River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model

The supersaturation of total dissolved gas (TDG) in rivers serves as a critical indicator of water quality downstream of high dams. This study models TDG levels at two monitoring stations in the Columbia and Snake River Basins (USA), where high TDG concentrations were recorded. Hourly data on water...

Full description

Saved in:
Bibliographic Details
Main Authors: Khabat Khosravi, Salim Heddam, Changhyun Jun, Sayed M. Bateni, Dongkyun Kim, Essam Heggy
Format: Article
Language:English
Published: Elsevier 2025-12-01
Series:Ecological Informatics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1574954125002006
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849233617075568640
author Khabat Khosravi
Salim Heddam
Changhyun Jun
Sayed M. Bateni
Dongkyun Kim
Essam Heggy
author_facet Khabat Khosravi
Salim Heddam
Changhyun Jun
Sayed M. Bateni
Dongkyun Kim
Essam Heggy
author_sort Khabat Khosravi
collection DOAJ
description The supersaturation of total dissolved gas (TDG) in rivers serves as a critical indicator of water quality downstream of high dams. This study models TDG levels at two monitoring stations in the Columbia and Snake River Basins (USA), where high TDG concentrations were recorded. Hourly data on water temperature, barometric pressure, dam spill, sensor depth, and discharge serve as input variables for deep-learning models. Several models are developed and tested, including long short-term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent unit (GRU), and an alternating model tree (AMT) hybridized with iterative absolute error regression (IAER) and iterative classifier optimizer (ICO) algorithms. A greedy stepwise feature selection technique (GSFST) is employed to identify the optimal model inputs. Each model is trained and evaluated at one station and validated at the second station to assess transferability and generalization capability. Model performance was compared using multiple quantitative and qualitative metrics, including the Nash–Sutcliffe Efficiency and uncertainty coefficient. Additionally, Friedman and Wilcoxon signed-rank tests confirmed statistically significant differences between models. Dam spills emerged as the most influential predictor of TDG levels at both sites. The GSFST selected the optimal input combination, including dam spill, water temperature, barometric pressure, and sensor depth. Among all models, GSFST-BiLSTM achieved the highest predictive accuracy, with Nash–Sutcliffe values of 0.95 (testing) and 0.90 (validation) and uncertainty coefficients of 5.2 % and 7.0 %, respectively. These findings demonstrate that GSFST-BiLSTM provides a robust and transferable framework for TDG prediction, with the potential for broader application pending further validation.
format Article
id doaj-art-d628f832ab6544e7bd7a3711d5aca254
institution Kabale University
issn 1574-9541
language English
publishDate 2025-12-01
publisher Elsevier
record_format Article
series Ecological Informatics
spelling doaj-art-d628f832ab6544e7bd7a3711d5aca2542025-08-20T05:05:00ZengElsevierEcological Informatics1574-95412025-12-019010319110.1016/j.ecoinf.2025.103191River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory modelKhabat Khosravi0Salim Heddam1Changhyun Jun2Sayed M. Bateni3Dongkyun Kim4Essam Heggy5Department of Natural Resources, College of Agriculture and Natural Resources, Razi University, Kermanshah, Iran; Corresponding authors.Faculty of Science, Agronomy Department, University 20 Août 1955 Skikda, Route El Hadaik, BP 26 Skikda, AlgeriaSchool of Civil, Environmental and Architectural Engineering, College of Engineering, Korea University, Seoul, Republic of Korea; Corresponding authors.Department of Civil, Environmental and Construction Engineering, and Water Resources Research Center, University of Hawaii at Manoa, Honolulu, HI, USA; UNESCO-UNISA Africa Chair in Nanoscience and Nanotechnology College of Graduates Studies, University of South Africa, Muckleneuk Ridge, Pretoria 392, South AfricaDepartment of Civil and Environmental Engineering, Hongik University, Seoul, South KoreaViterbi School of Engineering, University of Southern California, Los Angeles, CA 90089, USAThe supersaturation of total dissolved gas (TDG) in rivers serves as a critical indicator of water quality downstream of high dams. This study models TDG levels at two monitoring stations in the Columbia and Snake River Basins (USA), where high TDG concentrations were recorded. Hourly data on water temperature, barometric pressure, dam spill, sensor depth, and discharge serve as input variables for deep-learning models. Several models are developed and tested, including long short-term memory (LSTM), bidirectional LSTM (BiLSTM), gated recurrent unit (GRU), and an alternating model tree (AMT) hybridized with iterative absolute error regression (IAER) and iterative classifier optimizer (ICO) algorithms. A greedy stepwise feature selection technique (GSFST) is employed to identify the optimal model inputs. Each model is trained and evaluated at one station and validated at the second station to assess transferability and generalization capability. Model performance was compared using multiple quantitative and qualitative metrics, including the Nash–Sutcliffe Efficiency and uncertainty coefficient. Additionally, Friedman and Wilcoxon signed-rank tests confirmed statistically significant differences between models. Dam spills emerged as the most influential predictor of TDG levels at both sites. The GSFST selected the optimal input combination, including dam spill, water temperature, barometric pressure, and sensor depth. Among all models, GSFST-BiLSTM achieved the highest predictive accuracy, with Nash–Sutcliffe values of 0.95 (testing) and 0.90 (validation) and uncertainty coefficients of 5.2 % and 7.0 %, respectively. These findings demonstrate that GSFST-BiLSTM provides a robust and transferable framework for TDG prediction, with the potential for broader application pending further validation.http://www.sciencedirect.com/science/article/pii/S1574954125002006Total dissolved gas (TDG)PredictionDeep learningBidirectional long short-term memory (BiLSTM)Greedy stepwise feature selectionFriedman test
spellingShingle Khabat Khosravi
Salim Heddam
Changhyun Jun
Sayed M. Bateni
Dongkyun Kim
Essam Heggy
River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
Ecological Informatics
Total dissolved gas (TDG)
Prediction
Deep learning
Bidirectional long short-term memory (BiLSTM)
Greedy stepwise feature selection
Friedman test
title River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
title_full River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
title_fullStr River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
title_full_unstemmed River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
title_short River total dissolved gas prediction using a hybrid greedy-stepwise feature selection and bidirectional long short-term memory model
title_sort river total dissolved gas prediction using a hybrid greedy stepwise feature selection and bidirectional long short term memory model
topic Total dissolved gas (TDG)
Prediction
Deep learning
Bidirectional long short-term memory (BiLSTM)
Greedy stepwise feature selection
Friedman test
url http://www.sciencedirect.com/science/article/pii/S1574954125002006
work_keys_str_mv AT khabatkhosravi rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel
AT salimheddam rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel
AT changhyunjun rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel
AT sayedmbateni rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel
AT dongkyunkim rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel
AT essamheggy rivertotaldissolvedgaspredictionusingahybridgreedystepwisefeatureselectionandbidirectionallongshorttermmemorymodel