Lower Mekong river flood area monitored by multi-source remote sensing

Abstract

Remote sensing monitoring of submerged areas is an effective method to measure flooding, directly indicating severity of the disaster. This study uses MODIS, FY3A MERSI, HJ1A/B CCD and Landsat TM data to monitor time series of the flood inundation area in the Mekong River downstream in 2011, based on an optimal algorithm obtained from experiments. We evaluate the flood area with different types of vegetation using the MODIS land cover data. From the results the following recommendations are made. NWDI is the best algorithm for HJ1A/B CCD and FY3A MERSI. NDVI is more suitable for Landsat TM and MODIS data compared with other three algorithms. Cambodia and the Mekong Delta region had serious flooding disaster in October 2011, with the inundated area 6.5 times larger than the normal area. The Tonle Sap River basin was the worstaffected area, with the river widened by about 40 times. At the beginning of the flood, a large amount of water flowed into Tonle Sap Lake, which played an important role in storing flood water. Therefore, by combining advantages of multi-source remote sensing satellite data to monitor changes in the floods, we can acquire more detailed information and improve efficiency of flood detection.