5:15 PM - 7:15 PM
[SGD03-P03] Slip distribution estimation of the 2014-2016 long-term slow slip event in the Kii Channel using multiple GNSS positioning solutions: Examination of error reduction methods
Keywords:slow slip event, Kii Channel
To estimate slip distributions of Slow Slip Events (SSEs) using GNSS data, preprocessing is necessary to properly account for common mode error components and the effects of postseismic deformation from large earthquakes. In SSE research in Japan, methods widely used include spatial filtering for reducing common mode error components as proposed by Wdowinski et al. (1997), and regression analysis using logarithmic and exponential functions to model postseismic deformation from the 2011 Tohoku-oki earthquake (e.g., Tobita, 2016). Furthermore, there are studies that use multivariate analysis methods such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA) to separate common mode error components and postseismic deformation from GNSS time series data (e.g., Dong et al., 2006; Savage and Svarc, 2009; Gualandi et al., 2016).
This study focuses on the long-term SSE that occurred in the Kii Channel from 2014 to 2016, as reported by Kobayashi (2017). This long-term SSE has been discussed in relation to the gap in deep tremors and low-frequency earthquakes that are continuously distributed along the strike of the Nankai Trough (e.g., Mitsui et al., 2022; Ohta et al., 2023), making it worthy of more detailed analysis. Therefore, we attempt to reduce the effects of common mode error components and postseismic deformation from the 2011 Tohoku-oki earthquake using multiple approaches. For GNSS data, we use and compare not only the F5 solutions from the Geospatial Information Authority of Japan (Muramatsu et al., 2021) but also precise point positioning solutions from the Nevada Geodetic Laboratory (Blewitt et al., 2018), double-difference relative positioning solutions (by GAMIT) and precise point positioning solutions (by GipsyX) from ISTerre.
First, we corrected for the effects of postseismic deformation from the 2011 Tohoku-oki earthquake using logarithmic function approximation and estimated the SSE slip distribution based on Bayesian inversion analysis. As a result, using precise point positioning solutions, a concentrated slip distribution was estimated off the southwestern coast of the Kii Peninsula. This slip corresponds to a somewhat shallower area at the edge of the band-like region where deep tremors and low-frequency earthquakes occur beneath the Kii Peninsula, rather than the gap in deep tremors and low-frequency earthquakes. Near this slip distribution, low-frequency earthquake activity (based on Kato and Nakagawa, 2020) slightly increased during the SSE period, suggesting a phenomenon similar to the triggering of tremors by long-term SSEs reported in the Bungo Channel (Hirose et al., 2010).
On the other hand, when using F5 solutions and double-difference relative positioning solutions, slip was estimated not only off the southwestern coast of the Kii Peninsula but also in other areas such as off Cape Muroto. However, when inversion analysis was performed after applying common mode error reduction processing, these slips almost disappeared, suggesting they were likely ghosts due to common mode error. Since data preprocessing significantly affects SSE estimation results, it is necessary to seek more appropriate methods for removing postseismic deformation and further examine the validity of the slip distribution estimated off the southwestern coast of the Kii Peninsula.
This study focuses on the long-term SSE that occurred in the Kii Channel from 2014 to 2016, as reported by Kobayashi (2017). This long-term SSE has been discussed in relation to the gap in deep tremors and low-frequency earthquakes that are continuously distributed along the strike of the Nankai Trough (e.g., Mitsui et al., 2022; Ohta et al., 2023), making it worthy of more detailed analysis. Therefore, we attempt to reduce the effects of common mode error components and postseismic deformation from the 2011 Tohoku-oki earthquake using multiple approaches. For GNSS data, we use and compare not only the F5 solutions from the Geospatial Information Authority of Japan (Muramatsu et al., 2021) but also precise point positioning solutions from the Nevada Geodetic Laboratory (Blewitt et al., 2018), double-difference relative positioning solutions (by GAMIT) and precise point positioning solutions (by GipsyX) from ISTerre.
First, we corrected for the effects of postseismic deformation from the 2011 Tohoku-oki earthquake using logarithmic function approximation and estimated the SSE slip distribution based on Bayesian inversion analysis. As a result, using precise point positioning solutions, a concentrated slip distribution was estimated off the southwestern coast of the Kii Peninsula. This slip corresponds to a somewhat shallower area at the edge of the band-like region where deep tremors and low-frequency earthquakes occur beneath the Kii Peninsula, rather than the gap in deep tremors and low-frequency earthquakes. Near this slip distribution, low-frequency earthquake activity (based on Kato and Nakagawa, 2020) slightly increased during the SSE period, suggesting a phenomenon similar to the triggering of tremors by long-term SSEs reported in the Bungo Channel (Hirose et al., 2010).
On the other hand, when using F5 solutions and double-difference relative positioning solutions, slip was estimated not only off the southwestern coast of the Kii Peninsula but also in other areas such as off Cape Muroto. However, when inversion analysis was performed after applying common mode error reduction processing, these slips almost disappeared, suggesting they were likely ghosts due to common mode error. Since data preprocessing significantly affects SSE estimation results, it is necessary to seek more appropriate methods for removing postseismic deformation and further examine the validity of the slip distribution estimated off the southwestern coast of the Kii Peninsula.