Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction

Study region:: Hanjiang River Basin, China Study focus:: This study introduces DLGNet, an advanced water level prediction model that employs an adaptive multi-resolution decomposition to decouple multiple periodic components. The model incorporates a 2D overlap patch attention mechanism and separabl...

Full description

Saved in:
Bibliographic Details
Main Authors: Hui Wang, LiPing Wang, QiCang Qiu, Yuyan Gao, WenJuan Li, QiuShun Wang
Format: Article
Language:English
Published: Elsevier 2025-02-01
Series:Journal of Hydrology: Regional Studies
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2214581824004695
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832591746338390016
author Hui Wang
LiPing Wang
QiCang Qiu
Yuyan Gao
WenJuan Li
QiuShun Wang
author_facet Hui Wang
LiPing Wang
QiCang Qiu
Yuyan Gao
WenJuan Li
QiuShun Wang
author_sort Hui Wang
collection DOAJ
description Study region:: Hanjiang River Basin, China Study focus:: This study introduces DLGNet, an advanced water level prediction model that employs an adaptive multi-resolution decomposition to decouple multiple periodic components. The model incorporates a 2D overlap patch attention mechanism and separable dilation causal convolution to effectively capture local and global temporal variations. Additionally, DLGNet utilizes multiscale 2D separable convolutions to capture temporal variations among period components and employs a novel aggregation strategy for feature interaction analysis, ultimately enhancing prediction accuracy through direct decoding. New hydrologic insights for the region:: Compared to ten other advanced prediction models, DLGNet achieved an average reduction in mean square error (MSE) by 94. 82%, 93. 35%, and 83. 13% for short-term, mid-term, and long-term predictions, respectively. The adaptive period decomposition capability and multiscale temporal variation representation learning of DLGNet ensure its robustness and accuracy across various prediction tasks. Moreover, in all prediction tasks, the output of DLGNet more closely approximates the ground truth, exhibiting rich local detail changes and aligning more accurately with the actual fluctuations of water levels over time.
format Article
id doaj-art-0835eff9ea0f42b6891819728a7b2322
institution Kabale University
issn 2214-5818
language English
publishDate 2025-02-01
publisher Elsevier
record_format Article
series Journal of Hydrology: Regional Studies
spelling doaj-art-0835eff9ea0f42b6891819728a7b23222025-01-22T05:42:06ZengElsevierJournal of Hydrology: Regional Studies2214-58182025-02-0157102120Adaptive period decomposition network with muti-scale temporal variation learning for water level predictionHui Wang0LiPing Wang1QiCang Qiu2Yuyan Gao3WenJuan Li4QiuShun Wang5Computer Science & Technology, Zhejiang University of Technology, HangZhou, ChinaComputer Science & Technology, Zhejiang University of Technology, HangZhou, China; Corresponding author.Zhejiang Lab, No. 1818, Western Road of Wenyi, Hangzhou, Zhejiang, ChinaComputer Science & Technology, Zhejiang University of Technology, HangZhou, ChinaMeteorological Bureau of Zhejiang Province, HangZhou, ChinaZhejiang Institute of Hydraulics and Estuary, HangZhou, ChinaStudy region:: Hanjiang River Basin, China Study focus:: This study introduces DLGNet, an advanced water level prediction model that employs an adaptive multi-resolution decomposition to decouple multiple periodic components. The model incorporates a 2D overlap patch attention mechanism and separable dilation causal convolution to effectively capture local and global temporal variations. Additionally, DLGNet utilizes multiscale 2D separable convolutions to capture temporal variations among period components and employs a novel aggregation strategy for feature interaction analysis, ultimately enhancing prediction accuracy through direct decoding. New hydrologic insights for the region:: Compared to ten other advanced prediction models, DLGNet achieved an average reduction in mean square error (MSE) by 94. 82%, 93. 35%, and 83. 13% for short-term, mid-term, and long-term predictions, respectively. The adaptive period decomposition capability and multiscale temporal variation representation learning of DLGNet ensure its robustness and accuracy across various prediction tasks. Moreover, in all prediction tasks, the output of DLGNet more closely approximates the ground truth, exhibiting rich local detail changes and aligning more accurately with the actual fluctuations of water levels over time.http://www.sciencedirect.com/science/article/pii/S2214581824004695Time seriesWater level predictionDeep learningPattern recognition
spellingShingle Hui Wang
LiPing Wang
QiCang Qiu
Yuyan Gao
WenJuan Li
QiuShun Wang
Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
Journal of Hydrology: Regional Studies
Time series
Water level prediction
Deep learning
Pattern recognition
title Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
title_full Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
title_fullStr Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
title_full_unstemmed Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
title_short Adaptive period decomposition network with muti-scale temporal variation learning for water level prediction
title_sort adaptive period decomposition network with muti scale temporal variation learning for water level prediction
topic Time series
Water level prediction
Deep learning
Pattern recognition
url http://www.sciencedirect.com/science/article/pii/S2214581824004695
work_keys_str_mv AT huiwang adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction
AT lipingwang adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction
AT qicangqiu adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction
AT yuyangao adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction
AT wenjuanli adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction
AT qiushunwang adaptiveperioddecompositionnetworkwithmutiscaletemporalvariationlearningforwaterlevelprediction