Sensitivity Study of ALOS-2 Data to Floodwaters in Joso City in 2015 and its Application

Abstract

Synthetic aperture radar (SAR) plays a key role in monitoring water-damaged areas because of its flexible operability with a wide swath. To date, many studies have investigated floodwater detection by X-band SAR. However, uniform extraction of physical information is essential to realize seamless monitoring of a damaged area after a devastating disaster using a limited number of SAR satellites at various bands. In this paper, we examined and demonstrated the applicability of a spaceborne L-band SAR to flooded areas. A temporal change of Advanced Land Observation Satellite-2 (ALOS-2) images through flooding was carefully extracted by histogram analysis. A physical understanding of the experimental facts revealed that a temporal change of the strong backscattering cross sections in the histogram could be the most sensitive and reliable index. A floodwater-detection algorithm based on the criterion was applied to the ALOS-2 data sets for the devastating 2015 flood in Joso City, Japan, in day-by-day variation. The results were validated qualitatively with sufficient aerial photos. We found that the spaceborne L-band SAR has reasonable applicability to floodwater detection, not only in suburban but also in urban areas by means of the histogram analysis.