Detection of post-flood persistent Water Bodies and their Impact

Climate change and global warming have led to an increase of extreme weather events like extreme and unprecedented rainfall in recent years. Flooding near urban areas caused by intense rainfall have impacted the United Arab Emirates (UAE), damaging infrastructure and posing significant threats to communities. In light of the recent heavy rainfall events in the UAE, this paper presents a preliminary risk assessment of vulnerable areas in Dubai using satellite imagery and machine learning to identify persistent flooded areas that require future infrastructure planning. The analysis focuses on the biggest rain event that occurred in Dubai in the past 75 years, in April of 2024, assessing areas at high risk of flooding using a Support Vector Machine (SVM) classifier on Sentinel 2 images which resulted in above 95% overall accuracy. To collect training data, a Modified Normalized Difference Water Index (MNDWI) was produced for an image one day after the event (co-flood), a threshold of 0.25 was applied to filter out non-water features. A cleaned version of the MNDWI features were used to train the model. Using random point and ground truth from the original image, an accuracy assessment was made to verify the accuracy of the model. Employing the trained model on three other images, two pre-flood, and one 10 days after event (post-flood) to obtain their classifications. Permanent water bodies identified from the common feature in the pre-flood classification were excluded from the co- and post-flood classifications, and water bodies present in both were identified as persistent water bodies. Due to the heavy rainfall event, the flooded areas accumulated a total of 112 km<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>. The extensive efforts from public authorities to remove water from critical and urban areas reduced the flooded areas to 44 km<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> with 40% of those areas within 200 m of urban areas. The prolonged presence of these persistent water bodies without proper drainage on time will create a breeding ground for disease-carrying insects which will lead to health and environmental consequences.

Authors

Institutions

Publication Details

Published
2024-11-27
DOI
https://doi.org/10.1109/ihtc61819.2024.10855103
Primary Topic
Flood Risk Assessment and Management
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Detection of post-flood persistent Water Bodies and their Impact

Hussain Al-Ahmad, Saeed Al Mansoori, M. Sami Zitouni, Naseeb Asaad Albakri et al.
Flood Risk Assessment and Management
article

Detection of post-flood persistent Water Bodies and their Impact

Hussain Al-Ahmad, Saeed Al Mansoori, M. Sami Zitouni, Naseeb Asaad Albakri, Leena Elneel
article en

Abstract

Climate change and global warming have led to an increase of extreme weather events like extreme and unprecedented rainfall in recent years. Flooding near urban areas caused by intense rainfall have impacted the United Arab Emirates (UAE), damaging infrastructure and posing significant threats to communities. In light of the recent heavy rainfall events in the UAE, this paper presents a preliminary risk assessment of vulnerable areas in Dubai using satellite imagery and machine learning to identify persistent flooded areas that require future infrastructure planning. The analysis focuses on the biggest rain event that occurred in Dubai in the past 75 years, in April of 2024, assessing areas at high risk of flooding using a Support Vector Machine (SVM) classifier on Sentinel 2 images which resulted in above 95% overall accuracy. To collect training data, a Modified Normalized Difference Water Index (MNDWI) was produced for an image one day after the event (co-flood), a threshold of 0.25 was applied to filter out non-water features. A cleaned version of the MNDWI features were used to train the model. Using random point and ground truth from the original image, an accuracy assessment was made to verify the accuracy of the model. Employing the trained model on three other images, two pre-flood, and one 10 days after event (post-flood) to obtain their classifications. Permanent water bodies identified from the common feature in the pre-flood classification were excluded from the co- and post-flood classifications, and water bodies present in both were identified as persistent water bodies. Due to the heavy rainfall event, the flooded areas accumulated a total of 112 km<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>. The extensive efforts from public authorities to remove water from critical and urban areas reduced the flooded areas to 44 km<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> with 40% of those areas within 200 m of urban areas. The prolonged presence of these persistent water bodies without proper drainage on time will create a breeding ground for disease-carrying insects which will lead to health and environmental consequences.

University of Dubai (AE), Mohammed Bin Rashid School of Government (AE)
Openalex Percentile: Top 77%
Flood Risk Assessment and Management
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.