Harnessing multisatellite remote sensing data and machine learning for flood risk assessment in Nam Ngum River Basin, Lao PDR

Sackdavong Mangkhaseum; Sunil Duwal; Yogesh Bhattarai; Akitoshi Hanazawa

doi:10.1117/12.3022819

28 May 2024 Harnessing multisatellite remote sensing data and machine learning for flood risk assessment in Nam Ngum River Basin, Lao PDR

Sackdavong Mangkhaseum, Sunil Duwal, Yogesh Bhattarai, Akitoshi Hanazawa

Author Affiliations +

Proceedings Volume 13083, SPIE Future Sensing Technologies 2024; 130830B (2024) https://doi.org/10.1117/12.3022819
Event: SPIE Future Sensing Technologies, 2024, Yokohama, Japan

Abstract

The study focuses on flood susceptibility in the Nam Ngum River Basin, Lao PDR, an area prone to annual flooding due to monsoons and rainstorms. Flooding in this region significantly threatens human life, causes economic losses, and damages communities and agriculture. The study employs advanced remote sensing and machine learning techniques, including Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Networks (ANN), to address these issues and create detailed flood susceptibility maps. The machine learning models used historical flood data, Sentinel-1 SAR imagery from 2018 to 2020, and open-source flood data for training and validation. Eleven flood factors were considered. With 776 samples, 70% were trained, and 30% tested the model. Flood susceptibility map accuracy is assessed using statistical techniques such as multicollinearity, Kappa index, and Area Under the curve of Receiver Operating Characteristics (AUROC). The generated flood susceptibility map is used to analyze the possible effect on the different land use/land cover classes and populations. RF outperforms SVM and ANN, achieving higher accuracy based on Receiver Operating Characteristics. The resulting flood susceptibility map reveals that 25-44% of the basin area is highly susceptible, predominantly in low-elevation and low-slope regions. Likewise, 85 to 90% of the people are highly vulnerable to flooding within 260 to 280 km² of built-up area. The study proposes a new approach to using machine learning and readily available remote sensing data for flood susceptibility mapping. The findings of this study provide essential insights for policymakers, aiding in disaster risk reduction and facilitating sustainable development planning in Lao PDR.

Conference Presentation

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Sackdavong Mangkhaseum, Sunil Duwal, Yogesh Bhattarai, and Akitoshi Hanazawa "Harnessing multisatellite remote sensing data and machine learning for flood risk assessment in Nam Ngum River Basin, Lao PDR", Proc. SPIE 13083, SPIE Future Sensing Technologies 2024, 130830B (28 May 2024); https://doi.org/10.1117/12.3022819

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available