Flood assessment using multi-temporal remotely sensed data in Cambodia

This study aims to develop an approach for flood assessment using MODIS data in Cambodia. The data were processed for the period from 2000 to 2017 using vegetation and water surface indices. The results of flood mapping were verified with the ground reference data, indicating the overall accuracy of 88.7% and Kappa coefficient of 0.77, respectively. These findings were confirmed by close agreement between the flood-mapping area and ground reference data, with the correlation coefficient of determination (R ²) of 0.94. The flood-prone areas were generally concentrated along the Mekong River and Tonle Sap Lake. The seasonal flooded areas observed for 2010, 2015 and 2016 were remarkably smaller than other years, mainly attributed to the El Niño weather phenomenon exacerbated by impacts of climate change. Eventually, the probability of flood occurrence during this 18-year period with reference to percentage of population at the district level was calculated for flood management purposes.