Can machine learning accelerate process understanding and decision‐relevant predictions of river water quality?

Hydrological Processes, Journal Year: 2022, Volume and Issue: 36(4)

Published: March 29, 2022

Abstract The global decline of water quality in rivers and streams has resulted a pressing need to design new watershed management strategies. Water can be affected by multiple stressors including population growth, land use change, warming, extreme events, with repercussions on human ecosystem health. A scientific understanding factors affecting riverine predictions at local regional scales, sub‐daily decadal timescales are needed for optimal watersheds river basins. Here, we discuss how machine learning (ML) enable development more accurate, computationally tractable, scalable models analysis quality. We review relevant state‐of‐the art applications ML opportunities improve the emerging computational mathematical methods model selection, hyperparameter optimization, incorporating process knowledge into models, improving explainablity, uncertainty quantification, model‐data integration. then present considerations using address problems given their scale complexity, available data resources, stakeholder needs. When combined decades understanding, interdisciplinary advances knowledge‐guided ML, information theory, integration, analytics help fundamental science questions decision‐relevant

Language: Английский

---

Deep insight into daily runoff forecasting based on a CNN-LSTM model

Natural Hazards, Journal Year: 2022, Volume and Issue: 113(3), P. 1675 - 1696

Published: May 6, 2022

Language: Английский

---

Revolutionizing Groundwater Management with Hybrid AI Models: A Practical Review

Water, Journal Year: 2023, Volume and Issue: 15(9), P. 1750 - 1750

Published: May 2, 2023

Developing precise soft computing methods for groundwater management, which includes quality and quantity, is crucial improving water resources planning management. In the past 20 years, significant progress has been made in management using hybrid machine learning (ML) models as artificial intelligence (AI). Although various review articles have reported advances this field, existing literature must cover ML. This article aims to understand current state-of-the-art ML used achievements domain. It most cited employed from 2009 2022. summarises reviewed papers, highlighting their strengths weaknesses, performance criteria employed, highly identified. worth noting that accuracy was significantly enhanced, resulting a substantial improvement demonstrating robust outcome. Additionally, outlines recommendations future research directions enhance of including prediction related knowledge.

Language: Английский

---

Recent advancements of landslide hydrology

Wiley Interdisciplinary Reviews Water, Journal Year: 2023, Volume and Issue: 10(6)

Published: June 19, 2023

Abstract Occurrence of rainfall‐induced landslides is increasing worldwide, owing to land use and climate changes. Although the connection between hydrology might seem obvious, hydrological processes have been only marginally considered in landslide research for decades. In 2016, an advanced review paper published WIREs Water [Bogaard Greco (2016), , 3(3), 439–459] pointed out several challenging issues research: considering large‐scale assessment slope water balance; including antecedent information hazard assessment; understanding quantifying feedbacks deformation infiltration/drainage processes; overcoming conceptual mismatch soil mechanics models models. While little progress has made on latter two issues, a variety studies published, focusing role initiation prediction. The importance identification origin understand leading activation largely acknowledged. Techniques methodologies definition catchments balance are progressing fast, often hydraulic effect vegetation. prediction also progressed enormously. Empirical predictive tools, be implemented early warning systems shallow landslides, benefit from inclusion moisture, extracted different sources depending scale prediction, significant improvement their skill. However, this kind generally still missing operational LEWS. This article categorized under: Science > Hydrological Processes

Language: Английский

---

Forecasting of stage-discharge in a non-perennial river using machine learning with gamma test

Heliyon, Journal Year: 2023, Volume and Issue: 9(5), P. e16290 - e16290

Published: May 1, 2023

Knowledge of the stage-discharge rating curve is useful in designing and planning flood warnings; thus, developing a reliable fundamental crucial component water resource system engineering. Since continuous measurement often impossible, relationship generally used natural streams to estimate discharge. This paper aims optimize using generalized reduced gradient (GRG) solver test accuracy applicability hybridized linear regression (LR) with other machine learning techniques, namely, regression-random subspace (LR-RSS), regression-reduced error pruning tree (LR-REPTree), regression-support vector (LR-SVM) regression-M5 pruned (LR-M5P) models. An application these hybrid models was performed modeling Gaula Barrage problem. For this, 12-year historical data were collected analyzed. The daily flow (m3/s) stage (m) from during monsoon season, i.e., June October only 03/06/2007 31/10/2018, for discharge simulation. best suitable combination input variables LR, LR-RSS, LR-REPTree, LR-SVM, LR-M5P identified decided gamma test. GRG-based equations found be as effective more accurate conventional equations. outcomes GRG, compared observed values based on Nash Sutcliffe model efficiency coefficient (NSE), Willmott Index Agreement (d), Kling-Gupta (KGE), mean absolute (MAE), bias (MBE), relative percent (RE), root square (RMSE) Pearson correlation (PCC) determination (R2). LR-REPTree (combination 1: NSE = 0.993, d 0.998, KGE 0.987, PCC(r) 0.997, R2 0.994 minimum value RMSE 0.109, MAE 0.041, MBE −0.010 RE −0.1%; 2; 0.941, 0.984, 0. 923, 973, 947 331, 0.143, −0.089 −0.9%) superior all combinations testing period. It also noticed that performance alone LR its (i.e., LR-M5P) better than curve, including GRG method.

Language: Английский

---

A New Benchmark on Machine Learning Methodologies for Hydrological Processes Modelling: A Comprehensive Review for Limitations and Future Research Directions

Knowledge-Based Engineering and Sciences, Journal Year: 2023, Volume and Issue: 4(3), P. 65 - 103

Published: Dec. 31, 2023

The best practice of watershed management is through the understanding hydrological processes. As a matter fact, processes are highly associated with stochastic, non-linear, and non-stationary phenomena. Hydrological simulation modeling challenging issues in domains hydrology, climate environment. Hence, development machine learning (ML) models for solving those complex problems took essential place over past couple decades. It can be observed, data availability has increased remarkably, thus computational resources led to resurgence ML models’ development. been witnessed huge efforts on using facility several review researches have conducted. Literature studies approved capacity field hydrology classical “traditional models” based their forecastability, flexibility, precision, generalization, execution convergence speed. However, although potential merits were observed model’s development, limitations allied such as interpretability black-box models, practicality management, difficulty explain physical In this survey, an exhibition all published articles recognize research gaps direction. ultimate aim current survey establish new milestone interested environment researchers applications models.

Language: Английский

---

Application of machine learning and emerging remote sensing techniques in hydrology: A state-of-the-art review and current research trends

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 632, P. 130907 - 130907

Published: Feb. 16, 2024

Language: Английский

---

Toward interpretable LSTM-based modeling of hydrological systems

Hydrology and earth system sciences, Journal Year: 2024, Volume and Issue: 28(4), P. 945 - 971

Published: Feb. 27, 2024

Abstract. Several studies have demonstrated the ability of long short-term memory (LSTM) machine-learning-based modeling to outperform traditional spatially lumped process-based approaches for streamflow prediction. However, due mainly structural complexity LSTM network (which includes gating operations and sequential processing data), difficulties can arise when interpreting internal processes weights in model. Here, we propose test a modification architecture that is calibrated manner analogous hydrological system. Our architecture, called “HydroLSTM”, simulates updating Markovian storage while operation has access historical information. Specifically, modify how data are fed new representation facilitate simultaneous past lagged inputs consolidated information, which explicitly acknowledges importance trends patterns data. We compare performance HydroLSTM architectures using from 10 hydro-climatically varied catchments. further examine exploits information inputs, 588 catchments across USA. The HydroLSTM-based models require fewer cell states obtain similar their LSTM-based counterparts. Further, weight associated with input variables interpretable consistent regional hydroclimatic characteristics (snowmelt-dominated, recent rainfall-dominated, rainfall-dominated). These findings illustrate interpretability be enhanced by appropriate architectural modifications physically conceptually our understanding

Language: Английский

---

Cryosphere–groundwater connectivity is a missing link in the mountain water cycle

Nature Water, Journal Year: 2024, Volume and Issue: 2(7), P. 624 - 637

Published: July 19, 2024

Language: Английский

---

Advancing Hydrology through Machine Learning: Insights, Challenges, and Future Directions Using the CAMELS, Caravan, GRDC, CHIRPS, PERSIANN, NLDAS, GLDAS, and GRACE Datasets

Water, Journal Year: 2024, Volume and Issue: 16(13), P. 1904 - 1904

Published: July 3, 2024

Machine learning (ML) applications in hydrology are revolutionizing our understanding and prediction of hydrological processes, driven by advancements artificial intelligence the availability large, high-quality datasets. This review explores current state ML hydrology, emphasizing utilization extensive datasets such as CAMELS, Caravan, GRDC, CHIRPS, NLDAS, GLDAS, PERSIANN, GRACE. These provide critical data for modeling various parameters, including streamflow, precipitation, groundwater levels, flood frequency, particularly data-scarce regions. We discuss type methods used significant successes achieved through those models, highlighting their enhanced predictive accuracy integration diverse sources. The also addresses challenges inherent applications, heterogeneity, spatial temporal inconsistencies, issues regarding downscaling LSH, need incorporating human activities. In addition to discussing limitations, this article highlights benefits utilizing high-resolution compared traditional ones. Additionally, we examine emerging trends future directions, real-time quantification uncertainties improve model reliability. place a strong emphasis on citizen science IoT collection hydrology. By synthesizing latest research, paper aims guide efforts leveraging large techniques advance enhance water resource management practices.

Language: Английский

---