16:15 〜 16:30
[SSS05-10] Relationship between observation ability of GNSS-A and its detection ability for coupling condition and SSE
キーワード：GNSS-A、SSE、プレート間固着
The observation frequency and observation accuracy of GNSS-A seafloor geodetic observations carried out by the Japan Coast Guard are currently 4 times/year and 2 cm (1-σ) due to technological development in about 20 years [Ishikawa et al., 2020]. By accumulating long-term observation data, it has become possible to measure interplate coupling condition and shallow slow slip events along the Nankai Trough in recent years [Yokota et al., 2016; Yokota and Ishikawa, 2020], and it has become possible to interpret the time change of postseismic deformation following the 2011 Tohoku-oki earthquake [Watanabe et al., 2014]. However, the physical process of the plate boundary can be understood by understanding the more detailed time change of a coupling condition and the postseismic processes and improving the SSE detection accuracy. This consideration will also be useful for understanding the observation results of new observation points. In this presentation, we will understand how a steady state and an event can be detected by improving GNSS-A observation ability in the future, and set development goals for that purpose.
In order to understand the detection ability of the steady state, it is sufficient to generate pseudo data with frequency and accuracy and compare the accuracy of linear regression. On the other hand, in order to understand the ability to detect events, comparison is made using detection power. Attached figure compares effect size of event step (here, the magnitude of crustal deformation due to an event) and required sample size (here, annual observation frequency). σ means the accuracy of GNSS-A obseravtion. It shows how much an event can be detected regarding to the GNSS-A observing ability. From this figure, it can be seen that when the scale of crustal movement caused by the event approaches the same level as the error, it is difficult to detect it at the current observation frequency (several times per year).
The observation ability about time constant cannot be determined by this numerical test. In this presentation, we also discuss the accuracy of event time constant and magnitude detections through some numerical experiments.
In order to understand the detection ability of the steady state, it is sufficient to generate pseudo data with frequency and accuracy and compare the accuracy of linear regression. On the other hand, in order to understand the ability to detect events, comparison is made using detection power. Attached figure compares effect size of event step (here, the magnitude of crustal deformation due to an event) and required sample size (here, annual observation frequency). σ means the accuracy of GNSS-A obseravtion. It shows how much an event can be detected regarding to the GNSS-A observing ability. From this figure, it can be seen that when the scale of crustal movement caused by the event approaches the same level as the error, it is difficult to detect it at the current observation frequency (several times per year).
The observation ability about time constant cannot be determined by this numerical test. In this presentation, we also discuss the accuracy of event time constant and magnitude detections through some numerical experiments.