3:30 PM - 4:30 PM
[J06-P-05] Estimation of the spatiotemporal evolution of slow slip events in the Tokai region, central Japan, since 2013 using GNSS data
In the Tokai region, central Japan, the previous long-term slow slip event (L-SSE) occurred on the subducting Philippine Sea Plate (PSP) from 2000 to 2005. Moreover, many short-term SSEs (S-SSEs) have been observed in the Tokai region since 1996. Ozawa et al.(2016) reported that a SSE seems to have started in the similar area of the previous Tokai L-SSE in the beginning of 2013. We applied a time-dependent inversion method to GNSS data to obtain the spatiotemporal evolution of an L-SSE and S-SSEs on the PSP beneath the Tokai region, since 2013.
GNSS data from Jan. 1, 2008 to Dec. 30, 2015 were used in this study. The GIPSY-OASIS II software was used to estimate daily coordinates of 222 GNSS stations from the GEONET in the Tokai region. It is well known that GNSS time series have many systematic signals that do not result from SSEs. These systematic signals include, for example, seasonal variations and post-seismic deformation of the 2011 Tohoku-oki earthquake(Mw9.0). After removing these systematic signals, we applied a modified Network Inversion Filter (NIF) [Fukuda et al., 2008]. The original NIF [Segall & Matthews, 1997] assumes a constant hyperparameter for the temporal smoothing of slip rates and thus often results in oversmoothing of slip rates. The modified NIF assumes a time-variable hyperparameter, so that changes in slip rates are effectively extracted from GNSS time series.
The results indicate the moment magnitude and maximum cumulative slip of the SSE were estimated to be Mw～6.5 and ～6.5 cm from Jan. 1, 2013 to Dec. 30, 2015, respectively. In addition to the L-SSE from 2013 to 2015, two large S-SSEs were detected near the Ise Bay in the down-dip area of the L-SSE. Our results suggest that the slip peaks of the L-SSE and S-SSEs do not overlap and that the temporal variation of moment evolution in the central area of the L-SSE is smooth and is not affected by the S-SSEs. Low frequency tremors [Obara et al., 2010] do not occur near the center of the L-SSE.
GNSS data from Jan. 1, 2008 to Dec. 30, 2015 were used in this study. The GIPSY-OASIS II software was used to estimate daily coordinates of 222 GNSS stations from the GEONET in the Tokai region. It is well known that GNSS time series have many systematic signals that do not result from SSEs. These systematic signals include, for example, seasonal variations and post-seismic deformation of the 2011 Tohoku-oki earthquake(Mw9.0). After removing these systematic signals, we applied a modified Network Inversion Filter (NIF) [Fukuda et al., 2008]. The original NIF [Segall & Matthews, 1997] assumes a constant hyperparameter for the temporal smoothing of slip rates and thus often results in oversmoothing of slip rates. The modified NIF assumes a time-variable hyperparameter, so that changes in slip rates are effectively extracted from GNSS time series.
The results indicate the moment magnitude and maximum cumulative slip of the SSE were estimated to be Mw～6.5 and ～6.5 cm from Jan. 1, 2013 to Dec. 30, 2015, respectively. In addition to the L-SSE from 2013 to 2015, two large S-SSEs were detected near the Ise Bay in the down-dip area of the L-SSE. Our results suggest that the slip peaks of the L-SSE and S-SSEs do not overlap and that the temporal variation of moment evolution in the central area of the L-SSE is smooth and is not affected by the S-SSEs. Low frequency tremors [Obara et al., 2010] do not occur near the center of the L-SSE.