Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge

Flooding is one of the most devastating natural disasters worldwide, with increasing frequency due to climate change. Traditional hydrological models require extensive data and computational resources, while machine learning (ML) models struggle to capture spatial dependencies. To address this, we p...

Full description

Saved in:
Bibliographic Details
Main Authors: Jialou Wang, Jacob Sanderson, Sadaf Iqbal, Wai Lok Woo
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/9/1540
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850277585621090304
author Jialou Wang
Jacob Sanderson
Sadaf Iqbal
Wai Lok Woo
author_facet Jialou Wang
Jacob Sanderson
Sadaf Iqbal
Wai Lok Woo
author_sort Jialou Wang
collection DOAJ
description Flooding is one of the most devastating natural disasters worldwide, with increasing frequency due to climate change. Traditional hydrological models require extensive data and computational resources, while machine learning (ML) models struggle to capture spatial dependencies. To address this, we propose a modified U-Net architecture that integrates prior hydrological knowledge of permanent water bodies to improve flood susceptibility mapping in Northumberland County, UK. By embedding domain-specific insights, our model achieves a higher area under the curve (AUC) (0.97) compared to the standard U-Net (0.93), while also reducing training time by converging three times faster. Additionally, we integrate a Grad-CAM module to provide visualisations explaining the areas of attention from the model, enabling interpretation of its decision-making, thus reducing barriers to its practical implementation.
format Article
id doaj-art-dee32ea033b44f339461d88d7b12c2d6
institution OA Journals
issn 2072-4292
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-dee32ea033b44f339461d88d7b12c2d62025-08-20T01:49:50ZengMDPI AGRemote Sensing2072-42922025-04-01179154010.3390/rs17091540Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior KnowledgeJialou Wang0Jacob Sanderson1Sadaf Iqbal2Wai Lok Woo3Department of Computer and Information Sciences, Northumbria University, Newcastle NE1 8ST, UKDepartment of Computer and Information Sciences, Northumbria University, Newcastle NE1 8ST, UKDepartment of Computer and Information Sciences, Northumbria University, Newcastle NE1 8ST, UKDepartment of Computer and Information Sciences, Northumbria University, Newcastle NE1 8ST, UKFlooding is one of the most devastating natural disasters worldwide, with increasing frequency due to climate change. Traditional hydrological models require extensive data and computational resources, while machine learning (ML) models struggle to capture spatial dependencies. To address this, we propose a modified U-Net architecture that integrates prior hydrological knowledge of permanent water bodies to improve flood susceptibility mapping in Northumberland County, UK. By embedding domain-specific insights, our model achieves a higher area under the curve (AUC) (0.97) compared to the standard U-Net (0.93), while also reducing training time by converging three times faster. Additionally, we integrate a Grad-CAM module to provide visualisations explaining the areas of attention from the model, enabling interpretation of its decision-making, thus reducing barriers to its practical implementation.https://www.mdpi.com/2072-4292/17/9/1540flood susceptibility mappingdeep learningU-Nethydrology-aware deep learningremote sensingdigital terrain model (DTM)
spellingShingle Jialou Wang
Jacob Sanderson
Sadaf Iqbal
Wai Lok Woo
Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
Remote Sensing
flood susceptibility mapping
deep learning
U-Net
hydrology-aware deep learning
remote sensing
digital terrain model (DTM)
title Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
title_full Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
title_fullStr Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
title_full_unstemmed Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
title_short Accelerated and Interpretable Flood Susceptibility Mapping Through Explainable Deep Learning with Hydrological Prior Knowledge
title_sort accelerated and interpretable flood susceptibility mapping through explainable deep learning with hydrological prior knowledge
topic flood susceptibility mapping
deep learning
U-Net
hydrology-aware deep learning
remote sensing
digital terrain model (DTM)
url https://www.mdpi.com/2072-4292/17/9/1540
work_keys_str_mv AT jialouwang acceleratedandinterpretablefloodsusceptibilitymappingthroughexplainabledeeplearningwithhydrologicalpriorknowledge
AT jacobsanderson acceleratedandinterpretablefloodsusceptibilitymappingthroughexplainabledeeplearningwithhydrologicalpriorknowledge
AT sadafiqbal acceleratedandinterpretablefloodsusceptibilitymappingthroughexplainabledeeplearningwithhydrologicalpriorknowledge
AT wailokwoo acceleratedandinterpretablefloodsusceptibilitymappingthroughexplainabledeeplearningwithhydrologicalpriorknowledge