Features of coherence and coherence ratio images using PALSAR-2 data for the 2016 Kumamoto earthquake
Abstract
Recently, human and economic damages because of large-magnitude earthquakes have increased, particularly when densely populated cities are impacted. Satellite remote sensing is a powerful tool to spatially determine the scale and center of damages. Among them, the synthetic aperture radar is emphasized because of its ability to acquire data even in adverse weather conditions. Our study goal was to develop the methodology to acquire effective information needed for the initial operation and determine areas for detailed observation via airborne sensors or helicopters. Previously, we studied the interferometric features of several latest sensors. The object of the current study was to develop a method to extract the area of largely damaged buildings, and we need to reveal the features of coherence and coherence ratio for this purpose. This study concerned the earthquakes that struck Kumamoto, Japan, on April 14 and 16, 2016, during which many buildings collapsed in the town of Mashiki, southeast of Kumamoto city center. We used several pairs of advanced land observing satellite (ALOS)-2 phased array type L-band synthetic aperture radar (PALSAR)-2 datasets obtained before and after the event to conduct the coherence analysis. PALSAR-2, which began operation in 2014, is expected to be used in many applications, particularly disaster management. The orbit can be controlled with high accuracy, enabling improved coherence in PALSAR-2 compared with previous PALSAR. First, we classified the land cover into urban, natural, and water areas using the reference "high resolution land-use and land-cover map" produced by the advanced visible and near infrared radiometer (AVNIR-2) on the ALOS satellite operated during 2006-2011; however, classification errors occurred owing to product errors and differences in the acquisition time. Second, Stripmap mode data were acquired from February 2015 to May 2016 with different ascending and descending orbit directions to generate four coherence and two coherence ratio maps. Third, we quantitatively evaluated the range of both indices for each land-cover category. At last, we clipped the values of both indices to isolate the urban area and investigated the features of these indices. Results showed that both indices clearly classified the significantly damaged Mashiki town and the minimally damaged Kumamoto city center. The values of both indices around Mashiki town were approximately half of that in the Kumamoto city center. These features did not depend on the orbit directions, although the values for Kumamoto city center were slightly lower for descending than that for ascending orbit directions owing to the orientation of the buildings. Therefore, this study demonstrated that the proposed method can extract significantly damaged areas. Differences between the coherence and the coherence ratio were found in moving objects at the Kumamoto airport. The objects were not extracted as damaged by the coherence ratio but were extracted for coherence. In future studies, we will conduct further detailed analysis to acquire more knowledge regarding the relationships between the buildings' collapse rate and the values of the indices.
- Publication:
-
43rd COSPAR Scientific Assembly. Held 28 January - 4 February
- Pub Date:
- January 2021
- Bibcode:
- 2021cosp...43E.107N