ABSTRACT
Determining the location of coseismic landslides damaged buildings is critical to the understanding of human losses from earthquake-induced landslides during major earthquakes in mountainous regions. Because of the vast area of the earthquake affected region and the sparse locations of rural settlements, previous knowledge on buildings and human losses resulting from coseismic landslides mainly rely on rough estimations from limited field work. Application of high resolution satellite remote sensing opens a new way to map pre-seismic rural mountainous housing over large areas, and thus can accurately determine the location of coseismic landslide damaged housing spatially. In this work, we used the 2008 Wenchuan Earthquake as an example and use SPOT5 images taken before the earthquake for a seimi-automated mapping of buildings in rural areas. Then, by spatial overlay analysis with an existing coseismic landslide inventory, we mapped the location of landslide damaged buildings in this area. This semi-automatic method can be further applied to the entire Wenchuan Earthquake affected region to determine the locations of all landslides damaged buildings and finally improve the performance of earthquake loss estimation models.
