Tsunami Damage Detection Using Moderate-resolution Satellite Imagery

A moment-magnitude 9.0 earthquake struck the area off the western coast of northern Sumatra on 26 December 2004, which triggered massive tsunamis inundating the coastal areas in the countries along the Indian Ocean Rim. Various moderateto high-resolution satellite images were obtained before and after the tsunami attack and used for emergency management. High-resolution satellite (e.g. IKONOS) images allowed us to visually identify tsunami-affected areas with satisfactorily high accuracy, whereas moderate-resolution satellites (e.g. TerraASTER) would be more valuable for this kind of large-scale disasters. Since the change in land cover classes before and after the tsunami gave a good differentiation of tsunami inundation, multi-spectral satellite imagery is also thought to be useful to detect this change. In this study, the analysis on the thresholds of the indices to indicate surface characteristics such as the normalized difference vegetation index (NDVI), soil index (NDSI), and water index (NDWI) was carried out using moderate-resolution imagery. The thresholds were studied in two cases; one is for the tsunami damage detection based on only the data acquired after the disaster, and the other is for the detection based on the difference data of these indices before and after the tsunami.