Evaluation of Sentinel-1 data for flood mapping in the upstream of Sidi Salem dam (Northern Tunisia)

Abstract

Flood mapping is a powerful asset that allows drawing better strategies to contain possible economic repercussions and to rescue the affected population. This work is directly unfolded after the rainfall events that occurred in the north of the country, in February 2015, during which certain cities located in the vicinity of the Tunisian basin of Medjerda were flooded by the overflow of the Medjerda river, causing important damage to the towns of Jendouba and Bou Salem. The present research illustrates the potentiality of Sentinel-1 sensor in detecting flood areas in the upstream of Medjerda river. The Medjerda is the most important river in Tunisia, with an annual water potential reaching 0.8 billion m3. We compared the signature of flood water in vertical transmit and horizontal received (VH) and vertical transmit and vertical received (VV) polarizations of radar data. The study proves that the segregation of land/water areas with a threshold technique is better observed in VH polarization rather than VV polarization.