Rivers increasingly warmer: Prediction of changes in the thermal regime of rivers in Poland

Journal of Geographical Sciences, Год журнала: 2025, Номер 35(1), С. 139 - 172

Опубликована: Янв. 1, 2025

Язык: Английский

---

Prediction of daily river water temperatures using an optimized model based on NARX networks

Ecological Indicators, Год журнала: 2024, Номер 161, С. 111978 - 111978

Опубликована: Апрель 1, 2024

Water temperature is an important physical indicator of rivers because it impacts many other and biogeochemical processes controls the metabolism aquatic species in rivers. Having a good knowledge river thermal dynamics great importance. In this study, advanced machine learning based model that fast, accurate easy to use, namely nonlinear autoregressive network with exogenous inputs (NARX) neural network, was coupled Bayesian Optimization (BO) algorithm for optimizing number NARX hidden nodes lagged input/target values Regularization (BR) backpropagation training, forecast daily water temperatures (RWT). Long-term observed data from 18 Vistula River Basin, one largest Europe, were used testing, performance compared air2stream model. The results showed NARX-based performs significantly better than calibration validation stages, can capture seasonal pattern peak RWT. Input combinations impact RWT modeling, air day year (DOY) are major inputs, while streamflow rainfall play minor role on modeling at Basin. Considering future times series easily accessible climate models DOY be considered model, serve as promising tool investigate change dynamics.

Язык: Английский

---

Development of improved deep learning models for multi-step ahead forecasting of daily river water temperature

Engineering Applications of Computational Fluid Mechanics, Год журнала: 2025, Номер 19(1)

Опубликована: Янв. 17, 2025

Язык: Английский

---

Integrated metaheuristic algorithms with extreme learning machine models for river streamflow prediction

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Июнь 12, 2024

Accurate river streamflow prediction is pivotal for effective resource planning and flood risk management. Traditional forecasting models encounter challenges such as nonlinearity, stochastic behavior, convergence reliability. To overcome these, we introduce novel hybrid that combine extreme learning machines (ELM) with cutting-edge mathematical inspired metaheuristic optimization algorithms, including Pareto-like sequential sampling (PSS), weighted mean of vectors (INFO), the Runge-Kutta optimizer (RUN). Our comparative assessment includes 20 across eight categories, using data from Aswan High Dam on Nile River. findings highlight superior performance mathematically based models, which demonstrate enhanced predictive accuracy, robust convergence, sustained stability. Specifically, PSS-ELM model achieves a root square error 2.0667, Pearson's correlation index (R) 0.9374, Nash-Sutcliffe efficiency (NSE) 0.8642. Additionally, INFO-ELM RUN-ELM exhibit absolute percentage errors 15.21% 15.28% respectively, 1.2145 1.2105, high Kling-Gupta efficiencies values 0.9113 0.9124, respectively. These suggest adoption our proposed significantly enhances water management strategies reduces any risks.

Язык: Английский

---

A transfer learning-based long short-term memory model for the prediction of river water temperature

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 133, С. 108605 - 108605

Опубликована: Май 14, 2024

Язык: Английский

---

A long-term multivariate time series prediction model for dissolved oxygen

Ecological Informatics, Год журнала: 2024, Номер 82, С. 102695 - 102695

Опубликована: Июнь 20, 2024

Accurate and efficient long-term prediction of marine dissolved oxygen (DO) is crucial for the sustainable development aquaculture. However, multidimensional time dependency lag effects chemical variables present significant challenges when handling multiple inputs in univariate tasks. To address these issues, we designed a multivariate time-series model (LMFormer) based on Transformer architecture. The proposed decomposition strategy effectively leverages feature information at different scales, thereby reducing loss critical information. Additionally, dynamic variable selection gating mechanism was to optimize collinearity problem data extraction process. Finally, an two-stage attention architecture capture long-range dependencies between features. This study conducted high-precision 7-day advance DO predictions two case studies, environmentally stable Shandong Peninsula China San Juan Islands United States, which are affected by extreme conditions such as ocean currents. experimental results demonstrate superior performance generalizability model. In case, mean absolute error (MAE), root square (RMSE), coefficient determination (R2), Kling–Gupta efficiency (KGE) reached 0.0159, 0.126, 0.9743, 0.9625, respectively. MAE reduced average 42.34% compared that baseline model, RMSE 24.57%, R2 increased 22.54%, KGE improved 12.04%. Overall, achieves data, providing valuable references management decision-making

Язык: Английский

---

Forecasting Model for Danube River Water Temperature Using Artificial Neural Networks

Hydrology, Год журнала: 2025, Номер 12(2), С. 21 - 21

Опубликована: Янв. 21, 2025

The objective of this paper is to propose an artificial neural network (ANN) model forecast the Danube River temperature at Chiciu–Călărași, Romania, bordered by Romanian and Bulgarian ecological sites, situated upstream Cernavoda nuclear power plant. Given increase trend, potential thermal pollution rising, impacting aquatic terrestrial ecosystems. available data covered a period eight years, between 2008 2015. Using as input actual air water temperatures, discharge, well provided weather forecasts, ANN predicts one week in advance with root mean square deviation (RMSE) 0.954 °C for training 0.803 testing. uses Levenberg–Marquardt feedforward backpropagation algorithm. This feature useful irrigation systems plants area that use river different purposes. results are encouraging developing similar studies other locations extending include more parameters can have significant influence on temperature.

Язык: Английский

---

Research on Atlantic surface pCO2 reconstruction based on machine learning

Ecological Informatics, Год журнала: 2025, Номер unknown, С. 103094 - 103094

Опубликована: Март 1, 2025

Язык: Английский

---

Water Temperature Model to Assess Impact of Riparian Vegetation on Jucar River and Spain

Water, Год журнала: 2024, Номер 16(21), С. 3121 - 3121

Опубликована: Ноя. 1, 2024

Water temperature is a critical factor for aquatic ecosystems, influencing both chemical and biological processes, such as fish growth mortality; consequently, river lake ecosystems are sensitive to climate change (CC). Currently proposed CC scenarios indicate that air the Mediterranean Jucar River will increase higher in summer, 4.7 °C (SSP5-8.5), resulting water hotter month; July, 2.8 (SSP5-8.5). This have an impact on significantly reducing, fragmenting, or even eliminating natural cold-water species habitats, common trout. study consists of developing simulated model relates with shadow generated by riverside vegetation. The input data temperature, solar radiation, depth. only has one parameter, percentage. was calibrated representative stretch river, obtaining 0.93 Nash–Sutcliffe efficiency coefficient (NSE) indicates very good fit, 0.90 Kling–Gupta index (KGE), relative bias 0.04. also validated two other stretches same river. results show each 10% number shadows can reduce 1.2 and, applied, increasing from current status 62% 76–87% compensate CC. Generating shaded areas restorations be main measures rise due change.

Язык: Английский

---

Air Pollutant Concentration Forecasting with WTMP: Wavelet Transform-Based Multilayer Perceptron

Atmosphere, Год журнала: 2024, Номер 15(11), С. 1296 - 1296

Опубликована: Окт. 28, 2024

Atmospheric pollutants’ real-time changes and the internal interactions among various data make it challenging to efficiently predict concentration variations. In order extract more information from time series of pollutants improve accuracy prediction models, we propose a type Multilayer Perceptron model based on wavelet decomposition, named Wavelet Transform-based (WTMP) model. This decomposes pollutant through overlapping discrete transforms non-stationarity nonlinear dependencies in series. It combines decomposed with static covariate such as collection inputs them into an improved (MLP) model, reconstructing outputting results. Finally, is validated using atmospheric collected at specific location Ruian City, Zhejiang Province, China. The results indicate that performs well minimal errors.

Язык: Английский

---