A review of hybrid deep learning applications for streamflow forecasting

Journal of Hydrology, Journal Year: 2023, Volume and Issue: 625, P. 130141 - 130141

Published: Sept. 12, 2023

Language: Английский

---

Application of Machine Learning in Water Resources Management: A Systematic Literature Review

Water, Journal Year: 2023, Volume and Issue: 15(4), P. 620 - 620

Published: Feb. 5, 2023

In accordance with the rapid proliferation of machine learning (ML) and data management, ML applications have evolved to encompass all engineering disciplines. Owing importance world’s water supply throughout rest this century, much research has been concentrated on application strategies integrated resources management (WRM). Thus, a thorough well-organized review that is required. To accommodate underlying knowledge interests both artificial intelligence (AI) unresolved issues in WRM, overview divides core fundamentals, major applications, ongoing into two sections. First, basic are categorized three main groups, prediction, clustering, reinforcement learning. Moreover, literature organized each field according new perspectives, patterns indicated so attention can be directed toward where headed. second part, less investigated WRM addressed provide grounds for future studies. The widespread tools projected accelerate formation sustainable plans over next decade.

Language: Английский

---

Deep transfer learning based on transformer for flood forecasting in data-sparse basins

Journal of Hydrology, Journal Year: 2023, Volume and Issue: 625, P. 129956 - 129956

Published: July 19, 2023

Language: Английский

---

Machine-learning- and deep-learning-based streamflow prediction in a hilly catchment for future scenarios using CMIP6 GCM data

Hydrology and earth system sciences, Journal Year: 2023, Volume and Issue: 27(5), P. 1047 - 1075

Published: March 13, 2023

Abstract. The alteration in river flow patterns, particularly those that originate the Himalaya, has been caused by increased temperature and rainfall variability brought on climate change. Due to impending intensification of extreme events, as predicted Intergovernmental Panel Climate Change (IPCC) its Sixth Assessment Report, it is more essential than ever predict changes streamflow for future periods. Despite fact some research utilised machine-learning- deep-learning-based models patterns response change, very few studies have undertaken a mountainous catchment, with number western Himalaya being minimal. This study investigates capability five different machine learning (ML) one deep (DL) model, namely Gaussian linear regression model (GLM), generalised additive (GAM), multivariate adaptive splines (MARSs), artificial neural network (ANN), random forest (RF), 1D convolutional (1D-CNN), prediction over Sutlej River basin during periods 2041–2070 (2050s) 2071–2100 (2080s). Bias-corrected data downscaled at grid resolution 0.25∘ × from six general circulation (GCMs) Coupled Model Intercomparison Project Phase 6 GCM framework under two greenhouse gas (GHG) trajectories (SSP245 SSP585) were used this purpose. Four scenarios (R0, R1, R2, R3) applied trained daily (1979–2009) Kasol (the outlet basin) order better understand how catchment size geo-hydromorphological aspects affect runoff. predictive power each was assessed using statistical measures, i.e. coefficient determination (R2), ratio root mean square error standard deviation measured (RSR), absolute (MAE), Kling–Gupta efficiency (KGE), Nash–Sutcliffe (NSE), percent bias (PBIAS). RF scenario R3, which outperformed other training (R2 = 0.90; RSR 0.32; KGE 0.87; NSE PBIAS 0.03) testing 0.78; 0.47; 0.82; 0.71; −0.31) period, therefore chosen simulate 2050s 2080s SSP245 SSP585 scenarios. Bias correction further projected generate reliable times series discharge. ensemble results shows annual expected rise between 0.79 % 1.43 0.87 1.10 SSP245. In addition, will increase monsoon (9.70 11.41 11.64 12.70 %) both emission scenarios, but decrease pre-monsoon (−10.36 −6.12 −10.0 −9.13 %), post-monsoon (−1.23 −0.22 −5.59 −2.83 winter (−21.87 −21.52 −21.87 −21.11 %). highly correlated pattern precipitation CMIP6 GCMs physical processes operating within catchment. Predicted declines (April June) (December March) seasons might significant impact agriculture downstream river, already having problems due water restrictions time year. present assist strategy planning ensure sustainable use resources acquiring knowledge nature causes unpredictable patterns.

Language: Английский

---

A Comprehensive Survey of Machine Learning Methodologies with Emphasis in Water Resources Management

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(22), P. 12147 - 12147

Published: Nov. 8, 2023

This paper offers a comprehensive overview of machine learning (ML) methodologies and algorithms, highlighting their practical applications in the critical domain water resource management. Environmental issues, such as climate change ecosystem destruction, pose significant threats to humanity planet. Addressing these challenges necessitates sustainable management increased efficiency. Artificial intelligence (AI) ML technologies present promising solutions this regard. By harnessing AI ML, we can collect analyze vast amounts data from diverse sources, remote sensing, smart sensors, social media. enables real-time monitoring decision making applications, including irrigation optimization, quality monitoring, flood forecasting, demand enhance agricultural practices, distribution models, desalination plants. Furthermore, facilitates integration, supports decision-making processes, enhances overall sustainability. However, wider adoption faces challenges, heterogeneity, stakeholder education, high costs. To provide an management, research focuses on core fundamentals, major (prediction, clustering, reinforcement learning), ongoing issues offer new insights. More specifically, after in-depth illustration algorithmic taxonomy, comparative mapping all specific tasks. At same time, include tabulation works along with some concrete, yet compact, descriptions objectives at hand. leveraging tools, develop plans address world’s supply concerns effectively.

Language: Английский

---

Spatio-temporal deep learning model for accurate streamflow prediction with multi-source data fusion

Environmental Modelling & Software, Journal Year: 2024, Volume and Issue: 178, P. 106091 - 106091

Published: May 28, 2024

Language: Английский

---

River stream flow prediction through advanced machine learning models for enhanced accuracy

Results in Engineering, Journal Year: 2024, Volume and Issue: 22, P. 102215 - 102215

Published: May 4, 2024

The Narmada River basin, a significant water resource in central India, plays crucial role supporting agricultural, industrial, and domestic supply. Effective management of this basin requires accurate streamflow forecasting, which has become increasingly important. This study delves into forecasting using historical data from five major river stations, covering the upper reaches East middle sections. dataset spans 1978 to 2020 undergoes rigorous screening preparation, including normalization StandardScaler. research adopts comprehensive approach, developing models for training on 70% data, validation most current 15%, testing against future targets with another 15% data. To achieve precise predictions, suite machine learning is employed, CatBoost, LGBM (Light Gradient Boosting Machine), Random Forest, XGBoost. Performance evaluation these relies key indices such as mean squared error (MSE), absolute (MAE), root square (RMSE), percent (RMSPE), normalized (NRMSE), R-squared (R2). Notably, among models, Forest emerges robust prediction, showcasing its effectiveness handling complexities hydrological forecasting. contributes significantly field by providing insights performance various models. findings not only enhance our understanding watershed dynamics but also highlight pivotal that can play improving sustainable management.

Language: Английский

---

A new frontier in streamflow modeling in ungauged basins with sparse data: A modified generative adversarial network with explainable AI

Results in Engineering, Journal Year: 2024, Volume and Issue: 21, P. 101920 - 101920

Published: Feb. 15, 2024

Streamflow forecasting is crucial for effective water resource planning and early warning systems, especially in regions with complex hydrological behaviors uncertainties. While machine learning (ML) has gained popularity streamflow prediction, many studies have overlooked the predictability of future events considering anthropogenic, static physiographic, dynamic climate variables. This study, first time, used a modified generative adversarial network (GAN) based model to predict streamflow. The training concept modifies enhances existing data embed featureful information enough capture extreme rather than generating synthetic instances. was trained using (sparse data) combination variables obtained from an ungauged basin monthly GAN-based interpreted time local interpretable model-agnostic explanations (LIME), explaining decision-making process model. achieved R2 0.933 0.942 0.93–0.94 testing. Also, testing period been reasonably well captured. LIME generally adhere physical provided by related work. approach looks promising as it worked sparse basin. authors suggest this research work that focuses on learning-based predictions.

Language: Английский

---

Ensemble learning using multivariate variational mode decomposition based on the Transformer for multi-step-ahead streamflow forecasting

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 636, P. 131275 - 131275

Published: May 7, 2024

Language: Английский

---

A comparative study of Machine Learning and Deep Learning methods for flood forecasting in the Far-North region, Cameroon

Scientific African, Journal Year: 2023, Volume and Issue: 23, P. e02053 - e02053

Published: Dec. 27, 2023

Flood crises are the consequence of climate change and global warming, which lead to an increase in frequency intensity heavy rainfall. Floods are, remain, natural disasters that result huge loss lives material damage. risks threaten all countries globe general. The Far-North region Cameroon has suffered flood on several occasions, resulting significant human lives, infrastructural socio-economic damage, with destruction homes, crops grazing areas, halting economic activities. models used for forecasting this generally physical-based, produce unsatisfactory results. use artificial intelligence based methods order limit its consequences is a way be explored Cameroon. aims present research work design compare performance Machine Learning Deep such as one dimensional Convolutional Neural Network, Long Short Term Memory Multi Layer Perceptron short-term long-term designed take input temperature rainfall time series recorded region. Performance criteria evaluating Nash–SutcliffeEfficiency, Percent Bias, Coefficient Determination Root Mean Squared Error. As results comparison models, best model LSTM , still model. obtained from comparisons have satisfactory good generalization capabilities, reflected by criteria. our can implementation floods warning systems definition effective efficient risk management policies make more resilient crises.

Language: Английский

---