3:30 PM - 3:45 PM
[SSS14-19] Reconstruction of surface ruptures based on optical correlation using pre-and post-earthquake aerial photographs: The case of the 1995 Hyogo-ken Nanbu earthquake
★Invited Papers
Keywords:1995 Hyogo-ken Nanbu earthquake, Nojima Fault, Optical correlation, Aerial photograph, maturity of active fault
Recent advances in remote sensing technology, typified by InSAR and LiDAR DEM differential analysis, have made it possible to quickly and accurately determine the distribution and characteristics of surface ruptures associated with inland earthquakes. On the other hand, in the case of InSAR analysis, the interferometers may include deformations caused by secondary earthquakes other than the mainshock due to the observation cycle and satellite orbit. In addition, of course, the analysis cannot apply to earthquakes that occurred before the operation of the SAR satellite or during the period of switching satellites. As for the LiDAR DEM differential analysis, we cannot perform the high-precision analysis if the DEM before the earthquake is not obtained. To compensate for these problems, we focused on optical correlation analysis using aerial photographs that capture horizontal ground deformation. In this study, we attempted to detect surface ruptures by orthorectification of aerial photographs and matching them using the Cosi-Corr method, focusing on the 1995 Hyogo-ken Nanbu earthquake, for which pre- and post-aerial photographs were taken.
As a result of the analysis, discontinuities approximately 10 km long were detected along the Nojima Fault on the west coast of Awaji Island, although there was much noise due to the orthorectification process and artificial modification during the photography period. Based on field surveys of previous papers, near the Toshima River was thought to be the southern end of the surface rupture, but the optical correlation implies that the discontinuity continues another 1.3 km to the southwest. On the other hand, we cannot detect significant discontinuities exceeding the noise level along the Nadagawa Fault ruptures on the east coast of Awaji Island, and Kobe city side beyond the Akashi Strait. We measured the horizontal displacement along the discontinuities at cross-distances of ±25 m, ±35 m, ±50 m, and ±100 m apertures from the discontinuity and compared the results with the field measurements of previous studies. In particular, the displacement values for the cross-distances of ±50 m and ±100 m apertures were larger than the on-site measurement values. We took measurements at numerous points along the Nojima Fault, and when the ratio of the displacement value of ±100 m aperture was used as the denominator, suggesting that the near-fault displacements were responsible for approximately 50 to 70% of the total one.
Compared to similar attempts in arid and non-residential regions overseas, there are still issues to be addressed regarding detection limits and noise reduction. However, this study has shown that optical correlation using aerial photographs is also effective to some extent in terms of land cover conditions in Japan. In principle, it is possible to apply this method to earthquakes before the 1990s, starting from aerial photography by the US military after the war.
As a result of the analysis, discontinuities approximately 10 km long were detected along the Nojima Fault on the west coast of Awaji Island, although there was much noise due to the orthorectification process and artificial modification during the photography period. Based on field surveys of previous papers, near the Toshima River was thought to be the southern end of the surface rupture, but the optical correlation implies that the discontinuity continues another 1.3 km to the southwest. On the other hand, we cannot detect significant discontinuities exceeding the noise level along the Nadagawa Fault ruptures on the east coast of Awaji Island, and Kobe city side beyond the Akashi Strait. We measured the horizontal displacement along the discontinuities at cross-distances of ±25 m, ±35 m, ±50 m, and ±100 m apertures from the discontinuity and compared the results with the field measurements of previous studies. In particular, the displacement values for the cross-distances of ±50 m and ±100 m apertures were larger than the on-site measurement values. We took measurements at numerous points along the Nojima Fault, and when the ratio of the displacement value of ±100 m aperture was used as the denominator, suggesting that the near-fault displacements were responsible for approximately 50 to 70% of the total one.
Compared to similar attempts in arid and non-residential regions overseas, there are still issues to be addressed regarding detection limits and noise reduction. However, this study has shown that optical correlation using aerial photographs is also effective to some extent in terms of land cover conditions in Japan. In principle, it is possible to apply this method to earthquakes before the 1990s, starting from aerial photography by the US military after the war.