Enhancing river flow predictions: Comparative analysis of machine learning approaches in modeling stage-discharge relationship

Results in Engineering, Год журнала: 2024, Номер 22, С. 102017 - 102017

Опубликована: Март 19, 2024

Streamflow, a pivotal variable in water resources management, holds profound significance shaping the decision-making processes of hydrologic projects. This paper tries to delve into exploration stage-discharge relationship using three machine learning methods (MLMs) namely multi-layer neural networks (MLNN), radial basis (RBNN), and neuro-fuzzy systems (ANFIS) predict simulate mean daily data derived from two monitoring stations, Bulakbasi Karaozü, Kizilirmak River, Turkey. Root square error (RMSE), Mean absolute percentage (MAPE), coefficient determination (R2), Developed Discrepancy Ratio (DDR) metrics were utilized MLMs' performance assessment. The evaluation indices (RMSE, MAEP, R2, DDR) for preeminent MLNN model applied Bulakhbashi Karasu stations determined as (0.29, 1.57, 0.9998, 17.62) (1.71, 6.56, 0.9980, 6.65), respectively. contributed notable enhancement RMSE index aforementioned exhibiting improvements 87% 56%, These results affirm MLNN's proficiency accurately capturing at both stations.

Язык: Английский

---

Enhanced streamflow forecasting using hybrid modelling integrating glacio-hydrological outputs, deep learning and wavelet transformation

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 22, 2025

Abstract Understanding snow and ice melt dynamics is vital for flood risk assessment effective water resource management in populated river basins sourced inaccessible high-mountains. This study provides an AI-enabled hybrid approach integrating glacio-hydrological model outputs (GSM-SOCONT), with different machine learning deep techniques framed as alternative ‘computational scenarios, leveraging both physical processes data-driven insights enhanced predictive capabilities. The standalone (CNN-LSTM), relying solely on meteorological data, outperformed its counterpart equivalents. Hybrid models (CNN-LSTM1 to CNN-LSTM15) were trained using data augmented representing snow-melt contributions streamflow. (CNN-LSTM14), only glacier-derived features, performed best high NSE (0.86), KGE (0.80), R (0.93) values during calibration, the highest (0.83), (0.88), (0.91), lowest RMSE (892) MAE (544) validation. Finally, a multi-scale analysis feature permutations was explored wavelet transformation theory, these into final (CNN-LSTM19), which significantly enhances accuracy, particularly high-flow events, evidenced by improved (from 0.83 0.97) reduced 892 442) comparative illustrates how AI-enhanced hydrological improve accuracy of runoff forecasting provide more reliable actionable managing resources mitigating risks - despite paucity direct measurements.

Язык: Английский

---

Analyzing spatio-temporal variability of aquatic productive components in Northern Bay of Bengal using advanced machine learning models

Ocean & Coastal Management, Год журнала: 2024, Номер 251, С. 107074 - 107074

Опубликована: Март 1, 2024

Язык: Английский

---

Robust clustering-based hybrid technique enabling reliable reservoir water quality prediction with uncertainty quantification and spatial analysis

Journal of Environmental Management, Год журнала: 2024, Номер 362, С. 121259 - 121259

Опубликована: Июнь 1, 2024

Язык: Английский

---

Novel time-lag informed deep learning framework for enhanced streamflow prediction and flood early warning in large-scale catchments

Journal of Hydrology, Год журнала: 2024, Номер 631, С. 130841 - 130841

Опубликована: Фев. 5, 2024

Constrained by the sparsity of observational streamflow data, large-scale catchments face pressing challenges in prediction and flood management amid climate change. Deep learning excels simulation performance while flow lag information data-driven approaches is barely highlighted. In this study, we introduce a time-lag informed deep framework for catchments. Central to utilization between upstream downstream subbasins, enabling precise forecasting at outlet driven data. Taking monsoon-influenced Dulong-Irrawaddy River Basin (DIRB) as study area, determined peak (PFL) days relative annual scale (RAFS) defined subbasins. By incorporating with historical data different time intervals, developed optimal model DIRB. This was then applied evaluate processes 2008 2009, using selected indicators. The results indicate that led significant improvements, notably LSTM_PFL_RAFS Hkamti sub-basin which achieved Kling-Gupta Efficiency (KGE) 0.891 (Nash-Sutcliffe efficiency coefficient, NSE, 0.904), surpassing LSTM's 0.683 (NSE, 0.785). Further integration specific interval, model, H(15)_PFL utilizes reached an impressive KGE 0.948 0.940). outperformed standard LSTM accurately simulating key characteristics, including flows, initiation times, durations 2009 events. Notably, provides valuable 15-day lead forecasting, extending window emergency response preparations. Future research incorporates additional essential catchment features into holds great potential unraveling complex mechanisms hydrological responses human activities

Язык: Английский

---

Spatiotemporal assessment of groundwater quality and quantity using geostatistical and ensemble artificial intelligence tools

Journal of Environmental Management, Год журнала: 2024, Номер 355, С. 120495 - 120495

Опубликована: Март 1, 2024

Язык: Английский

---

A hybrid deep learning approach for streamflow prediction utilizing watershed memory and process-based modeling

Hydrology Research, Год журнала: 2024, Номер 55(4), С. 498 - 518

Опубликована: Март 25, 2024

Abstract Accurate streamflow prediction is essential for optimal water management and disaster preparedness. While data-driven methods’ performance often surpasses process-based models, concerns regarding their ‘black-box’ nature persist. Hybrid integrating domain knowledge process modeling into a framework, offer enhanced capabilities. This study investigated watershed memory modeling-based hybridizing approaches across diverse hydrological regimes – Korean Ethiopian watersheds. Following analysis, the Soil Water Assessment Tool (SWAT) was calibrated using recession constant other relevant parameters. Three hybrid incorporating residual error, were developed evaluated against standalone long short-term (LSTM) models. Hybrids outperformed LSTM all The memory-based approach exhibited superior consistent training, evaluation periods, regions, achieving 17–66% Nash–Sutcliffe efficiency coefficient improvement. error-based technique showed varying regions. hybrids improved extreme event predictions, particularly peak flows, models struggled at low flow. watersheds’ significant improvements highlight models’ effectiveness in regions with pronounced temporal variability. underscores importance of selecting specific based on desired objectives rather than solely relying statistical metrics that reflect average performance.

Язык: Английский

---

Integrating machine learning with process-based glacio-hydrological model for improving the performance of runoff simulation in cold regions

Journal of Hydrology, Год журнала: 2025, Номер unknown, С. 132963 - 132963

Опубликована: Фев. 1, 2025

Язык: Английский

---

Evaluating agricultural non-point source pollution with high-resolution remote sensing technology and SWAT model: A case study in Ningxia Yellow River Irrigation District, China

Ecological Indicators, Год журнала: 2024, Номер 166, С. 112578 - 112578

Опубликована: Сен. 1, 2024

Язык: Английский

---

Analysis of the responses of surface water resources to climate change in arid and semi-arid area

Agricultural Water Management, Год журнала: 2024, Номер 295, С. 108751 - 108751

Опубликована: Март 2, 2024

About 70% of water withdrawals in the Yellow River Basin (YRB) are used for irrigation, and deeply explanation effects climate change on runoff YRB provides a guarantee agricultural production. Analysis prediction were implemented according to meteorological hydrological data from 1967 2016, responses catchments six stations different combinations precipitation temperature conditions explored adopting Budyko framework, then results based scenario simulation elasticity compared. Our revealed that indicated downward trend (p>0.05) while showed upward (p<0.01), both which predicted climb future; sensitivities upstream, midstream downstream gradually increased catchment characteristics acted decisive role determining rather than climatic factors, generally, by 17.1–30.2% with only 10% increase decreased 4.2%-12.4% 1℃ temperature; compared elasticity, tend be more accurate as it captured changes when change. The this study provide foundation regional development utilization resources under influences

Язык: Английский

---