2:00 PM - 2:15 PM
[SGD02-02] Shallow interplate slip deficit inferred from GNSS-Acoustic observations off Nemuro
Keywords:GNSS-Acoustic observation, Kuril Trench, Seafloor geodetic observation, interplate locking
Along the southern Kuril trench, M7–8 class interplate earthquakes segmented along the trench have repeatedly occurred. Moreover, Ioki & Tanioka (2016) suggested that Mw8.8 giant interplate earthquake with a large shallow rupture might have occurred along the multiple segments from Tokachi to Nemuro in the 17th century. Then, the possibility of a future large interplate earthquake with a large rupture reaching the trench (i.e., Slip To the Trench; e.g., the 2011 Tohoku earthquake) is concerned to be caused. However, it is difficult to detect actual accumulation of the slip deficit near the trench from onshore geodetic observations. Hence, Tohoku university and Hokkaido university deployed three GNSS-Acoustic (GNSS-A) observation sites off Nemuro in 2019 and have repeatedly conducted GNSS-A surveys using a research vessel and a Wave Glider operated by JAMSTEC.
Two of the above three sites locate on the landward slope (G21: 100 km from the trench, G22: 35 km from the trench), and the other locates on the incoming Pacific plate (G23: 35 km from the trench). We conducted six campaign surveys (2019/7, 2020/10, 2021/4, 2022/5, 2023/4, 2023/10) at G21 and five campaign surveys (2019/7, 2020/10, 2021/4, 2022/5, 2023/10) at G22 and G23. Among these campaign surveys, we conducted simultaneous GNSS-A observations using both the research vessel and the Wave Glider at the 2021/4 campaign for G22, the 2022/5 campaign for G22, and the 2023/10 for G23.
We analyzed GNSS data by kinematic relative positioning (onshore reference points: Akkeshi and Erimo) using RTKLIB v2.4.2 (Takasu, 2013) and obtained travel-times from the acoustic data using the phase-only correlation and the template matching techniques proposed by Honsho et al. ( 2021). We then estimated an array displacement for each campaign using the GNSS-A positioning software SeaGap v1.1 (Tomita and Kido, 2023; Tomita, 2023). In the GNSS-A positioning, we assumed an underwater sound speed structure with a single gradient layer and performed the static positioning for each campaign dataset by a MCMC technique. Note that constraints (prior distributions) for the parameters expressing the sound speed gradient (gradient depth and shallow gradient) were imposed to obtain robust positioning results. Regarding the gradient depth, we provided a prior distribution which is a normal distribution with mean of 650 m depth that was proposed in our previous study (Tomita and Kido, 2023). For the campaign datasets with simultaneous GNSS-A observations using both the research vessel and the Wave Glider, it is preferable to perform the joint positioning using both of them. However, this study separately analyzed them due to the current inversion strategy.
The results show displacement rates of ~6–9 cm/yr toward the north-west direction relative to the Okhotsk plate motion at the three sites (G21: ~5.59 cm/yr, N288°; G22: ~9.22 cm/yr, N294°; G23: ~7.34 cm/yr, N283°). We additionally performed the GNSS-A positioning under different inversion conditions (no prior distribution and another prior distribution with zero mean normal distribution on the gradient depth) and obtained over 8.2 cm/yr at G22 in the any cases. This significant displacement rate at G22 is comparable with a surface displacement rate calculated from a simple model assuming a full-locking condition on the shallow plate interface with a slip deficit rate of 9.1 cm/yr (N293°) which corresponds to the subduction rate of the Pacific plate (MOVEL: DeMets et al., 2010), and it suggests strong slip deficit near the trench off Nemuro.
As future works, we would like to perform the joint analysis of the simultaneous observational datasets from the multiple sea-surface platforms and to estimate a detailed slip deficit distribution using the above GNSS-A observational results.
Two of the above three sites locate on the landward slope (G21: 100 km from the trench, G22: 35 km from the trench), and the other locates on the incoming Pacific plate (G23: 35 km from the trench). We conducted six campaign surveys (2019/7, 2020/10, 2021/4, 2022/5, 2023/4, 2023/10) at G21 and five campaign surveys (2019/7, 2020/10, 2021/4, 2022/5, 2023/10) at G22 and G23. Among these campaign surveys, we conducted simultaneous GNSS-A observations using both the research vessel and the Wave Glider at the 2021/4 campaign for G22, the 2022/5 campaign for G22, and the 2023/10 for G23.
We analyzed GNSS data by kinematic relative positioning (onshore reference points: Akkeshi and Erimo) using RTKLIB v2.4.2 (Takasu, 2013) and obtained travel-times from the acoustic data using the phase-only correlation and the template matching techniques proposed by Honsho et al. ( 2021). We then estimated an array displacement for each campaign using the GNSS-A positioning software SeaGap v1.1 (Tomita and Kido, 2023; Tomita, 2023). In the GNSS-A positioning, we assumed an underwater sound speed structure with a single gradient layer and performed the static positioning for each campaign dataset by a MCMC technique. Note that constraints (prior distributions) for the parameters expressing the sound speed gradient (gradient depth and shallow gradient) were imposed to obtain robust positioning results. Regarding the gradient depth, we provided a prior distribution which is a normal distribution with mean of 650 m depth that was proposed in our previous study (Tomita and Kido, 2023). For the campaign datasets with simultaneous GNSS-A observations using both the research vessel and the Wave Glider, it is preferable to perform the joint positioning using both of them. However, this study separately analyzed them due to the current inversion strategy.
The results show displacement rates of ~6–9 cm/yr toward the north-west direction relative to the Okhotsk plate motion at the three sites (G21: ~5.59 cm/yr, N288°; G22: ~9.22 cm/yr, N294°; G23: ~7.34 cm/yr, N283°). We additionally performed the GNSS-A positioning under different inversion conditions (no prior distribution and another prior distribution with zero mean normal distribution on the gradient depth) and obtained over 8.2 cm/yr at G22 in the any cases. This significant displacement rate at G22 is comparable with a surface displacement rate calculated from a simple model assuming a full-locking condition on the shallow plate interface with a slip deficit rate of 9.1 cm/yr (N293°) which corresponds to the subduction rate of the Pacific plate (MOVEL: DeMets et al., 2010), and it suggests strong slip deficit near the trench off Nemuro.
As future works, we would like to perform the joint analysis of the simultaneous observational datasets from the multiple sea-surface platforms and to estimate a detailed slip deficit distribution using the above GNSS-A observational results.