CHARACTERISTICS OF SATELLITE IMAGES OF DAMAGED AREAS DUE TO THE 1995 KOBE EARTHQUAKE

Several satellites have on-board optical sensors and/or synthetic aperture radar (SAR), with which the Kobe area before and after the 1995 Hyogoken-Nanbu (Kobe) Earthquake in Japan was observed. Since a part of the damage survey results of this earthquake is maintained as GIS data, a quantitative analysis of the surface changes in damaged areas is possible. Spectral characteristics and backscattering signatures of the area damaged by the earthquake were investigated using Landsat, JERS and ERS images taken before and after the earthquake, to examine the possibility of extracting earthquake damage distribution by satellite remote sensing. Introduction To gather information on damage due to natural disasters, several methods exist, such as field survey, aerial videography and photography, and satellite imagery. Satellite remote sensing, by which a large area is easily monitored, may provide effective information at the time of recovery activity, e.g., developing a restoration plan, if it is possible to determine the distribution of damage due to disasters at an early stage. Several satellites observed the Kobe area before and after the 1995 Hyogoken-Nanbu (Kobe) Earthquake which occurred on January 17, 1995 [1]. Multispectral characteristics were different between images of liquefied areas and burned areas taken by airborne remote sensing just after the earthquake occurred [2]. A study suggested the possibility of interpreting the damaged area based on the spectral pattern changes between optical satellite images taken before and after this earthquake [3]. Synthetic aperture radar (SAR) observations can be performed night and day. This feature could be useful for effective postdisaster assessment when a field survey for a large area is difficult. Massonnet et al. (1993) applied SAR interferometric analysis to estimate the distribution of ground displacement due to the 1992 Landars Earthquake [4]. Earth observation satellites equipped with SAR, containing information on the amplitude and phase of microwave backscattering from objects on the earth surface, were also used to observe the Kobe area. The spatial distribution of ground displacement extracted from SAR interferograms using phase images and a feasibility study on interpreting damaged built-up areas using SAR intensity images were reported [5][6]. Applications of satellite remote sensing to identify the damaged areas were attempted, but no quantitative approach was found for examining the relationship between the spectrum characteristics and backscattering properties of the damaged area using satellite images and detailed damage survey results. Since a part of the damage survey results of this earthquake was maintained as GIS data, a quantitative analysis of the surface changes in the damaged area is possible. The authors have already reported preliminary studies on the possibility of identifying earthquake damage using images from Landsat and SPOT as optical images and from ERS as SAR images [7][8]. In this study, the spectral characteristics and the backscattering signatures of the damage due to this earthquake were investigated using the satellite images from Landsat/TM, ERS/SAR and JERS/SAR, and modified analytical procedures were performed. Earthquake Damage Survey Data Liquefaction and building damage were focused on as forms of earthquake damage in this study. Sand deposits boiled by the earthquake were digitized by authors on the 1/50,000-scale ground-failure survey map [9] and used as liquefied area data. The building damage data based on detailed survey results compiled by AIJ (the Architectural Institute of Japan) and CPIJ (the City Planning Institute of Japan), and digitized by BRI (Building Research Institute, Ministry of Construction) were utilized as GIS data. In the GIS data, the building damage level was classified into the five categories: damage by fire, severe structural damage, moderate damage, slight damage and no damage, and the numbers of damaged buildings were totaled for each block in each ward [10]. Satellite Images The Landsat/TM, which is an optical sensor, was used to observe the area of interest on January 24, 1995. The JERS-1/SAR and ERS-1/SAR performed observations on February 5 and May 23, 1995, respectively. Because the sensitivity of the sensor installed in the SPOT satellite to detect earthquake damage is low in spite of its spatial high resolution [7], the SPOT image was not used in this study. We used the images taken on August 17, 1994 by Landsat, on May 30, 1993 by JERS and on October 12, 1994 by ERS for data before the earthquake, and aimed to examine the change in the spectrum and backscattering characteristics of the damaged area. The main characteristics of optical satellites and SAR satellites are listed in Table 1 and Table 2, respectively. The areas covered by the satellite images are shown in Fig. 1. Satellite Image Processing Optical Image Modification: Because the digitized values in the satellite images were different depending on the observation and surface conditions, digital number (DN) modification is required before starting this study. The characteristics of the reflection of Table 1 Characteristics of Landsat TM Number of bands Visible: 3 bands Nearand mid-infrared: 3 bands Thermal infrared: 1 band