Abstract
The disaster information system, the Early Damaged Area Estimation System (EDES), was developed to estimate damaged areas of natural disaster using the night-time imagery of the Defense Meteorological Satellite Program Operational Linescan System (DMSP-OLS). The system employs two estimation methods to detect the city lights loss or reduction as possible impacted areas; one is the bi-temporal images (BTI) method and the other is the time-series images (TSI) method. Both methods are based on significance tests assuming that brightness of city lights fluctuates as normal random variables, and the BTI method is simplified by introducing the assumption that the standard deviation of city lights fluctuation is constant. The validity of the estimation method is discussed based on the result of the application to the 2001 Western India earthquake disaster. The estimation results identify the damaged areas distant from the epicentre fairly well, especially when using the TSI method. The system is designed to estimate the global urban damage and to provide geographic information through the Internet within 24 h after a severe disaster event. The information is expected to support the disaster response and relief activities of governments and non-governmental organizations.
Acknowledgments
Development of the EDES was conducted as a joint research project of the National Oceanic and Atmospheric Administration, National Geophysical Data Center (NOAA/NGDC) and the Earthquake Disaster Mitigation Research Center, National Research Institute for Earth Science and Disaster Prevention (EDM/NIED). The DMSP-OLS images, Stable City Lights Image, and Radiance Calibrated Image were provided by NOAA/NGDC. DMSP-OLS images of India were distributed through the Ministry of Agriculture, Forestry and Fisheries Research Network's Satellite Image Database System in the Ministry of Agriculture, Forestry and Fisheries (MAFFIN-SIDaB) (Kodama and Song Citation2000). The authors would like to acknowledge the generosity of these organizations.