A hybrid monthly hydrological prediction model based on LSTM-EBLS and improved VMD DOI Creative Commons
Boya Zhou, Lehao Wang, Ying Han

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 11, 2024

Abstract Scarce of large sample data makes deep learning based monthly hydrological prediction still challenging. Compared with methods, broad learn-ing system (BLS) has the advantages fast operation and small suita-bility. While, using BLS alone to predict, accuracy is relatively low. Using weights between input vector output gate in long short-term memory (LSTM) as initial BLS, extended (EBLS) constructed temporal feature extraction module for prediction. Considering time-consuming problem resulting by variational mode decomposition (VMD), an improved version VMD (IVMD) presented this paper. Finally, a hybrid forecast model on LSTM, EBLS IVMD proposed. The trained validated prediction, results demonstrated that: (1) For multi-month ahead outperforms discussed state art models. Meawhile, peak fitting also enhanced. (2) CNN-LSTM structure, LSTM-EBLS improves accuracy. (3) Efficient parameter selection method high correlation signals further enhance computation efficiency.

Language: Английский

The dynamics of lowland river sections of Danube and Tisza in the Carpathian basin DOI Creative Commons
Imre M. Jánosi,

István Zsuffa,

Tibor Bíró

et al.

Frontiers in Earth Science, Journal Year: 2025, Volume and Issue: 13

Published: Feb. 12, 2025

The paper presents a detailed statistical analysis of data from 41 hydrometric stations along the Danube (section in Carpathian Basin) and its longest tributary, Tisza River. Most records cover 2–3 decades with an automated high temporal sampling frequency (15 min), few span 120 years daily or half-daily records. is not even exhibits strong irregularities. demonstrates that cubic spline fits down-sampling (where necessary) produce reliable, evenly sampled time series smoothly reconstruct water level river discharge data. Almost all indicate decadal decreasing trend for annual maximum values. timing (day year) maxima minima evaluated. While minimum values do show coherent tendencies, exhibit increasing trends but (earlier onset). Various possibilities explanations these observations are listed. empirical histograms changes can be well-fitted by piecewise-exponential functions containing four three sections, consistent understanding deterministic rather than stochastic processes, as well known hydrology. Such tests serve benchmarks modeling levels discharges. Extracted periods Lomb-Scargle algorithm (suitable unevenly series) long-time means expected seasonality. Resampled (1-hour frequency) were evaluated standard Fourier Welch procedures, revealing some secondary peaks spectra indicating quasi-periodic components signals. Further significance progress, attempts at explanations. Secondary may environmental changes, future investigation which could reveal important correlations.

Language: Английский

Citations

0
MVIE-LSTM: a deep learning-based method for water quality assessment using monthly river data DOI

Sha Xiong,

Junjie Cui, Feifei Hou

et al.

Stochastic Environmental Research and Risk Assessment, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 26, 2025

Language: Английский

Citations

0
TSWS: An observation-based streamflow dataset of Tianshan Mountains watersheds (1901–2019) DOI Creative Commons
Shuai Li, Wei Wei, Yaning Chen

et al.

Scientific Data, Journal Year: 2025, Volume and Issue: 12(1)

Published: April 29, 2025

Due to scarcity of data and complex hydrological conditions in the Tianshan region, long-term complete streamflow are lacking. This study produced a multi-basin dataset, named watershed (TSWS) by comparing results Hydrologiska Byråns Vattenavdelning Long Short-Term Memory models, analyzed spatiotemporal variation streamflow.TSWS dataset provides daily for 56 watersheds monthly 89 Mountains 1901-2019. The simulations 40 (daily scale) 70 (monthly passed S-tests (Nash-Sutcliffe efficiency ≥0.5, percent bias ≤25%, ratio root-mean-square error standard deviation measured ≤0.7). showed an overall increasing trend streamflow, especially from 1990 2019; spatially, it higher west south, lower east north. first comprehensive simulation its long time series will provide important reference climatic studies.

Language: Английский

Citations

0
Effects of Climate Change and Human Activities on Runoff in the Upper Reach of Jialing River, China DOI Creative Commons

Wei-Zhao Shi,

Yi He, Yiting Shao

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(13), P. 2481 - 2481

Published: July 6, 2024

In recent years, the runoff of numerous rivers has experienced substantial changes owing to dual influences climate change and human activities. This study focuses on Lixian hydrological station’s controlled basin, located in upper reaches Jialing River China. The objective is assess quantify impacts activities variations. analyzed variations from 1960 2016 employed Soil Water Assessment Tool (SWAT) model, long short-term memory (LSTM) eight Budyko framework formulations factors influencing runoff. Additionally, it used patch-generating land use simulation (PLUS) SWAT models simulate future scenarios under various conditions. results indicate following. (1) area witnessed a significant decline (p < 0.01), while potential evapotranspiration shows upward trend 0.01). Precipitation displays nonsignificant decreasing > 0.1). An abrupt point occurred 1994, dividing period into baseline periods. (2) reveal that contributed 50% 60% changes. According LSTM models, contribution rates are 63.21% 52.22%, respectively. Human thus identified as predominant factor (3) primarily influence through cover Conservation measures led notable increase forested areas 1990 2010, representing most among types. (4) Future suggest highest simulated occurs comprehensive development scenario, lowest observed an ecological conservation scenario. Among 32 scenarios, increases significantly with 10% precipitation decreases substantially 15% reduction precipitation. These findings underscore impact River, highlighting importance incorporating both water resource management planning.

Language: Английский

Citations

3
Filling the gap between GRACE and GRACE-FO data using a model integrating variational mode decomposition and long short-term memory: a case study of Northwest China DOI
Jiangdong Chu, Xiaoling Su, Tianliang Jiang

et al.

Environmental Earth Sciences, Journal Year: 2023, Volume and Issue: 82(1)

Published: Jan. 1, 2023

Language: Английский

Citations

9
A Comparative Analysis of Multiple Machine Learning Methods for Flood Routing in the Yangtze River DOI Open Access
Liwei Zhou,

Ling Kang

Water, Journal Year: 2023, Volume and Issue: 15(8), P. 1556 - 1556

Published: April 15, 2023

Obtaining more accurate flood information downstream of a reservoir is crucial for guiding regulation and reducing the occurrence disasters. In this paper, six popular ML models, including support vector regression (SVR), Gaussian process (GPR), random forest (RFR), multilayer perceptron (MLP), long short-term memory (LSTM) gated recurrent unit (GRU) were selected compared their effectiveness in routing two complicated reaches located at upper middle main stream Yangtze River. The results suggested that performance MLP, LSTM GRU models all gradually improved then slightly decreased as time lag increased. Furthermore, outperformed SVR, GPR RFR model demonstrated superior across range efficiency criteria, mean absolute percentage error (MAPE), root square (RMSE), Nash–Sutcliffe coefficient (NSE), Taylor skill score (TSS) Kling–Gupta (KGE). Specifically, achieved reductions MAPE RMSE least 7.66% 3.80% first case study 19.51% 11.76% second study. paper indicated was most appropriate choice

Language: Английский

Citations

9
Review of Machine Learning Methods for River Flood Routing DOI Open Access

Li Li,

Kyung Soo Jun

Published: Jan. 3, 2024

River flood routing computes changes in shape of a wave over time as it travels downstream along river. Conventional models, especially hydrodynamic models require high quality and quantity input data such measured hydrologic series, geometric data, hydraulic structures hydrological parameters. Unlike physically based machine learning algorithms, which are driven do not much knowledge about underlying physical processes can identify complex nonlinearity between inputs outputs. Due to the higher performance, less complexity, low computation cost, novel methods single application or hybrid were introduced by researchers achieve more accurate efficient routing. This paper reviews recent river

Language: Английский

Citations

2
Quantifying the impacts of climate change and human activities on ecological flow security based on a new framework DOI
Hongxiang Wang, Siyuan Cheng,

Xiangyu Bai

et al.

Ecohydrology, Journal Year: 2024, Volume and Issue: 17(6)

Published: April 24, 2024

Abstract Climate change and human activities combine to alter river hydrology, thereby threatening the health of ecosystems. Quantifying impacts climate on ecological flow assurance is essential for water resource management protection. However, fewer studies quantify based a complete set frameworks. The present study introduces an integrated assessment framework designed security. includes following steps: (1) natural runoff reconstruction utilizing semi‐distributed hydrological model (SWAT), (2) calculation most suitable stream watershed ecosystem by using non‐parametric kernel density estimation method, (3) safety security levels under minimum appropriate conditions in (4) quantification influences through application quantitative attribution method. impact level was analysed three stations Xiangtan, Hengyang Laobutou, which are main tributaries Xiangjiang River Basin, as case study. findings indicated substantial decrease across basin during period (1991–2019). results suggest that predominantly drive degradation throughout impact, accounting 57.05% total impact. Extensive gradient reservoir scheduling anthropogenic withdrawals were factors contributing basin. methodology presented this offer insights into evolutionary characteristics driving forces behind dynamic environment. Furthermore, they establish scientific foundation local

Language: Английский

Citations

2
An improved nonlinear dynamical model for monthly runoff prediction for data scarce basins DOI
Longxia Qian,

Nanjun Liu,

Mei Hong

et al.

Stochastic Environmental Research and Risk Assessment, Journal Year: 2024, Volume and Issue: 38(10), P. 3771 - 3798

Published: Aug. 17, 2024

Language: Английский

Citations

2
Drought driving mechanism and risk situation prediction based on machine learning models in the Yellow River Basin, China DOI Creative Commons

Ling Kang,

Yunliang Wen,

Liwei Zhou

et al.

Geomatics Natural Hazards and Risk, Journal Year: 2023, Volume and Issue: 14(1)

Published: Nov. 10, 2023

Under global warming, the acceleration of water cycle has increased risk drought in Yellow River Basin. Revealing driving mechanisms basin and understanding situation have become particularly important. This paper uses wavelet analysis transfer entropy to analyze mechanisms. In addition, an Improved Particle Swarm Optimization (IPSO) coupled with Long Short-Term Memory (LSTM) is used for prediction. The results are as follows: (1) Hydrological lags behind meteorological by 2–3 months, they show two main periods on different time scales, which 5–6 months 8–14 respectively. (2) Rainfall, runoff, temperature, humidity, vapor pressure factors, rainfall humidity having most significant impact. (3) IPSO-LSTM model improved process selecting parameters based empirical experiences LSTM model, improving prediction accuracy average 3.1%. provides a scientific basis resource management assessment basin, better cope future climate challenges.

Language: Английский

Citations

6