Proceedings of the 2nd International Conference on Nature-Based Solution in Climate Change, RESILIENCE 2023, 24 November 2023, Jakarta, Indonesia

Research Article

Assessing Landslide Vulnerability in Deforested Areas of Sumatra Island Using Remote Sensing and Machine Learning

Download30 downloads
  • @INPROCEEDINGS{10.4108/eai.24-11-2023.2346387,
        author={Ifandra Kusuma Cahya Reynaldi and Anjar Dimara Sakti and Deni  Suwardhi and Agung Budi Harto},
        title={Assessing Landslide Vulnerability in Deforested Areas of Sumatra Island Using Remote Sensing and Machine Learning},
        proceedings={Proceedings of the 2nd International Conference on Nature-Based Solution in Climate Change, RESILIENCE 2023, 24 November 2023, Jakarta, Indonesia},
        publisher={EAI},
        proceedings_a={RESILIENCE},
        year={2024},
        month={7},
        keywords={landslide deforestation remote sensing machine learning hazard},
        doi={10.4108/eai.24-11-2023.2346387}
    }
    
  • Ifandra Kusuma Cahya Reynaldi
    Anjar Dimara Sakti
    Deni Suwardhi
    Agung Budi Harto
    Year: 2024
    Assessing Landslide Vulnerability in Deforested Areas of Sumatra Island Using Remote Sensing and Machine Learning
    RESILIENCE
    EAI
    DOI: 10.4108/eai.24-11-2023.2346387
Ifandra Kusuma Cahya Reynaldi1, Anjar Dimara Sakti1,*, Deni Suwardhi1, Agung Budi Harto1
  • 1: Institut Teknologi Bandung
*Contact email: anjar@itb.ac.id

Abstract

Sumatra is among 11 regions in the world that significantly contribute to global deforestation. In the period 2001-2020, deforestation in Sumatra has reached more than three million hectares. This has made hydrometeorological disasters a reality in Sumatra due to the loss of rainwater catchment. One of them is landslides that often occur and have endangered residents and productive areas in Sumatra. The development of ecosystem-based mitigation can be done to reduce risks and losses that may occur. However, it is necessary to conduct preliminary studies to assess the likelihood of future hazards. This can be done by utilizing remote sensing technology and geographic information systems using various geospatial data. Supported by a machine learning approach, it can improve the quality of hazard assessment. It was found that Sumatera has landslide hazard vulnerability dominated from medium to high level in mountain area, but not in deforested area.