) and less recharge in April, May, and October relative to 2016 – 2021. The lowest mean monthly base flows from the future emission scenarios all remain larger than the minimum recommended streamflow and indicate no long-term declines in flow relative to historical conditions. However, simulated base flows during hypothetical 3-year drought scenarios declined below minimum recommended streamflow during the summer months in the stabilized- and high-emission scenarios in the mid-21st century. Although water is generally plentiful in the Wards Brook valley, reduced pumping may be needed to maintain streamflows in Wards Brook under future climate conditions similar to modeled drought scenarios.

EarthArXiv↗

USGS2025

A spatial analysis of the groundwater emergence flood hazard in Long Island, New York and near coastal areas surrounding Long Island Sound in New York, Connecticut, and Rhode Island

Long Island, New York and near coastal areas surrounding Long Island Sound are densely populated and, like other coastal areas, are susceptible to flooding from several potential sources, including stormwater from precipitation events, tidal flooding and storm surge, and groundwater inundation or groundwater emergence flooding. The latter refers to the intersection of a rising water table with land surface or critical infrastructure. Many studies of flood drivers either neglect or only briefly discuss how shallow groundwater conditions may contribute to or exacerbate flood conditions. As part of a comprehensive study of compound flood hazards in the near coastal areas surrounding Long Island and Long Island Sound, a spatial analysis was completed, in cooperation with the Environmental Protection Agency’s Long Island Sound Study, using available regional datasets to characterize the potential hazard for groundwater emergence flooding. The approximately 3,100 square mile study area was subdivided into 11,407 900-meter by 900-meter (approximately 3,000-feet by 3,000-feet) grid cells, for the purposes of integrating the spatial datasets to calculate and map the groundwater emergence flood hazard. The depth to the water table, hydrologic soil groups, and National Land Cover Database were harmonized to the common grid. A groundwater emergence flood hazard rank was calculated for each grid cell for current average conditions following a set of rules accounting for the depth to the water table and the percent of area within each cell with slow infiltrating soils. A higher sea level position scenario was also calculated for the Long Island part of the study area. The calculated groundwater emergence flood hazard rank was reviewed in concert with the National Land Cover Data Base to identify developed areas and associated infrastructure that may be at risk to groundwater emergence flooding. Study results indicate that the groundwater emergence flood hazard is highest in coastal areas and near surface water where the water table is close to ground surface. Inland areas away from surface water bodies are not likely to be exposed to groundwater emergence flooding. For Long Island, under a scenario with higher sea level position, a greater groundwater emergence flood hazard is calculated in some locations closer to the coast and where land is submerged. Away from the coast and surface-water drainage, the groundwater emergence flood hazard is similar between the current average sea level condition and a higher sea level position scenario.

EarthArXiv↗

USGS2025

Multi-Scale Graph Learning for anti-sparse downscaling

Water temperature can vary substantially even across short distances within the same sub-watershed. Accurate prediction of stream water temperature at fine spatial resolutions (i.e., fine scales, ≤ 1 km) enables precise interventions to maintain water quality and protect aquatic habitats. Although spatiotemporal models have made substantial progress in spatially coarse time series modeling, challenges persist in predicting at fine spatial scales due to the lack of data at that scale. To address the problem of insufficient fine-scale data, we propose a Multi-Scale Graph Learning (MSGL) method. This method employs a multi-task learning framework where coarse-scale graph learning, bolstered by larger datasets, simultaneously enhances fine-scale graph learning. Although existing multi-scale or multi-resolution methods integrate data from different spatial scales, they often overlook the spatial correspondences across graph structures at various scales. To address this, our MSGL introduces an additional learning task, cross-scale interpolation learning, which leverages the hydrological connectedness of stream locations across coarse- and fine-scale graphs to establish cross-scale connections, thereby enhancing overall model performance. Furthermore, we have broken free from the mindset that multi-scale learning is limited to synchronous training by proposing an Asynchronous Multi-Scale Graph Learning method (ASYNC-MSGL). Extensive experiments demonstrate the state-of-the-art performance of our method for anti-sparse downscaling of daily stream temperatures in the Delaware River Basin, USA, highlighting its potential utility for water resources monitoring and management.

Conference Paper↗

USGS2025

Base-flow sampling to enhance understanding of the groundwater flow component of nitrogen loading in small watersheds draining into Long Island Sound

Excessive nitrogen discharge is a major concern for the Long Island Sound. Programs have been implemented to reduce point sources of nitrogen to the sound, but little is known about the nonpoint sources. This study aims to better understand the current groundwater contributions of nitrogen from nonpoint sources in the Long Island Sound watershed. During the spring and summer of 2022, the U.S. Geological Survey, in cooperation with the U.S. Environmental Protection Agency, collected water-quality samples to analyze nutrients (nitrogen and phosphorus), chloride, and bromide at 45 stations in the Long Island Sound watershed in Connecticut, New York, and Rhode Island. The stations were in small drainage watersheds (5 to 30 square kilometers) in the southern part of the Long Island Sound watershed. During two separate synoptic sampling events, water-quality samples and instantaneous streamflow measurements were collected under base-flow conditions (where the streamflow is dominated by groundwater inputs rather than overland flow or runoff flow). One sampling event was in the nongrowing season (April 24–25, 2022), and the other was in the growing season (June 30–July 1, 2022). To calculate instantaneous nitrogen loads and yields, streamflow was measured at the time of sample collection. Nitrogen concentrations, loads, and yields varied among sampling stations and by season. Total filtered nitrogen concentrations were generally lower in the nongrowing season (from less than 0.14 to 1.9 milligrams per liter) than in the growing season (from less than 0.23 to 3.0 milligrams per liter). Nitrate plus nitrite concentrations showed little variation between the nongrowing and growing seasons. Unfiltered ammonia plus organic nitrogen concentrations were generally lower in the nongrowing season (from less than 0.07 to 0.83 milligram per liter) than in the growing season (from 0.11 to 0.98 milligram per liter). In contrast, total filtered and unfiltered nitrogen loads and yields were higher in the nongrowing season than during the growing season, likely because streamflows were higher during the nongrowing season. Total unfiltered nitrogen yields during the nongrowing season ranged from less than 0.15 to 5.0 kilograms per square kilometer per day. Total unfiltered nitrogen yields during the growing season ranged from less than 0.12 to 2.5 kilograms per square kilometer per day. Total filtered nitrogen yields during the nongrowing season ranged from less than 0.13 to 5.2 kilograms per square kilometer per day. Total filtered nitrogen yields during the growing season ranged from less than 0.06 to 2.5 kilograms per square kilometer per day.

Connecticut, New York, Rhode Island↗

USGS2024

Simulated mean monthly groundwater-transported nitrogen loads in watersheds on the north shore of Long Island Sound, 1993–2022

Elevated nitrogen loads are pervasive in the Long Island Sound, an estuary that receives freshwater and nutrients from both surface-water and groundwater discharge. Surface-water nitrogen loads to the Long Island Sound are relatively well characterized, but less is known about groundwater-transported nitrogen loads. Prior work on the northern shore of Long Island Sound (Connecticut and areas of New York and Rhode Island) suggested that groundwater travel times are relatively short (median less than 2 years) and that decade-long nutrient legacies are not widespread. Because the travel times are short, groundwater flow and nutrient loads likely vary substantially between months. In the current study, the U.S. Geological Survey, in cooperation with the U.S. Environmental Protection Agency’s Long Island Sound Study and the Connecticut Department of Energy and Environmental Protection, developed a set of models to better characterize spatial and temporal patterns of groundwater-transported nitrogen loading from atmospheric deposition, septic systems, and fertilizers within the study area. The models provide an estimate, with uncertainty, of groundwater-transported nitrogen loads in the study area, filling a key gap in the nitrogen budget for Long Island Sound. The models also highlight the spatial and temporal variation in nitrogen loading throughout the study area. The modeling workflow involved four models. (1) A soil-water-balance model was developed by using the Soil-Water-Balance software to simulate groundwater recharge across the study area for water years 2005 through 2022. The simulated mean monthly recharge from the soil-water-balance model was used as input into a groundwater-flow model. (2) The groundwater-flow model was developed by using the MODFLOW 6 software and data for water years 1993 through 2022 and simulates average monthly hydrologic conditions. The groundwater-flow model was calibrated by using the Iterative Ensemble Smoother method within the PEST++ software. The Iterative Ensemble Smoother method generates an ensemble of sets of parameter values, with each set producing reasonable simulated hydrologic parameter values. (3) An ensemble of MODPATH particle-tracking simulations were run to generate particle flow paths and travel times, with each simulation using a different set of the flow model parameters. (4) A nitrogen load model uses the MODPATH simulation outputs to track nitrogen from the land surface through multiple attenuation zones until it discharges into fresh or saline surface water. As with the groundwater-flow model, the nitrogen model simulated average monthly groundwater-transported nitrogen loads for water years 1993 through 2022. One novel aspect of the nitrogen load model is that the nitrogen attenuation parameters were calibrated to observed nitrogen loads. Across the ensemble of simulated nitrogen loads, the median study-area-wide monthly simulated nitrogen loads from the aquifer to Long Island Sound throughout the year ranged from 900 to 18,600 kilograms of nitrogen per day, with a median load of 5,100 kilograms of nitrogen per day. The simulated loads were based on average monthly conditions for water years 1993 through 2022. Loads were highest during the winter and early spring and lowest during the late summer. However, simulated travel times for groundwater and nitrogen loads discharged to Long Island Sound during summer were longer than travel times for groundwater and loads discharged during the winter, indicating that, on average, groundwater discharged during summer traveled along different, and longer, flow paths, than groundwater discharged during winter. This indicates that summer loads would respond more slowly to changes in nitrogen inputs at the water table than winter loads. Over the entire study area, approximately 15 percent of the simulated load is from atmospheric deposition sources, 30 to 40 percent is from fertilizer, and 50 to 60 percent is from septic systems. The final analysis of the study involved simulating the change in groundwater-transported nitrogen load in response to upgrading septic systems or reducing fertilizing inputs to areas of turf grass. Both management interventions reduced the groundwater-transported nitrogen load, and reductions were greater in areas with greater loads from septic systems or turf-grass fertilizers. The delay between management actions and substantial reductions in groundwater-transported nitrogen loads varied seasonally; loads during the late summer months remained elevated longer than the winter loads.

Connecticut, Rhode Island↗

USGS2024

Exploring landscape and geologic controls on spatial patterning of streambank groundwater discharge in a mixed land use watershed

Preferential groundwater discharge features along stream corridors are ecologically important at local and stream network scales, yet we lack quantification of the multiscale controls on the spatial patterning of groundwater discharge. Here we identify physical attributes that best explain variation in the presence and lateral extent of preferential groundwater discharges along two 5th order streams, the Housatonic and Farmington Rivers, and 32 1st to 4th order reaches across the Farmington River network. We mapped locations of preferential groundwater discharge exposed along streambanks using handheld thermal infrared cameras paired with high-resolution topographic and land use land cover datasets, surficial soil characteristic maps, and depth-to-bedrock geophysical measurements. The unconfined Housatonic River, MA, USA (12 km) had fewer discharge locations and less lateral extent (41 discharge locations with 38 m of active discharge/km of river) compared to the partially confined Farmington River, CT, USA (26 km; 169 discharge locations with 129 m of active discharge/km of river). Using a moving window analysis, we found along both rivers that discharge was more likely to occur where bank slopes were steeper, floodplain extent was narrower, and degree of confinement was higher. Along the Farmington River, groundwater discharge was more likely to occur where saturated hydraulic conductivity was higher and depth-to-bedrock was shallower. Among the 32 stream reaches surveyed (33.2 km of total stream length) within the Farmington River watershed, preferential discharge was observed in all but two stream reaches, varied from 0 to 25% of lateral extent along stream banks (mean = 6%), and was more likely to occur where stream reach slopes were steep, saturated hydraulic conductivity was high, and watershed urbanization was low. Our results show that, though both surface (e.g., topographic, land use land cover) and subsurface (e.g., soil characteristics, bedrock depth) factors control the prevalence of streambank preferential groundwater discharge, the dominant controls vary across valley settings and stream sizes.

Connecticut, Massachusetts↗

USGS2023

Train, inform, borrow, or combine? Approaches to process-guided deep learning for groundwater-influenced stream temperature prediction

Although groundwater discharge is a critical stream temperature control process, it is not explicitly represented in many stream temperature models, an omission that may reduce predictive accuracy, hinder management of aquatic habitat, and decrease user confidence. We assessed the performance of a previously-described process-guided deep learning model of stream temperature in the Delaware River Basin (USA). We found lower accuracy (root mean square error [RMSE] of 1.71 versus 1.35°C) and stronger seasonal bias (absolute mean monthly bias of 1.06 vs. 0.68°C) for reaches primarily influenced by deep groundwater as compared to atmospheric conditions. We then tested four approaches for improving groundwater process representation: (a) a custom loss function leveraging the unique patterns of air and water temperature coupling characteristic of different temperature drivers, (b) inclusion of additional groundwater-relevant catchment attributes, (c) incorporation of additional process model outputs, and (d) a composite model. The custom loss function and the additional attributes significantly improved the predictive accuracy in groundwater-dominated reaches (RMSE of 1.37 and 1.26°C) and reduced the seasonal bias (absolute mean monthly bias of 0.44 and 0.48°C), but neither approach could identify holdout groundwater reaches. Variable importance analysis indicates the custom loss function nudges the model to use the existing inputs more efficiently, whereas with the added features the model relies on a broader suite of inputs. This analysis is a substantial step toward more accurately representing groundwater discharge processes in stream temperature models and will improve predictive accuracy and inform habitat management.

Delaware River Basin↗

USGS2023

Where the past meets the present: Connecting nitrogen from watersheds to streams through groundwater flowpaths

Groundwater discharge to streams is a nonpoint source of nitrogen (N) that confounds N mitigation efforts and represents a significant portion of the annual N loading to watersheds. However, we lack an understanding of where and how much groundwater N enters streams and watersheds. Nitrogen concentrations at the end of groundwater flowpaths are the culmination of biogeochemical and physical processes from the contributing land area where groundwater recharges, within the aquifer system, and in the near-stream riparian area where groundwater discharges to streams. Our research objectives were to quantify the spatial distribution of N concentrations at groundwater discharges throughout a mixed land-use watershed and to evaluate how relationships among contributing and riparian land cover, modeled aquifer characteristics, and groundwater discharge biogeochemistry explain the spatial variation in groundwater discharge N concentrations. We accomplished this by integrating high-resolution thermal infrared surveys to locate groundwater discharge, biogeochemical sampling of groundwater, and a particle tracking model that links groundwater discharge locations to their contributing area land cover. Groundwater N loading from groundwater discharges within the watershed varied substantially between and within streambank groundwater discharge features. Groundwater nitrate concentrations were spatially heterogeneous ranging from below 0.03–11.45 mg-N/L, varying up to 20-fold within meters. When combined with the particle tracking model results and land cover metrics, we found that groundwater discharge nitrate concentrations were best predicted by a linear mixed-effect model that explained over 60% of the variation in nitrate concentrations, including aquifer chemistry (dissolved oxygen, Cl − , SO 4 2− ), riparian area forested land cover, and modeled physical aquifer characteristics (discharge, Euclidean distance). Our work highlights the significant spatial variability in groundwater discharge nitrate concentrations within mixed land-use watersheds and the need to understand groundwater N processing across the many spatiotemporal scales within groundwater cycling.

Connecticut↗

USGS2023

Stream temperature prediction in a shifting environment: The influence of deep learning architecture

Stream temperature is a fundamental control on ecosystem health. Recent efforts incorporating process guidance into deep learning models for predicting stream temperature have been shown to outperform existing statistical and physical models. This performance is in part because deep learning architectures can actively learn spatiotemporal relationships that govern how water and energy propagate through a river network. However, exploration of how spatiotemporal awareness and process guidance influence a model's generalizability under shifting environmental conditions such as climate change is limited. Here, we use Explainable Artificial Intelligence (XAI) to interrogate how differing deep learning architectures affect a model's learned spatial and temporal dependencies, and how those learned dependencies affect a model's ability to maintain high accuracy when applied to unseen environmental conditions. Using the Delaware River Basin in the northeastern United States as a test case, we compare two spatiotemporally aware process-guided deep learning models for predicting stream temperature (a recurrent graph convolution network—RGCN, and a temporal convolution graph model—Graph WaveNet). Both models achieve equally high predictive performance when testing data are well represented in the training data (test root mean squared errors of 1.64°C and 1.65°C); however, Graph WaveNet significantly outperforms RGCN in 4 out of 5 experiments where test partitions represent different types of unseen environmental conditions. XAI results show that the architecture of Graph WaveNet leads to learned spatial relationships with greater fidelity to physical processes, and that this fidelity improves the generalizability of the model when applied to shifting and/or unseen environmental conditions.

Delaware River Basin↗

USGS2022

Where groundwater seeps: Evaluating modeled groundwater discharge patterns with thermal infrared surveys at the river-network scale

Predicting baseflow dynamics, protecting aquatic habitat, and managing legacy contaminants requires explicit characterization and prediction of groundwater discharge patterns throughout river networks. Using handheld thermal infrared (TIR) cameras, we surveyed 47 km of stream length across the Farmington River watershed (1,570 km 2 ; CT and MA, USA), mapping locations of bank and waterline groundwater discharges based on their thermal signature. Using the observed groundwater discharge locations and predicted groundwater discharge rates from 6 variations of a numerical groundwater-flow model (MODFLOW-NWT), we compared 1) predicted groundwater-discharge rates in areas with and without observed groundwater discharge, 2) spatial patterns of observed and predicted groundwater discharge locations, and 3) density of observed groundwater discharge locations with predicted discharge rates. Five of six models reasonably predicted the spatial patterns of discharge locations along the 5th order mainstem, but fewer models predicted groundwater discharge patterns in smaller streams. Our results highlight 1) the feasibility of using TIR observations to evaluate groundwater models, 2) model parameters that influence discharge prediction accuracy (riverbed sediment and bedrock hydraulic conductivity and river-aquifer connections), and 3) current strengths and future opportunities for improved modeling of groundwater-discharge patterns.

Connecticut, Massachusetts↗

USGS2021

Simulation of groundwater budgets and travel times for watersheds on the north shore of Long Island Sound, with implications for nitrogen-transport studies

Aquatic systems in and around the Long Island Sound (LIS) provide a variety of ecological and economic benefits, but in some areas of the LIS, aquatic ecosystems have become degraded by excess nitrogen. A substantial fraction of the nitrogen inputs to the LIS are transported through the groundwater-flow system. Because groundwater travel times in surficial aquifers can exceed 100 years, multiyear lags are introduced between inputs at the water table in recharge areas and discharge to inland or coastal receiving waters. The U.S. Geological Survey, in cooperation with the Connecticut Department of Energy and Environmental Protection and the U.S. Environmental Protection Agency’s Long Island Sound Study, developed a steady-state groundwater model of the watersheds draining from the northern shore of the LIS for the purpose of calculating groundwater budgets and travel times to coastal waters. The model was developed by using the MODFLOW–NWT software and existing spatial data on aquifers, river networks, land-surface altitudes, land cover, groundwater recharge, and water use. Coastal waters were delineated on the basis of the National Wetland Inventory; all non-coastal waters were collectively termed “inland waters.” A coarse-resolution model was calibrated by using the PEST++ software, long-term records of water levels in 65 wells, stream altitudes from 477 streams, base-flow records for 14 streamgages that are relatively unaffected by withdrawals, and error metrics based on incorrectly simulated flooding and incorrectly simulated dry streams. The calibrated values were used in a fine-resolution model in which the mean absolute residuals were 4.5 meters for groundwater levels, 1.3 meters for stream altitudes, and 7,200 cubic meters per day (2.9 cubic feet per second) for base flow. About 89 percent of the terrestrial cells were correctly simulated with the water table below land surface, and nearly 90 percent of the cells representing streams were correctly simulated as having the water table above the stream bottom. Together, these metrics suggest that this model is robust for simulating regional-scale groundwater patterns. Simulated groundwater budgets were compiled for the entire study area, for each HUC12 (Hydrologic Unit Code no. 12) watershed and its adjacent coastal waters, if applicable, within the study area, and for 14 coastal-embayment watersheds. Most groundwater (90.6 percent of inflows) discharged to inland waters, with smaller fractions to coastal waters (7.0 percent) and well withdrawals (2.4 percent). When computed for HUC12 watersheds with coastal discharge, the portions of groundwater discharging to coastal waters ranged from 0.02 to 66 percent of groundwater outflows, with a median of 13 percent. Within priority-embayment watersheds, the portions of groundwater discharging to coastal waters ranged from 2 to 56 percent, with a median of 15 percent. Groundwater travel times also were simulated for the entire study area, for each HUC12 watershed and its adjacent coastal waters, if applicable, within the study area and for 14 priority coastal embayments. Within the entire study area, the median groundwater travel time was 1.9 years, with an interquartile range of 0.1 to 5.9 years. Sensitivity analysis of groundwater travel times within a subbasin in the study area indicates that the travel times are a function of the grid resolution, with coarser grids resulting in shorter median travel times. Travel times for groundwater discharging to coastal waters were similar to travel times for groundwater discharging to inland waters, with a median of 1.9 years. Median travel times for the HUC12 watersheds ranged from 0.9 to 53.5 years, with a median of 1.8 years. Among HUC12 watersheds that include coastal areas, travel times for groundwater discharging to coastal waters ranged from less than 1 to 61.6 years, with a median of 2.8 years. The HUC12 watersheds with the longest simulated travel times were in the urban area near New York City where the model performance is less accurate. Median travel times for groundwater discharging to coastal waters within the priority-embayment watersheds ranged from less than 1 to 18.6 years, with a median of 2.3 years. A more focused analysis was conducted for the Niantic River watershed to demonstrate the applicability of the regional model to local-scale nitrogen-transport analyses by using nitrogen-input and -attenuation rates from literature sources. Nitrogen inputs were estimated by using land-cover-based loading factors, and attenuation was estimated by using attenuation factors based on geologic zones and soil properties. Based on this analysis, groundwater transports an estimated 22,000 kilograms of nitrogen per year (2.9 kilograms of nitrogen per hectare per year) to streams, rivers, and coastal waters within the Niantic River watershed. Approximately 36 percent of discharging nitrogen is from atmospheric-deposition sources, 38 percent is from fertilizers, and 26 percent is from septic systems. Most of the groundwater-transported nitrogen (88 percent) discharges first to streams and rivers, with only 12 percent discharging directly to coastal waters. Travel times for groundwater-transported nitrogen ranged from less than 1 day to more than 100 years, with a median of 1.6 years.

Connecticut, New York, Rhode Island↗

USGS2020

2020 drought in New England

Below average and infrequent rainfall from May through September 2020 led to an extreme hydrologic drought across much of New England, with some areas experiencing a flash drought, reflecting its quick onset. The U.S. Geological Survey (USGS) recorded record-low streamflow and groundwater levels throughout the region. In September, the U.S. Department of Agriculture (2020) declared Aroostook County in Maine and Hillsborough and Merrimack Counties in New Hampshire as crop disaster areas. By the beginning of October, 166 community water systems and 5 municipalities in New Hampshire, more than 100 municipalities in Massachusetts, and several community water supplies in Connecticut, Maine, and Rhode Island had mandatory water restrictions in place.

Connecticut, Maine, Massachusetts, New Hampshire, ↗

USGS2020

Improved prediction of management-relevant groundwater discharge characteristics throughout river networks

Groundwater discharge zones connect aquifers to surface water, generating baseflow and serving as ecosystem control points across aquatic ecosystems. The influence of groundwater discharge on surface flow connectivity, fate and transport of contaminants and nutrients, and thermal habitat depends strongly on hydrologic characteristics such as the spatial distribution, age, and depth of source groundwater flow paths. Groundwater models have the potential to predict spatial discharge characteristics within river networks, but models are often not evaluated against these critical characteristics and model equifinality with respect to discharge processes is a known challenge. We quantify discharge characteristics across a suite of groundwater models with commonly used frameworks and calibration data. We developed a base model (MODFLOW‐NWT) for a 1,570‐km 2 watershed in the northeastern United States and varied the calibration data, control of river‐aquifer exchange directionality, and resolution. Most models ( n = 11 of 12) fit similarly to calibration metrics, but patterns in discharge location, flow path depth, and subsurface travel time varied substantially. We found (1) a 15% difference in the percent of discharge going to first‐order streams, (2) threefold variations in flow path depth, and (3) sevenfold variations in the subsurface travel times among the models. We recalibrated three models using a synthetic discharge location data set. Calibration with discharge location data reduced differences in simulated discharge characteristics, suggesting an approach to improved equifinality based on widespread field‐based mapping of discharge zones. Our work quantifying variation across common modeling approaches is an important step toward characterizing and improving predictions of groundwater discharge characteristics.

Connecticut, Massachusetts↗

USGS2020

Updating data inputs, assessing trends, and evaluating a method to estimate probable high groundwater levels in selected areas of Massachusetts

A method to estimate the probable high groundwater level in Massachusetts, excluding Cape Cod and the islands, was developed in 1981. The method uses a groundwater measurement from a test site, groundwater measurements from an index well, and a distribution of high groundwater levels from wells in similar geologic and topographic settings. The U.S. Geological Survey, in cooperation with the Massachusetts Department of Environmental Protection, conducted an update to the Frimpter method for estimating the probable high groundwater levels in Massachusetts. The study evaluated the potential changes to the method resulting from four decades of additional groundwater-level data and the expansion of the network of wells for monitoring groundwater levels. The differences and potential benefits of daily, as opposed to monthly, measurements in the application of the method were examined because of the increased availability of high-frequency (subdaily) groundwater-level data. The study also considered long-term trends in groundwater levels that may alter the accuracy of the method. Finally, the accuracy of the estimated high groundwater levels was evaluated, and improved implementation guidance was prepared. For this study, groundwater levels in 153 wells in Massachusetts and surrounding States with records with lengths of 16 to 78 years were analyzed. The highest recorded groundwater levels ranged from 1.2 feet (ft) above land surface (flooded conditions) to 45.8 ft below land surface, with a median of 4.6 ft below land surface. The maximum annual groundwater-level range was 1.4 to 17.9 ft, with a median of 5.5 ft. The within-month variation, maximum annual groundwater-level range, and highest recorded groundwater level were computed using daily mean groundwater-level values from 28 wells with continuous records. The use of daily data resulted in larger maximum annual groundwater-level ranges (0.02 to 2.94 ft larger, with a median of 0.58 ft larger) and shallower highest-recorded groundwater levels (0.0 to 1.60 ft shallower, with a median of 0.18 ft shallower) than computations based on monthly measurements in the same wells. Statistical tests showed moderate to strong evidence of trends in measurements of both high and low groundwater levels within most of the periods during which water levels were analyzed. High groundwater levels rose beneath the land surface at most sites during four of the six periods used for analysis (1966–2015, 1986–2015, 1991–2010, and 1981–2010). Low groundwater levels also increased at many sites during most of the periods evaluated, but this trend was less widespread than the similar trends in high groundwater levels, and the trend was to deeper low groundwater levels at more sites than the trend to deeper high groundwater levels. There was no clear trend in annual groundwater-level ranges at most sites during the six periods analyzed. In general, the Frimpter method predicted shallower (higher) high groundwater levels than were observed but correctly classified sites according to their suitabilities for unmounded septic systems. The mean error of the predictions (difference between the estimated and observed groundwater levels) ranged from −3.23 ft to −1.40 ft for various approaches to estimating the groundwater-level range and selecting an index well. The method correctly classified 83 to 86 percent of monitoring-well sites according to their suitability for an unmounded septic system for many approaches to estimating the annual groundwater-level range and selecting an index well. The approach selected for estimating the annual groundwater-level range and selecting an index well will depend upon the importance of an accurate estimate of the high groundwater level as compared to the importance of an estimated high groundwater level that is less likely to be exceeded.

Connecticut, Massachusetts, New Hampshire, Rhode I↗

USGS2020

Hydrogeology and interactions of groundwater and surface water near Mill Creek and the Herring River, Wellfleet, Massachusetts, 2017–18

Groundwater levels and stream stage were monitored by the U.S. Geological Survey, in cooperation with the Friends of Herring River, at 19 sites in the Mill Creek Basin, a tributary of the Herring River in Wellfleet, Massachusetts, on outer Cape Cod, to provide baseline data prior to a proposed restoration of tidal flow to the Herring River estuary at the Cape Cod National Seashore. Tidal flow in the Herring River has been restricted by a tide-control structure since 1909. Baseline data are necessary to understand current conditions and provide information on water levels for comparison to future water levels under the proposed Herring River restoration, which includes restoration of salt marshes by enhancing tidal flow to the Herring River and construction of a tide-control structure on Mill Creek to prevent the flooding of upstream private properties, including a golf course. Analysis of data collected during monitoring-well installation at eight locations on or near the golf course and Mill Creek, along with analysis of existing information, determined that parts of the study area are underlain by salt marsh deposits up to 18 feet (ft) thick. These marsh deposits are directly underlain by estuarine sediments, and adjacent upland areas are underlain by medium to very coarse sand. The freshwater lens on the golf course is 70 ft thick or more. Groundwater levels at individual wells in the study area fluctuated by 1.3 to 2.6 ft during the study period (June 1, 2017, to June 14, 2018). Total precipitation during this period was 60.8 inches, about 10 inches greater than the long-term (2000–17) annual average (50.3 inches). Groundwater levels on Cape Cod generally were normal to above normal during the study owing to the higher than normal precipitation. Tidal amplitudes of groundwater levels caused by daily fluctuations at nearby tidal waterbodies (M2 tidal harmonic) were as large as 0.12 ft at a well 105 ft from the tidally restricted Herring River and as large as 0.06 ft at a well 575 ft from Wellfleet Harbor. Tidal fluctuations in groundwater levels were generally limited to areas about 1,500 ft from the nearest tidal waterbody. Under the initial proposed restoration, where mean tides would be maintained similar to current conditions, tidal fluctuations would be restored to parts of Mill Creek, and subsequent tidal fluctuations in groundwater levels could increase at some of the areas closest to the proposed tide-control structure, but the fluctuations would be less than about 0.06 ft in magnitude. Regression models were used to describe the variability of daily mean tidally filtered groundwater levels and daily maximum stream stage in Mill Creek. Significant independent variables for the groundwater-level model included daily tidally filtered Wellfleet Harbor stage with a lag time of zero to 2 days, 7-day precipitation, the growing degree days (50 degrees Fahrenheit), and the quartile of groundwater levels relative to a long period of record at a nearby observation well. Significant independent variables to predict the Mill Creek stage included daily mean groundwater levels in nearby wells, 7-day precipitation, growing degree days (50 degrees Fahrenheit), and a binary indicator of either a flooded or nonflooded condition on the golf course near Mill Creek. Flooding in Mill Creek occurred primarily when groundwater levels at nearby wells reached certain thresholds, when the precipitation in the preceding 7 days was at least 0.92–1.04 inches, and during the nongrowing season.

Massachusetts↗