Evaluation of remote sensing-based seismic site characterization using earthquake damage data

Preparing seismic site characterization maps requires extensive geological and geotechnical data collection and interpretation. This study evaluates a remote sensing-based method to generate seismic site characterization map at 30-m spatial resolution. Geographical object-based image analysis was applied to topographic attributes computed from an Advance Spaceborne Thermal Emission and Reflection Radiometer Digital Elevation Model to demarcate terrane units of mountains, piedmonts and basins. Classified terrane units were verified through a field-based landform map resulting in an accuracy of 70%. To evaluate the importance of classified terrane units for seismic site characterization and their role in earthquake-induced building damage, the classified terrane units were compared with 2005 Kashmir earthquake damage data derived from field visits and SPOT-5 imagery. This shows that 81% of the less, moderate and severe damage intensity zones correspond with mountains, piedmonts and basins respectively. Classified terrane units were assigned Vs30 to generate a seismic site characterization map.