4.4 Article

Spatial aggregation effects on the performance of machine learning metamodels for predicting transit time to baseflow

Related references

Note: Only part of the references are listed.
Article Engineering, Civil

Modeling groundwater transit time distributions and means across a Nebraska watershed: Effects of heterogeneity in the aquifer, riverbed, and recharge parameters

Caner Zeyrek et al.

Summary: The study investigates the travel times of groundwater discharge to the Upper Middle Loup River in Nebraska, USA, using a 3D-steady-state model. The results show significant differences in mean transit times and transit time distributions along the river due to variations in recharge, aquifer heterogeneity, and riverbed hydraulic conductivity.

JOURNAL OF HYDROLOGY (2023)

Article Engineering, Civil

Flood susceptible prediction through the use of geospatial variables and machine learning methods

Navid Mahdizadeh Gharakhanlou et al.

Summary: This study assessed the spatial patterns of flood susceptibility using six machine learning models and selected thirteen variables related to flood susceptibility. The results showed that the Random Forest model had the best predictive potential, and flood susceptibility was mainly influenced by lithology, drainage density, distance from rivers, soil, rainfall, land cover, and Normalized Difference Moisture Index (NDMI). According to this model, flood susceptibility in the study area was classified into five levels, providing valuable resources for flood management.

JOURNAL OF HYDROLOGY (2023)

Article Environmental Sciences

Constraining Bedrock Groundwater Residence Times in a Mountain System With Environmental Tracer Observations and Bayesian Uncertainty Quantification

Nicholas E. E. Thiros et al.

Summary: This study investigates the groundwater residence times in a mountain catchment in the East River Watershed, Colorado, USA. The results show that the groundwater in the upslope well has a residence time of hundreds of years, while the groundwater in the toe-slope well has a residence time of thousands of years. This highlights the importance of recharge dynamics and flow path mixing in the groundwater system.

WATER RESOURCES RESEARCH (2023)

Article Environmental Sciences

A Machine Learning Approach to Predict Watershed Health Indices for Sediments and Nutrients at Ungauged Basins

Ganeshchandra Mallya et al.

Summary: Effective water quality management relies on accurate water quality data, which is often sparse in both time and space. In this study, machine learning models were used to predict watershed health and other risk metrics at ungauged locations, providing a valuable tool for decision makers and water quality monitoring agencies.

WATER (2023)

Review Engineering, Environmental

Application of machine learning in groundwater quality modeling-A comprehensive review

Ryan Haggerty et al.

Summary: This review assesses the development of machine learning techniques in predicting groundwater quality over the past two decades. It evaluates the application of various supervised, semi-supervised, unsupervised, and ensemble ML models, with neural networks being the most widely used. The usage of neural networks has decreased in recent years, leading to the emergence of more accurate or advanced techniques like deep learning or unsupervised algorithms. Iran and the United States lead in the number of modeled areas, with extensive historical data available. Future advancements include further implementation of deep learning, explainable artificial intelligence, and other cutting-edge techniques, as well as modeling of sparsely studied variables and unique study areas, and the application of ML techniques for groundwater quality management.

WATER RESEARCH (2023)

Article Meteorology & Atmospheric Sciences

Accelerating the Lagrangian Particle Tracking in Hydrologic Modeling to Continental-Scale

Chen Yang et al.

Summary: Unprecedented climate change and anthropogenic activities have led to increasing ecohydrological problems, prompting the development of large-scale hydrologic modeling for solutions. However, scientific progress in tracking water parcels at large-scale with high spatiotemporal resolutions is lacking due to the absence of powerful modeling tools. In this study, a parallel framework for the particle tracking model EcoSLIM is demonstrated, showing significant speedup and excellent parallel performance. This parallel framework can be applied to other particle tracking models and is a promising tool for accelerating our understanding of the terrestrial water cycle and upscaling subsurface hydrology to Earth System Models.

JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS (2023)

Article Environmental Sciences

Machine learning predictions of nitrate in groundwater used for drinking supply in the conterminous United States

K. M. Ransom et al.

Summary: The study developed a three-dimensional machine learning model to predict the distribution of nitrate in groundwater, providing important insights into monitoring nitrate in groundwater in the conterminous United States. Predictions showed that nitrate concentrations were mostly below 1 mg/L in most areas, but about 1.4 million people still rely on groundwater for their drinking supplies.

SCIENCE OF THE TOTAL ENVIRONMENT (2022)

Article Multidisciplinary Sciences

A hydrological simulation dataset of the Upper Colorado River Basin from 1983 to 2019

Hoang Tran et al.

Summary: This article presents a hydrological reconstruction of the Upper Colorado River Basin using hourly temporal resolution and 1-km spatial resolution from October 1982 to September 2019. The validated dataset includes various hydrologic variables such as streamflow, water table depth, snow water equivalent (SWE), and evapotranspiration (ET). These datasets provide a long-term simulation of natural flow for one of the most over-allocated basins in the world.

SCIENTIFIC DATA (2022)

Article Water Resources

Coevolution of machine learning and process-based modelling to revolutionize Earth and environmental sciences: A perspective

Saman Razavi et al.

Summary: Machine learning applications in Earth and environmental sciences have evolved separately from traditional process-based modeling paradigms. Overcoming cultural barriers and exploring the strengths and weaknesses of both approaches are essential for developing a coevolutionary approach to model building.

HYDROLOGICAL PROCESSES (2022)

Article Engineering, Environmental

Regional Scale Assessment of Shallow Groundwater Vulnerability to Contamination from Unconventional Hydrocarbon Extraction

Mario A. Soriano et al.

Summary: This study evaluated the vulnerability of groundwater to contamination from unconventional oil and gas (UOG) extraction in a large region of the Appalachian Basin in northeastern USA. By simulating groundwater flow and contaminant transport processes, the researchers estimated that approximately 21,000-30,000 individuals rely on domestic water wells that are at risk of contamination from UOG activities. The findings of this study can be used to guide groundwater monitoring, inform public health studies, and address environmental justice issues.

ENVIRONMENTAL SCIENCE & TECHNOLOGY (2022)

Review Environmental Sciences

Transit Time Estimation in Catchments: Recent Developments and Future Directions

Paolo Benettin et al.

Summary: This study reviews the recent developments in water transit time research, focusing on water-age-balance approaches and data-based approaches. Various methods, including StorAge-Selection functions, well/partially mixed compartments, and direct transit time estimates, are discussed and compared. The study aims to unify the heterogeneity in the literature and highlight the remaining open questions in transit time research.

WATER RESOURCES RESEARCH (2022)

Review Environmental Sciences

A Review of Current Capabilities and Science Gaps in Water Supply Data, Modeling, and Trends for Water Availability Assessments in the Upper Colorado River Basin

Fred D. Tillman et al.

Summary: The Colorado River is a critical water resource in the southwestern United States, providing drinking water for 40 million people and irrigation water for 2.2 million hectares of land. Challenges in water availability due to extended drought and potential climate change in the Upper Colorado River Basin are highlighted. The manuscript reviews the current state of science for water availability components in the UCOL, emphasizing the importance of data, modeling, and trends.

WATER (2022)

Article Engineering, Civil

Assessing spatial transferability of a random forest metamodel for predicting drainage fraction

Elisa Bjerre et al.

Summary: Fully distributed hydrological models are widely used in groundwater management, but their speed and data requirements hinder their use for decision support. Metamodels provide a simpler and faster model that emulates the complex model using machine learning techniques. However, metamodel predictions beyond the ranges of training data are highly uncertain, so assessing model transferability is crucial. We present a novel methodology for evaluating model transferability based on differences in covariate distributions between training and testing data. This approach can be used as a screening tool to assess the suitability of a metamodel for spatial prediction beyond its training domain.

JOURNAL OF HYDROLOGY (2022)

Article Environmental Sciences

Random forest-based modeling of stream nutrients at national level in a data-scarce region

Holger Virro et al.

Summary: Water quality modeling is essential for effective mitigation efforts, and recent advancements in machine learning methods have shown promising results in predicting nutrient concentrations.

SCIENCE OF THE TOTAL ENVIRONMENT (2022)

Article Environmental Sciences

Over a third of groundwater in USA public-supply aquifers is Anthropocene-age and susceptible to surface contamination

Bryant C. Jurgens et al.

Summary: The distribution of groundwater age is important for assessing the vulnerability and sustainability of groundwater resources. This study computes the aquifer-scale cumulative distribution function to characterize the age distribution for 21 Principal Aquifers in the United States. The results show that the groundwater is predominantly Anthropocene, Holocene, and Pleistocene, with varying degrees of susceptibility to land-surface contamination. Pumping sustainability depends on recharge and release rates of groundwater stored in fine-grained layers.

COMMUNICATIONS EARTH & ENVIRONMENT (2022)

Article Water Resources

Assessing the contribution of groundwater to catchment travel time distributions through integrating conceptual flux tracking with explicit Lagrangian particle tracking

Miao Jing et al.

Summary: The study investigates the characteristics and influencing factors of travel time distributions in a catchment area, finding that topography and aquifer structure have significant impacts on groundwater travel times, while meteorological drivers also affect transport processes in soil zones and groundwater. The catchment travel time exhibits power-law shape and fractal behavior, with prediction uncertainty primarily influenced by deep groundwater. Multiple summary statistics are recommended to minimize predictive uncertainty introduced by tailing behavior in catchment travel time distributions.

ADVANCES IN WATER RESOURCES (2021)

Article Geosciences, Multidisciplinary

How Will Baseflow Respond to Climate Change in the Upper Colorado River Basin?

Olivia L. Miller et al.

Summary: This study provides estimates of projected baseflow changes in the Upper Colorado River Basin under different climatic conditions, showing potential declines in total baseflow supplied to the Lower Colorado River Basin, with a possible increase under warm/wet conditions. Additionally, the results highlight the impacts of climate-driven changes in high-elevation hydrology on basinwide water availability.

GEOPHYSICAL RESEARCH LETTERS (2021)

Article Environmental Sciences

Three-Dimensional Distribution of Groundwater Residence Time Metrics in the Glaciated United States Using Metamodels Trained on General Numerical Simulation Models

J. J. Starn et al.

Summary: Residence time distribution (RTD) is a critical characteristic of groundwater flow systems, inferred from tracer data and numerical models. A novel automated approach utilizing General Simulation Models (GSMs) and Boosted Regression Tree metamodels successfully predicted RTD metrics, matching with numerical model results. The study identified aquifer thickness, recharge rate, Multi-Order Hydrologic Position (MOHP), and hydrogeologic terrane as important predictors for RTD metrics.

WATER RESOURCES RESEARCH (2021)

Article Environmental Sciences

Toward catchment hydro-biogeochemical theories

Li Li et al.

Summary: Headwater catchments are fundamental units connecting land and ocean, with hydrological flow and biogeochemical processes intricately linked yet lacking integration in their scientific development. The absence of integrated theories hinders understanding and forecasting the future of the Critical Zone under human and climate-induced perturbations.

WILEY INTERDISCIPLINARY REVIEWS-WATER (2021)

Article Engineering, Civil

Machine learning predictions of mean ages of shallow well samples in the Great Lakes Basin, USA

C. T. Green et al.

Summary: The travel time or age of groundwater affects catchment responses to hydrologic changes, geochemical reactions, and time lags between management actions and responses at down-gradient streams and wells. Machine learning was used to predict ages in wells across a large region around the Great Lakes Basin, showing a broad distribution of ages among wells with a median of 32.9 years for the Basin.

JOURNAL OF HYDROLOGY (2021)

Article Water Resources

River network travel time is correlated with dissolved organic matter composition in rivers of the contiguous United States

Jacob D. Hosen et al.

Summary: Most allochthonous organic matter in river networks comes from headwater streams during high flow events. As this material moves downstream, the composition of dissolved organic matter changes, with autochthonous characteristics becoming more prevalent. The time spent traveling through inland water systems appears to play a key role in determining DOM composition.

HYDROLOGICAL PROCESSES (2021)

Article Environmental Sciences

Development of a Deep Learning Emulator for a Distributed Groundwater-Surface Water Model: ParFlow-ML

Hoang Tran et al.

Summary: This study demonstrates how a deep learning model can emulate complex physical processes in hydrologic models at a fraction of the computational expense, with results that closely match those of the original model simulations. The emulator is up to 42 times faster than the original model and opens the door to future applications at larger scales.

WATER (2021)

Article Environmental Sciences

A Physics-Informed, Machine Learning Emulator of a 2D Surface Water Model: What Temporal Networks and Simulation-Based Inference Can Help Us Learn about Hydrologic Processes

Reed M. Maxwell et al.

Summary: This study demonstrates a 2D emulator using deep learning to predict hydrological problems, comparing different deep learning frameworks and hyperparameters. The research focuses on evaluating model performance and the match between machine learning calibrated input parameters and physics-based simulation.

WATER (2021)

Article Environmental Sciences

Substantial Declines in Salinity Observed Across the Upper Colorado River Basin During the 20th Century, 1929-2019

Christine A. Rumsey et al.

Summary: The study quantified long-term trends in salinity across the Upper Colorado River Basin and found that there have been large, widespread, and sustained downward trends in dissolved-solids concentrations and loads over the past 50-90 years, with the steepest rates of decline occurring from 1980 to 2000. The results also indicate that significant decreases in salinity occurred as early as the 1940s at some streams, suggesting that changes in land cover, land use, and/or climate have substantial effects on salinity transport in the UCRB.

WATER RESOURCES RESEARCH (2021)

Article Environmental Sciences

Assessment of groundwater well vulnerability to contamination through physics-informed machine learning

Mario A. Soriano et al.

Summary: The study utilized a hybrid approach combining physical-based and machine learning models to evaluate groundwater well vulnerability to contamination. Important predictors such as topography, hydrology, and proximity to contaminant sources were identified for accurate predictions. Results suggest that future groundwater contamination incidents are likely to increase if safeguards against contaminant releases are relaxed.

ENVIRONMENTAL RESEARCH LETTERS (2021)

Article Environmental Sciences

Vertical Connectivity Regulates Water Transit Time and Chemical Weathering at the Hillslope Scale

Dacheng Xiao et al.

Summary: The study shows that permeability variations in hillslope structure have significant effects on vertical connectivity, mean transit times, and chemical weathering rates. However, in arid climates, the influence of long transit times on weathering is less significant. Three hypotheses were proposed and it was highlighted that hillslope structure plays a crucial role in regulating the hydrogeochemical response of surface water to changing climate and human activities.

WATER RESOURCES RESEARCH (2021)

Review Computer Science, Information Systems

Machine Learning of Spatial Data

Behnam Nikparvar et al.

Summary: Spatially explicit data properties are often overlooked in machine learning for spatial domains; there is a lag in resources to discover and handle them. Research can be divided into handling properties in the observation matrix or learning algorithm; the latter, although less explored, holds promising prospects for the future of spatial machine learning.

ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION (2021)

Article Computer Science, Artificial Intelligence

Machine Learning Analysis of Hydrologic Exchange Flows and Transit Time Distributions in a Large Regulated River

Huiying Ren et al.

Summary: This study developed simplified models of hydrologic exchange flows and transit time distributions using machine learning methods, integrating field observations and simulation data. The results showed that Random Forest and Extreme Gradient Boosting models achieved 70% to 80% accuracy in the Columbia River Corridor, serving as effective alternatives to computationally demanding numerical models.

FRONTIERS IN ARTIFICIAL INTELLIGENCE (2021)

Review Astronomy & Astrophysics

Global Groundwater Sustainability, Resources, and Systems in the Anthropocene

Tom Gleeson et al.

ANNUAL REVIEW OF EARTH AND PLANETARY SCIENCES, VOL 48, 2020 (2020)

Article Geosciences, Multidisciplinary

Simulating Groundwater-Streamflow Connections in the Upper Colorado River Basin

Hoang Tran et al.

GROUNDWATER (2020)

Article Multidisciplinary Sciences

Global threat of arsenic in groundwater

Joel Podgorski et al.

SCIENCE (2020)

Article Statistics & Probability

Visualizing the effects of predictor variables in black box supervised learning models

Daniel W. Apley et al.

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY (2020)

Article Geosciences, Multidisciplinary

On the shape of forward transit time distributions in low-order catchments

Ingo Heidbuechel et al.

HYDROLOGY AND EARTH SYSTEM SCIENCES (2020)

Article Computer Science, Artificial Intelligence

From local explanations to global understanding with explainable AI for trees

Scott M. Lundberg et al.

NATURE MACHINE INTELLIGENCE (2020)

Article Engineering, Civil

Variability of transit time distributions with climate and topography: A modelling approach

Federica Remondi et al.

JOURNAL OF HYDROLOGY (2019)

Review Geochemistry & Geophysics

Global Isotope Hydrogeology?Review

Scott Jasechko

REVIEWS OF GEOPHYSICS (2019)

Review Geochemistry & Geophysics

The Demographics of Water: A Review of Water Ages in the Critical Zone

Matthias Sprenger et al.

REVIEWS OF GEOPHYSICS (2019)

Article Computer Science, Information Systems

The scale effects of the spatial autocorrelation measurement: aggregation level and spatial resolution

Boen Zhang et al.

INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE (2019)

Article Environmental Sciences

Sensitivity of flood damage estimation to spatial resolution

A. A. Komolafe et al.

JOURNAL OF FLOOD RISK MANAGEMENT (2018)

Article Engineering, Civil

Metamodeling and mapping of nitrate flux in the unsaturated zone and groundwater, Wisconsin, USA

Bernard T. Nolan et al.

JOURNAL OF HYDROLOGY (2018)

Article Environmental Sciences

Regionalization of Groundwater Residence Time Using Metamodeling

J. J. Starn et al.

WATER RESOURCES RESEARCH (2018)

Article Environmental Sciences

Metamodeling for Groundwater Age Forecasting in the Lake Michigan Basin

Michael N. Fienen et al.

WATER RESOURCES RESEARCH (2018)

Article Environmental Sciences

Performance of National Maps of Watershed Integrity at Watershed Scales

Anne Kuhn et al.

WATER (2018)

Article Geosciences, Multidisciplinary

Sensitivity of young water fractions to hydro-climatic forcing and landscape properties across 22 Swiss catchments

Jana von Freyberg et al.

HYDROLOGY AND EARTH SYSTEM SCIENCES (2018)

Article Geosciences, Multidisciplinary

Scale effect challenges in urban hydrology highlighted with a distributed hydrological model

Abdellah Ichiba et al.

HYDROLOGY AND EARTH SYSTEM SCIENCES (2018)

Article Environmental Sciences

How Misapplication of the Hydrologic Unit Framework Diminishes the Meaning of Watersheds

James M. Omernik et al.

ENVIRONMENTAL MANAGEMENT (2017)

Article Water Resources

The role of baseflow in dissolved solids delivery to streams in the Upper Colorado River Basin

Christine A. Rumsey et al.

HYDROLOGICAL PROCESSES (2017)

Article Geosciences, Multidisciplinary

Global aquifers dominated by fossil groundwaters but wells vulnerable to modern contamination

Scott Jasechko et al.

NATURE GEOSCIENCE (2017)

Article Geosciences, Multidisciplinary

Changes in groundwater recharge under projected climate in the upper Colorado River basin

Fred D. Tillman et al.

GEOPHYSICAL RESEARCH LETTERS (2016)

Article Geosciences, Multidisciplinary

The imprint of climate and geology on the residence times of groundwater

Reed M. Maxwell et al.

GEOPHYSICAL RESEARCH LETTERS (2016)

Article Environmental Sciences

Transit timesthe link between hydrology and water quality at the catchment scale

Markus Hrachowitz et al.

WILEY INTERDISCIPLINARY REVIEWS-WATER (2016)

Article Computer Science, Interdisciplinary Applications

Evaluating machine learning and statistical prediction techniques for landslide susceptibility modeling

J. N. Goetz et al.

COMPUTERS & GEOSCIENCES (2015)

Review Environmental Sciences

A review of surrogate models and their application to groundwater modeling

M. J. Asher et al.

WATER RESOURCES RESEARCH (2015)

Article Geosciences, Multidisciplinary

Are we missing the tail (and the tale) of residence time distributions in watersheds?

Marty D. Frisbee et al.

GEOPHYSICAL RESEARCH LETTERS (2013)

Article Geochemistry & Geophysics

The dependence of chemical weathering rates on fluid residence time

K. Maher

EARTH AND PLANETARY SCIENCE LETTERS (2010)

Article Water Resources

Catchment transit times and landscape controls-does scale matter?

M. Hrachowitz et al.

HYDROLOGICAL PROCESSES (2010)

Review Engineering, Civil

A review and evaluation of catchment transit time modeling

Kevin J. McGuire et al.

JOURNAL OF HYDROLOGY (2006)

Article Environmental Sciences

The role of topography on catchment-scale water residence time

KJ McGuire et al.

WATER RESOURCES RESEARCH (2005)

Article Computer Science, Artificial Intelligence

Random forests

L Breiman

MACHINE LEARNING (2001)