5:15 PM - 7:15 PM
[SGD03-P14] Spatio-temporal Changes in Seismic Attenuation in the Forearc Region Following the 2011 Tohoku-oki Earthquake
Keywords:Seismic Attenuation, Viscoelastic Relaxation, Poroelastic Rebound, Afterslip
Post-seismic surface displacement following the 2011 Tohoku-oki earthquake is thought to be explained by three mechanisms: afterslip, viscoelastic relaxation, and poroelastic rebound. However, these components have a trade-off relationship in their contributions to surface displacement, making it difficult to separate them using the observations of surface displacement alone. This study aims to provide a constraint of seismological observation on this trade-off through temporal changes in seismic attenuation.
Our analysis focuses on M3.0-5.0 earthquakes that occurred in the source region of Tohoku-oki earthquake during 2011-2023. First, we applied hypoDD to the JMA unified earthquake catalog as initial locations. Then, we selected the first one-third of all earthquakes in chronological order as the initial dataset. In our analysis, we first selected earthquake pairs with hypocentral distances <5 km within the dataset and calculated spectral ratios of P-direct wave at identical stations. These ratios were then normalized by the average travel-time of the two earthquakes to remove source distance effects. Subsequently, we stacked all earthquake pairs where the denominator event was older than the numerator event to obtain observed spectral ratios, which were then fitted with theoretical spectral ratios. The slope of spectral ratio becomes zero if there is no change in attenuation between two earthquakes, while it shows a positive slope when attenuation decreases over time and a negative slope when it increases. We tracked temporal changes in attenuation by repeating this analysis while sequentially replacing earthquakes in the dataset with the oldest 1/100 of the total number of events.
The results show a continuous decrease in attenuation over long-term scales (several years or more). However, short-term variations (less than several months) are difficult to interpret due to their strong dependence on the analysis region and period. Since the long-term attenuation changes may reflect structural changes due to viscoelastic relaxation in the lower crust, we will address methods to reliably extract short-term attenuation changes immediately after the mainshock for comparison with viscoelastic models.
Our analysis focuses on M3.0-5.0 earthquakes that occurred in the source region of Tohoku-oki earthquake during 2011-2023. First, we applied hypoDD to the JMA unified earthquake catalog as initial locations. Then, we selected the first one-third of all earthquakes in chronological order as the initial dataset. In our analysis, we first selected earthquake pairs with hypocentral distances <5 km within the dataset and calculated spectral ratios of P-direct wave at identical stations. These ratios were then normalized by the average travel-time of the two earthquakes to remove source distance effects. Subsequently, we stacked all earthquake pairs where the denominator event was older than the numerator event to obtain observed spectral ratios, which were then fitted with theoretical spectral ratios. The slope of spectral ratio becomes zero if there is no change in attenuation between two earthquakes, while it shows a positive slope when attenuation decreases over time and a negative slope when it increases. We tracked temporal changes in attenuation by repeating this analysis while sequentially replacing earthquakes in the dataset with the oldest 1/100 of the total number of events.
The results show a continuous decrease in attenuation over long-term scales (several years or more). However, short-term variations (less than several months) are difficult to interpret due to their strong dependence on the analysis region and period. Since the long-term attenuation changes may reflect structural changes due to viscoelastic relaxation in the lower crust, we will address methods to reliably extract short-term attenuation changes immediately after the mainshock for comparison with viscoelastic models.