Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed

ABSTRACT Floods are major natural disasters that present considerable challenges to socioeconomic and ecological systems. Flash floods are highly nonlinear and exhibit rapid spatiotemporal variability. Existing methods struggle to capture these features, leading to suboptimal long‐term and peak floo...

Full description

Saved in:
Bibliographic Details
Main Authors: Rukai Wang, Ximin Yuan, Fuchang Tian, Minghui Liu, Xiujie Wang, Xiaobin Li, Minrui Wu
Format: Article
Language:English
Published: Wiley 2025-06-01
Series:Journal of Flood Risk Management
Subjects:
Online Access:https://doi.org/10.1111/jfr3.70022
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849433019044069376
author Rukai Wang
Ximin Yuan
Fuchang Tian
Minghui Liu
Xiujie Wang
Xiaobin Li
Minrui Wu
author_facet Rukai Wang
Ximin Yuan
Fuchang Tian
Minghui Liu
Xiujie Wang
Xiaobin Li
Minrui Wu
author_sort Rukai Wang
collection DOAJ
description ABSTRACT Floods are major natural disasters that present considerable challenges to socioeconomic and ecological systems. Flash floods are highly nonlinear and exhibit rapid spatiotemporal variability. Existing methods struggle to capture these features, leading to suboptimal long‐term and peak flood prediction accuracy. This study proposes a hierarchical flood prediction model based on clustering to enhance forecasting accuracy in the Heshengxi watershed. We employ STGCN and GWN models with the spatiotemporal attention mechanism. Enhanced loss functions further refine flood prediction accuracy. Results show that the hierarchical prediction method is an effective means of extracting flood features by addressing the variability of prediction parameters for different flood magnitudes. The integration of Graph Convolutional and Time Aware models enables the model to recognize the spatiotemporal flood characteristics, overcoming limitations of prevailing methods and ensuring long‐term forecast accuracy. The optimized loss function further improves the prediction performance, resulting in a significant improvement in the accuracy of flood peak prediction, with a reduction of 0.26% in the relative error of the peak prediction by the GWN model. This framework provides an effective solution for flood warning, emergency response, and optimal scheduling. It also demonstrates the potential of deep learning models in the field of intelligent hydrological forecasting.
format Article
id doaj-art-72d1df7fc67e486dafb4b49df4d312ce
institution Kabale University
issn 1753-318X
language English
publishDate 2025-06-01
publisher Wiley
record_format Article
series Journal of Flood Risk Management
spelling doaj-art-72d1df7fc67e486dafb4b49df4d312ce2025-08-20T03:27:11ZengWileyJournal of Flood Risk Management1753-318X2025-06-01182n/an/a10.1111/jfr3.70022Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain WatershedRukai Wang0Ximin Yuan1Fuchang Tian2Minghui Liu3Xiujie Wang4Xiaobin Li5Minrui Wu6State Key Laboratory of Hydraulic Engineering Intelligent Construction and Operation Tianjin University Tianjin ChinaState Key Laboratory of Hydraulic Engineering Intelligent Construction and Operation Tianjin University Tianjin ChinaState Key Laboratory of Hydraulic Engineering Intelligent Construction and Operation Tianjin University Tianjin ChinaResearch Institute of Soil and Water Conservation Chengde ChinaState Key Laboratory of Hydraulic Engineering Intelligent Construction and Operation Tianjin University Tianjin ChinaWater Conservancy Workstation Chengde ChinaState Key Laboratory of Hydraulic Engineering Intelligent Construction and Operation Tianjin University Tianjin ChinaABSTRACT Floods are major natural disasters that present considerable challenges to socioeconomic and ecological systems. Flash floods are highly nonlinear and exhibit rapid spatiotemporal variability. Existing methods struggle to capture these features, leading to suboptimal long‐term and peak flood prediction accuracy. This study proposes a hierarchical flood prediction model based on clustering to enhance forecasting accuracy in the Heshengxi watershed. We employ STGCN and GWN models with the spatiotemporal attention mechanism. Enhanced loss functions further refine flood prediction accuracy. Results show that the hierarchical prediction method is an effective means of extracting flood features by addressing the variability of prediction parameters for different flood magnitudes. The integration of Graph Convolutional and Time Aware models enables the model to recognize the spatiotemporal flood characteristics, overcoming limitations of prevailing methods and ensuring long‐term forecast accuracy. The optimized loss function further improves the prediction performance, resulting in a significant improvement in the accuracy of flood peak prediction, with a reduction of 0.26% in the relative error of the peak prediction by the GWN model. This framework provides an effective solution for flood warning, emergency response, and optimal scheduling. It also demonstrates the potential of deep learning models in the field of intelligent hydrological forecasting.https://doi.org/10.1111/jfr3.70022deep learningHesheng Riverspatial‐temporal featureswater level forecasting
spellingShingle Rukai Wang
Ximin Yuan
Fuchang Tian
Minghui Liu
Xiujie Wang
Xiaobin Li
Minrui Wu
Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
Journal of Flood Risk Management
deep learning
Hesheng River
spatial‐temporal features
water level forecasting
title Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
title_full Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
title_fullStr Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
title_full_unstemmed Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
title_short Flood Classification and Improved Loss Function by Combining Deep Learning Models to Improve Water Level Prediction in a Small Mountain Watershed
title_sort flood classification and improved loss function by combining deep learning models to improve water level prediction in a small mountain watershed
topic deep learning
Hesheng River
spatial‐temporal features
water level forecasting
url https://doi.org/10.1111/jfr3.70022
work_keys_str_mv AT rukaiwang floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT ximinyuan floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT fuchangtian floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT minghuiliu floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT xiujiewang floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT xiaobinli floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed
AT minruiwu floodclassificationandimprovedlossfunctionbycombiningdeeplearningmodelstoimprovewaterlevelpredictioninasmallmountainwatershed