Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods

Abstract Floods are widespread and dangerous natural hazards worldwide. It is essential to grasp the causes of floods to mitigate their severe effects on people and society. The key drivers of flood susceptibility in rapidly urbanizing areas can vary depending on the specific context and require fur...

Full description

Saved in:
Bibliographic Details
Main Authors: Kaili Zhu, Zhaoli Wang, Chengguang Lai, Shanshan Li, Zhaoyang Zeng, Xiaohong Chen
Format: Article
Language:English
Published: SpringerOpen 2024-10-01
Series:International Journal of Disaster Risk Science
Subjects:
Online Access:https://doi.org/10.1007/s13753-024-00590-6
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850162804421558272
author Kaili Zhu
Zhaoli Wang
Chengguang Lai
Shanshan Li
Zhaoyang Zeng
Xiaohong Chen
author_facet Kaili Zhu
Zhaoli Wang
Chengguang Lai
Shanshan Li
Zhaoyang Zeng
Xiaohong Chen
author_sort Kaili Zhu
collection DOAJ
description Abstract Floods are widespread and dangerous natural hazards worldwide. It is essential to grasp the causes of floods to mitigate their severe effects on people and society. The key drivers of flood susceptibility in rapidly urbanizing areas can vary depending on the specific context and require further investigation. This research developed an index system comprising 10 indicators associated with factors and environments that lead to disasters, and used machine learning methods to assess flood susceptibility. The core urban area of the Yangtze River Delta served as a case study. Four scenarios depicting separate and combined effects of climate change and human activity were evaluated using data from various periods, to measure the spatial variability in flood susceptibility. The findings demonstrate that the extreme gradient boosting model outperformed the decision tree, support vector machine, and stacked models in evaluating flood susceptibility. Both climate change and human activity were found to act as catalysts for flooding in the region. Areas with increasing susceptibility were mainly distributed to the northwest and southeast of Taihu Lake. Areas with increased flood susceptibility caused by climate change were significantly larger than those caused by human activity, indicating that climate change was the dominant factor influencing flood susceptibility in the region. By comparing the relationship between the indicators and flood susceptibility, the rising intensity and frequency of extreme precipitation as well as an increase in impervious surface areas were identified as important reasons of heightened flood susceptibility in the Yangtze River Delta region. This study emphasized the significance of formulating adaptive strategies to enhance flood control capabilities to cope with the changing environment.
format Article
id doaj-art-ea5c00e9bf614d9d8ab08b647ea8ad69
institution OA Journals
issn 2095-0055
2192-6395
language English
publishDate 2024-10-01
publisher SpringerOpen
record_format Article
series International Journal of Disaster Risk Science
spelling doaj-art-ea5c00e9bf614d9d8ab08b647ea8ad692025-08-20T02:22:28ZengSpringerOpenInternational Journal of Disaster Risk Science2095-00552192-63952024-10-0115573875310.1007/s13753-024-00590-6Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning MethodsKaili Zhu0Zhaoli Wang1Chengguang Lai2Shanshan Li3Zhaoyang Zeng4Xiaohong Chen5School of Civil Engineering and Transportation, State Key Laboratory of Subtropical Building and Urban Science, South China University of TechnologySchool of Civil Engineering and Transportation, State Key Laboratory of Subtropical Building and Urban Science, South China University of TechnologySchool of Civil Engineering and Transportation, State Key Laboratory of Subtropical Building and Urban Science, South China University of TechnologySchool of Civil Engineering and Transportation, State Key Laboratory of Subtropical Building and Urban Science, South China University of TechnologySchool of Civil Engineering and Transportation, State Key Laboratory of Subtropical Building and Urban Science, South China University of TechnologyCenter for Water Resources and Environment, Sun Yat-sen UniversityAbstract Floods are widespread and dangerous natural hazards worldwide. It is essential to grasp the causes of floods to mitigate their severe effects on people and society. The key drivers of flood susceptibility in rapidly urbanizing areas can vary depending on the specific context and require further investigation. This research developed an index system comprising 10 indicators associated with factors and environments that lead to disasters, and used machine learning methods to assess flood susceptibility. The core urban area of the Yangtze River Delta served as a case study. Four scenarios depicting separate and combined effects of climate change and human activity were evaluated using data from various periods, to measure the spatial variability in flood susceptibility. The findings demonstrate that the extreme gradient boosting model outperformed the decision tree, support vector machine, and stacked models in evaluating flood susceptibility. Both climate change and human activity were found to act as catalysts for flooding in the region. Areas with increasing susceptibility were mainly distributed to the northwest and southeast of Taihu Lake. Areas with increased flood susceptibility caused by climate change were significantly larger than those caused by human activity, indicating that climate change was the dominant factor influencing flood susceptibility in the region. By comparing the relationship between the indicators and flood susceptibility, the rising intensity and frequency of extreme precipitation as well as an increase in impervious surface areas were identified as important reasons of heightened flood susceptibility in the Yangtze River Delta region. This study emphasized the significance of formulating adaptive strategies to enhance flood control capabilities to cope with the changing environment.https://doi.org/10.1007/s13753-024-00590-6Climate changeFlood susceptibilityHuman activityMachine learning methodsYangtze River Delta core urban agglomeration
spellingShingle Kaili Zhu
Zhaoli Wang
Chengguang Lai
Shanshan Li
Zhaoyang Zeng
Xiaohong Chen
Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
International Journal of Disaster Risk Science
Climate change
Flood susceptibility
Human activity
Machine learning methods
Yangtze River Delta core urban agglomeration
title Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
title_full Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
title_fullStr Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
title_full_unstemmed Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
title_short Evaluating Factors Affecting Flood Susceptibility in the Yangtze River Delta Using Machine Learning Methods
title_sort evaluating factors affecting flood susceptibility in the yangtze river delta using machine learning methods
topic Climate change
Flood susceptibility
Human activity
Machine learning methods
Yangtze River Delta core urban agglomeration
url https://doi.org/10.1007/s13753-024-00590-6
work_keys_str_mv AT kailizhu evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods
AT zhaoliwang evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods
AT chengguanglai evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods
AT shanshanli evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods
AT zhaoyangzeng evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods
AT xiaohongchen evaluatingfactorsaffectingfloodsusceptibilityintheyangtzeriverdeltausingmachinelearningmethods