AUTOMATED GROUND MOTION CHARACTERIZATION USING SATELLITE IMAGERY

Estimation of local site effects requires precise information about geomorphic and geologic conditions. On the basis of our ongoing effort (Yong et al., 2007), we report on our progress using multi-resolution, remote sensing data to provide preliminary site characterizations for estimating ground motion intensities in future earthquakes. Use of remote sensing to determine site effects is especially important in many regions of the world, such as Pakistan, Mozambique and Turkey, where local geologic information is either sparse or are not readily available. Even where local geologic information is available, details in the traditional maps often lack the precision or the level of information required for effective microzonation. We use readily available satellite-based data and state-of-the-art object-based image analysis methods to address this problem. Our data sets consist of (1) optical imagery, that include regions in the short-wave infrared (SWIR) and thermal infrared (TIR) domains, and (2) relative digital elevation models (DEM), based on stereoscopic-correlation methods, derived from NASA's ASTER (Advanced Space-borne Thermal Emission and Reflection Radiometer) sensors. On the basis of geomorphology and geology, we apply automated feature extraction methods to determine the local terrain. Then, on the basis of the site classification schemes from the Wills et al (2000) and Wills and Clahan (2006) maps of California, we assign Vs30 values to selected regions in California, Mozambique, Pakistan, and Turkey. We compare our results to available empirical data and discuss the applicability of our site class assignments in each region and the implications of its effects on seismic hazards assessment.