11:30 AM - 11:45 AM
[SCG45-10] GNSS/Acoustic geodetic measurement at the southern end of Ryukyu Trench 2014-2020
Keywords:Yaeyama Tsunami, Ryukyu Trench, GNSS/Acoustic measurement, Seafloor crustal deformation, Inter-plate coupling
We present the results of six years of GNSS/acoustic measurement conducted from October 2014 to September 2020 in the forearc basin of the Yaeyama Islands, the southernmost part of the Ryukyu Trench.
To the south of the Yaeyama Islands, the Philippine Sea Plate is subducting under the Eurasian Plate (Yangtze Plate) at a rate of about 8 cm/yr along the Ryukyu Trench. Behind the Yaeyama Islands, the Okinawa Trough back arc basin is opening at a rate of about 4-5 cm/yr. As a result, the convergence rate between the Yaeyama Islands and the Philippine Sea Plate is fast at about 12-13 cm/yr. The Ryukyu subduction zone beneath the Yaeyama Islands is an extensional plate boundary where back-arc opening occurs, and it has been thought that the interplate coupling is weak and does not capable of huge earthquakes (Scholz and Campos, 2012, JGR).
However, there are several old legends of tsunamis in Miyakojima and Ishigakijima islands, and it is known that the Yaeyama earthquake of April 24, 1771 caused a devastating tsunami in these islands (Nakamura, 2009, GRL).
Nakamura (2009, GRL) conducted a numerical simulation of the tsunami based on the Tsunami inundation data by Nakata and Kawana (1995), and estimated the Yaeyama earthquake as a tsunamigenic earthquake of Mw 8.0 that occurred in the Ryukyu Trench. In addition, trenching of tsunami deposits in Ishigaki Island estimated that the Yaeyama earthquake was a massive earthquake with severe strong ground motion, and that tsunamis of approximately the same size as the 1771 earthquake occurred four times in the past 2000 years at intervals of about 600 years (Ando et al., 2018, Tectonophysics). In order to estimate the presence or absence of the inter-plate coupling around the tsunami fault area estimated by Nakamura (2009, GRL), seafloor transponders were deployed in the forearc region at a depth of about 3300 m, about 40 km off the south coast of Hateruma Island, by Shizuoka University and the University of the Ryukyus in October 2014. In this study, we estimated the motion of the transponders over a period of six years from the observation data taken seven times in total from October 2014 to September 2020.
The position of the transponders was estimated using the method of Ikuta et al. (2008, JGR) using kinematic GPS, ship attitude, two-way travel time of acoustic ranging, and XCTD data. To estimate the kinematic GPS positions, we used Interferometry Trajectory (IT), a long baseline analysis software developed by NASA/GSFC in six of the seven campaigns expect for the campaign in 2014. In 2014 campaign, the results of the IT analysis showed a large discrepancy between the astronomical tide and the GPS antenna height change. We reanalyzed the kinematic GPS for the 2014 campaign only using RTKLIB (Takasu, 2006). For the estimation of the three transponders positions, we introduced the assumption that the sound velocity structure changes smoothly with time and that the geometries of them do not change through all campaigns. And we estimated the positions of the three transponders for all campaigns simultaneously by penalized least squares method.
As a result, the average velocity vectors calculated from the positions of the transponders for all campaigns were estimated to be 64.2±11.2 mm/yr in the south direction, 25.8±9.8 mm/yr in the east direction, and 24.8±8.0 mm/yr in the subsidence direction relative to the Eurasian plate. The rms of the residuals between the predicted and actual transponders positions for each campaign at the calculated velocities were 54.3 mm north-south, 47.8 mm east-west, and 39.0 mm up-down; before the reanalysis of the 2014 GNSS data by RTKLIB, these were 61.0 mm, 66.4 mm, and 40.7 mm, respectively. It can be seen that the improvement of GNSS data in 2014 affected the analysis results.
In this presentation, based on the above results, we discuss the extent of inter-plate coupling at this southernmost tip of the Ryukyu Trench.
To the south of the Yaeyama Islands, the Philippine Sea Plate is subducting under the Eurasian Plate (Yangtze Plate) at a rate of about 8 cm/yr along the Ryukyu Trench. Behind the Yaeyama Islands, the Okinawa Trough back arc basin is opening at a rate of about 4-5 cm/yr. As a result, the convergence rate between the Yaeyama Islands and the Philippine Sea Plate is fast at about 12-13 cm/yr. The Ryukyu subduction zone beneath the Yaeyama Islands is an extensional plate boundary where back-arc opening occurs, and it has been thought that the interplate coupling is weak and does not capable of huge earthquakes (Scholz and Campos, 2012, JGR).
However, there are several old legends of tsunamis in Miyakojima and Ishigakijima islands, and it is known that the Yaeyama earthquake of April 24, 1771 caused a devastating tsunami in these islands (Nakamura, 2009, GRL).
Nakamura (2009, GRL) conducted a numerical simulation of the tsunami based on the Tsunami inundation data by Nakata and Kawana (1995), and estimated the Yaeyama earthquake as a tsunamigenic earthquake of Mw 8.0 that occurred in the Ryukyu Trench. In addition, trenching of tsunami deposits in Ishigaki Island estimated that the Yaeyama earthquake was a massive earthquake with severe strong ground motion, and that tsunamis of approximately the same size as the 1771 earthquake occurred four times in the past 2000 years at intervals of about 600 years (Ando et al., 2018, Tectonophysics). In order to estimate the presence or absence of the inter-plate coupling around the tsunami fault area estimated by Nakamura (2009, GRL), seafloor transponders were deployed in the forearc region at a depth of about 3300 m, about 40 km off the south coast of Hateruma Island, by Shizuoka University and the University of the Ryukyus in October 2014. In this study, we estimated the motion of the transponders over a period of six years from the observation data taken seven times in total from October 2014 to September 2020.
The position of the transponders was estimated using the method of Ikuta et al. (2008, JGR) using kinematic GPS, ship attitude, two-way travel time of acoustic ranging, and XCTD data. To estimate the kinematic GPS positions, we used Interferometry Trajectory (IT), a long baseline analysis software developed by NASA/GSFC in six of the seven campaigns expect for the campaign in 2014. In 2014 campaign, the results of the IT analysis showed a large discrepancy between the astronomical tide and the GPS antenna height change. We reanalyzed the kinematic GPS for the 2014 campaign only using RTKLIB (Takasu, 2006). For the estimation of the three transponders positions, we introduced the assumption that the sound velocity structure changes smoothly with time and that the geometries of them do not change through all campaigns. And we estimated the positions of the three transponders for all campaigns simultaneously by penalized least squares method.
As a result, the average velocity vectors calculated from the positions of the transponders for all campaigns were estimated to be 64.2±11.2 mm/yr in the south direction, 25.8±9.8 mm/yr in the east direction, and 24.8±8.0 mm/yr in the subsidence direction relative to the Eurasian plate. The rms of the residuals between the predicted and actual transponders positions for each campaign at the calculated velocities were 54.3 mm north-south, 47.8 mm east-west, and 39.0 mm up-down; before the reanalysis of the 2014 GNSS data by RTKLIB, these were 61.0 mm, 66.4 mm, and 40.7 mm, respectively. It can be seen that the improvement of GNSS data in 2014 affected the analysis results.
In this presentation, based on the above results, we discuss the extent of inter-plate coupling at this southernmost tip of the Ryukyu Trench.