A river network model using a weight-based merged LSTM for multi-source monitoring integration

Rivers typically exhibit spatial connectivity from upstream to downstream, with various heterogeneous monitoring systems operating concurrently across basins. While graph neural networks (GNNs) have shown promise in modeling spatial connectivity, they remain limited by reliance on features common to...

Full description

Saved in:
Bibliographic Details
Main Authors: Jonggyu Jung, Taeseung Park, Jaegwan Park, Dogeon Lee, YoonKyung Cha
Format: Article
Language:English
Published: Elsevier 2025-12-01
Series:Ecological Informatics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1574954125003292
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849233572542545920
author Jonggyu Jung
Taeseung Park
Jaegwan Park
Dogeon Lee
YoonKyung Cha
author_facet Jonggyu Jung
Taeseung Park
Jaegwan Park
Dogeon Lee
YoonKyung Cha
author_sort Jonggyu Jung
collection DOAJ
description Rivers typically exhibit spatial connectivity from upstream to downstream, with various heterogeneous monitoring systems operating concurrently across basins. While graph neural networks (GNNs) have shown promise in modeling spatial connectivity, they remain limited by reliance on features common to all nodes. As the diversity of monitoring systems increases, the overlap of measured variables decreases, reducing usable input features and limiting model applicability. To address this issue, this study proposes a river network model based on a weight-based merged long short-term memory (LSTM) architecture for forecasting daily total organic carbon (TOC) concentrations by integrating multi-source data from spatially connected monitoring sites. The river network is divided into upstream, midstream, and downstream segments, with each input processed independently and merged into a unified representation. Segment models are connected in sequence to represent directional flow. Trainable scalar weights are included to quantify the relative contribution of each site and enhance spatial interpretability. The model is applied to a section of the Han River in South Korea, which flows through the Seoul metropolitan area in South Korea. The model demonstrates strong forecasting performance, with mean absolute errors ranging from 0.055 to 0.518, root mean squared errors from 0.075 to 0.784, and coefficients of determination between 0.424 and 0.721. Scenario analyses using site-specific contributions are conducted to evaluate changes in TOC concentrations under pollution reduction scenarios. This river network modeling framework is adaptable to a wide range of applications and provides practical utility for watershed-scale water quality forecasting and management.
format Article
id doaj-art-c6e9d2f540af417899ccbb6707e1ecbb
institution Kabale University
issn 1574-9541
language English
publishDate 2025-12-01
publisher Elsevier
record_format Article
series Ecological Informatics
spelling doaj-art-c6e9d2f540af417899ccbb6707e1ecbb2025-08-20T05:05:45ZengElsevierEcological Informatics1574-95412025-12-019010332010.1016/j.ecoinf.2025.103320A river network model using a weight-based merged LSTM for multi-source monitoring integrationJonggyu Jung0Taeseung Park1Jaegwan Park2Dogeon Lee3YoonKyung Cha4School of Environmental Engineering, University of Seoul, Dongdaemun-gu, Seoul 02504, Republic of KoreaSchool of Environmental Engineering, University of Seoul, Dongdaemun-gu, Seoul 02504, Republic of KoreaSchool of Environmental Engineering, University of Seoul, Dongdaemun-gu, Seoul 02504, Republic of KoreaSchool of Environmental Engineering, University of Seoul, Dongdaemun-gu, Seoul 02504, Republic of KoreaCorresponding author.; School of Environmental Engineering, University of Seoul, Dongdaemun-gu, Seoul 02504, Republic of KoreaRivers typically exhibit spatial connectivity from upstream to downstream, with various heterogeneous monitoring systems operating concurrently across basins. While graph neural networks (GNNs) have shown promise in modeling spatial connectivity, they remain limited by reliance on features common to all nodes. As the diversity of monitoring systems increases, the overlap of measured variables decreases, reducing usable input features and limiting model applicability. To address this issue, this study proposes a river network model based on a weight-based merged long short-term memory (LSTM) architecture for forecasting daily total organic carbon (TOC) concentrations by integrating multi-source data from spatially connected monitoring sites. The river network is divided into upstream, midstream, and downstream segments, with each input processed independently and merged into a unified representation. Segment models are connected in sequence to represent directional flow. Trainable scalar weights are included to quantify the relative contribution of each site and enhance spatial interpretability. The model is applied to a section of the Han River in South Korea, which flows through the Seoul metropolitan area in South Korea. The model demonstrates strong forecasting performance, with mean absolute errors ranging from 0.055 to 0.518, root mean squared errors from 0.075 to 0.784, and coefficients of determination between 0.424 and 0.721. Scenario analyses using site-specific contributions are conducted to evaluate changes in TOC concentrations under pollution reduction scenarios. This river network modeling framework is adaptable to a wide range of applications and provides practical utility for watershed-scale water quality forecasting and management.http://www.sciencedirect.com/science/article/pii/S1574954125003292River networkWeight-based merged LSTMSpatial connectivityFeature-flexibilityWater quality forecastingTotal organic carbon
spellingShingle Jonggyu Jung
Taeseung Park
Jaegwan Park
Dogeon Lee
YoonKyung Cha
A river network model using a weight-based merged LSTM for multi-source monitoring integration
Ecological Informatics
River network
Weight-based merged LSTM
Spatial connectivity
Feature-flexibility
Water quality forecasting
Total organic carbon
title A river network model using a weight-based merged LSTM for multi-source monitoring integration
title_full A river network model using a weight-based merged LSTM for multi-source monitoring integration
title_fullStr A river network model using a weight-based merged LSTM for multi-source monitoring integration
title_full_unstemmed A river network model using a weight-based merged LSTM for multi-source monitoring integration
title_short A river network model using a weight-based merged LSTM for multi-source monitoring integration
title_sort river network model using a weight based merged lstm for multi source monitoring integration
topic River network
Weight-based merged LSTM
Spatial connectivity
Feature-flexibility
Water quality forecasting
Total organic carbon
url http://www.sciencedirect.com/science/article/pii/S1574954125003292
work_keys_str_mv AT jonggyujung arivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT taeseungpark arivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT jaegwanpark arivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT dogeonlee arivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT yoonkyungcha arivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT jonggyujung rivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT taeseungpark rivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT jaegwanpark rivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT dogeonlee rivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration
AT yoonkyungcha rivernetworkmodelusingaweightbasedmergedlstmformultisourcemonitoringintegration