5:15 PM - 7:15 PM
[U03-P06] Geospatial Analysis of Flood Hazards Using Remote Sensing Techniques in the Lower Mekong Region
Keywords:MODIS data, Flood hazards, Remote sensing, Lower Mekong Region
Flooding ranks among the most frequent and costly natural hazards worldwide, inflicting severe human, agricultural, and infrastructural losses that exacerbate poverty. Cambodia, located in the Lower Mekong Region and one of the globe's most climate-vulnerable nations, was ranked 13th among 181 countries affected by extreme weather events between 1996 and 2015. This study presents a novel approach to accurately delineate flood-hazard areas in Cambodia using the multi-temporal Moderate Resolution Imaging Spectroradiometer (MODIS) data. Initially, by integrating three key vegetation indices, including Enhanced Vegetation Index (EVI), Land Surface Water Index (LSWI), and the differential index between LSWI and EVI (DVLE), we developed a two-step methodology to first map flood-prone areas. The water-related pixels were identified by thresholds (EVI < 0.3 and DVLE < 0.05). Pixels with EVI values below 0.1 were subsequently classified as flood-prone areas. Mixed land-surface areas were distinguished when EVI values ranged between 0.1 and 0.3. The results validated against the ground reference data demonstrated an overall classification accuracy exceeding 88.7%. Finally, flood probability analyses for 2000–2017 were performed, and the results revealed a high likelihood of flood events in districts adjacent to the Tonle Sap Lake and the Mekong River, with at least 32 districts experiencing floods on more than 40% of their land area when flood probability exceeded 70%. These research findings provide robust, spatially explicit data that can inform effective flood management strategies, optimize rescue operations, and guide policymaking to mitigate the adverse impacts of flooding on Cambodia’s communities and agricultural sectors.