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

---

Performance analysis and modelling of circular jets aeration in an open channel using soft computing techniques

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Feb. 7, 2024

Abstract Dissolved oxygen (DO) is an important parameter in assessing water quality. The reduction DO concentration the result of eutrophication, which degrades quality water. Aeration best way to enhance concentration. In current study, aeration efficiency (E 20 ) various numbers circular jets open channel was experimentally investigated for different angle inclination (θ), discharge (Q), number (J n ), Froude ( Fr and hydraulic radius each jet (HR Jn ). statistical results show that from 8 64 significantly provide channel. input parameters are modelled into a linear relationship. Additionally, utilizing WEKA software, three soft computing models predicting were created with Artificial Neural Network (ANN), M5P, Random Forest (RF). Performance evaluation box plot have shown ANN outperforming model correlation coefficient (CC) = 0.9823, mean absolute error (MAE) 0.0098, root square (RMSE) 0.0123 during testing stage. order assess influence factors on E jets, sensitivity analysis conducted using most effective model, i.e., ANN. indicate influential variable , followed by jets.

Language: Английский

---

The Modeling of a River Impacted with Tailings Mudflows Based on the Differentiation of Spatiotemporal Domains and Assessment of Water–Sediment Interactions Using Machine Learning Approaches

Water, Journal Year: 2024, Volume and Issue: 16(3), P. 379 - 379

Published: Jan. 23, 2024

The modeling of metal concentrations in large rivers is complex because the contributing factors are numerous, namely, variation sources across spatiotemporal domains. By considering both domains, this study modeled derived from interaction river water and sediments contrasting grain size chemical composition, regions seasonal precipitation. Statistical methods assessed processes partitioning transport, while artificial intelligence structured dataset to predict evolution as a function environmental changes. methodology was applied Paraopeba River (Brazil), divided into sectors coarse aluminum-rich natural enriched fine iron- manganese-rich mine tailings, after collapse B1 dam Brumadinho, with 85–90% rainfall occurring October March. prediction capacity random forest regressor for aluminum, iron manganese concentrations, average precision > 90% accuracy < 0.2.

Language: Английский

---

Assessing the reliability of a physical-based model and a convolutional neural network in an ungauged watershed for daily streamflow calculation: a case study in southern Portugal

Environmental Earth Sciences, Journal Year: 2024, Volume and Issue: 83(7)

Published: March 25, 2024

Abstract The main goal of this study was to estimate inflows the Maranhão reservoir, southern Portugal, using two distinct modeling approaches: a one-dimensional convolutional neural network (1D-CNN) model and physically based model. 1D-CNN previously trained, validated, tested in sub-basin area where observed streamflow values were available. trained here subject an improvement applied entire watershed by replacing forcing variables (accumulated delayed precipitation) make them correspond watershed. same way, MOHID-Land calibrated validated for sub-basin, parameters then Inflow estimated both models considering mass balance at reservoir. demonstrated better performance simulating daily values, peak flows, wet period. showed estimating during dry periods monthly analysis. Hence, results show adequateness solutions integrating decision support system aimed supporting decision-makers management water availability subjected increasing scarcity.

Language: Английский

---

Ensemble learning of decomposition-based machine learning models for multistep-ahead daily streamflow forecasting in northwest China

Hydrological Sciences Journal, Journal Year: 2024, Volume and Issue: 69(11), P. 1501 - 1522

Published: July 1, 2024

Accurate daily streamflow forecasts remain challenging in arid regions. A Bayesian Model Averaging (BMA) ensemble learning strategy was proposed to forecast 1-, 2-, and 3-day ahead Dunhuang Oasis, northwest China. The efficiency of BMA compared with four decomposition-based machine deep models. Satisfactory were achieved all models at lead times; however, based on NSE values 0.976, 0.967, 0.957, the greatest accuracy for forecasts, respectively. Uncertainty analysis confirmed reliability yielding consistently accurate forecasts. Thus, could provide an efficient alternative approach multistep-ahead forecasting. incorporation data decomposition techniques (e.g. Variational mode decomposition) algorithms Deep belief network) into BMA, may serve as worthy technical references supervised systems scare

Language: Английский

---

Comparison of Multilayer Perceptron with an Optimal Activation Function and Long Short-Term Memory for Rainfall-Runoff Simulations and Ungauged Catchment Runoff Prediction

Water Resources Management, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 30, 2024

Language: Английский

---

Spatial Downscaling of Streamflow Data with Attention Based Spatio-Temporal Graph Convolutional Networks

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: June 12, 2023

Abstract Accurate streamflow data is vital for various climate modeling applications, including flood forecasting. However, many streams lack sufficient monitoring due to the high operational costs involved. To address this issue and promote enhanced disaster preparedness, management, response, our study introduces a neural network-based method estimating historical hourly in two spatial downscaling scenarios. The targets types of ungauged locations: (1) those without sensors sparsely gauged river networks, (2) that previously had sensor, but gauge no longer available. For both cases, we propose ScaleGNN, graph network architecture. We evaluate performance ScaleGNN against Long Short-Term Memory (LSTM) baseline persistence discharge values over 36-hour period. Our findings indicate surpasses first scenario, while approaches demonstrate their effectiveness compared second scenario.

Language: Английский

---

Spatial Downscaling of Streamflow Data with Attention Based Spatio-Temporal Graph Convolutional Networks

EarthArXiv (California Digital Library), Journal Year: 2023, Volume and Issue: unknown

Published: March 31, 2023

Accurate streamflow data is vital for various climate modeling applications, including flood forecasting. However, many streams lack sufficient monitoring due to the high operational costs involved. To address this issue and promote enhanced disaster preparedness, management, response, our study introduces a neural network-based method estimating historical hourly in two spatial downscaling scenarios. The targets types of ungauged locations: (1) those without sensors sparsely gauged river networks, (2) that previously had sensor, but gauge no longer available. For both cases, we propose ScaleGNN, graph network based on Attention-Based Spatio-Temporal Graph Convolutional Networks (ASTGCN). We evaluate performance ScaleGNN against Long Short-Term Memory (LSTM) baseline persistence discharge values over 36-hour period. Our findings indicate surpasses first scenario, while approaches demonstrate their effectiveness compared second scenario.

Language: Английский

---

Direct integration of reservoirs' operations in a hydrological model for streamflow estimation: coupling a CLSTM model with MOHID-Land

Hydrology and earth system sciences, Journal Year: 2023, Volume and Issue: 27(21), P. 3875 - 3893

Published: Nov. 2, 2023

Abstract. Knowledge about streamflow regimes and values is essential for different activities situations in which justified decisions must be made. However, behavior commonly assumed to non-linear, being controlled by various mechanisms that act on temporal spatial scales, making its estimation challenging. An example the construction operation of infrastructures such as dams reservoirs rivers. The challenges faced modelers correctly describe impact hydrological systems are considerable. In this study, an already implemented solution MOHID-Land (where MOHID stands HYDrodinamic MOdel, or MOdelo HIDrodinâmico Portuguese) model a natural flow regime Ulla River basin was considered baseline. watershed referred includes three reservoirs. Outflow were estimated considering basic rule two them (run-of-the-river dams) data-driven convolutional long short-term memory (CLSTM) type other (high-capacity dam). outflow obtained with CLSTM imposed model, while fed level inflow reservoir. This coupled system evaluated daily using hydrometric stations located downstream reservoirs, resulting improved performance compared baseline application. analysis modeled without further demonstrated dams' operations resulted increase during dry season decrease wet but no differences average streamflow. thus promising improving estimates modified catchments.

Language: Английский

---

Modelagem de redes neurais artificiais MLP para previsão de vazões na bacia do rio Miranda afluente do Pantanal

Published: Feb. 16, 2024

---