5:15 PM - 6:45 PM
[SCG55-P04] Paraneterization of outer-rise seismic normal fault and its implication to subduction zone tectonics
Keywords:Outer-rise, Seismic fault, Liner discriminant analysis, Seafloor roughness
On the surface of oceanic plate within 100 km from the trench axis, we generally observe a topographic rise called the outer rise. Since the outer rise is in a tensile field, normal faults are formed intensively. As a result, horst and graben structures up to 100 km scale are developed along the trench. It is expected that structures affect frictional properties and fluid behavior on the plate interface after subduction. In this study, we attempt to parameterize normal faults in outer rises, and compare them to seismic and geodetic coupling in these regions.
Elevation data are obtained from ETOPO2022 (15 sec grid) and multibeam surveys. We use two methods to parameterize faults. First, we detect the fault scarp from the elevation data and parameterized it based on the individual fault geometry (Vega-Ramírez et al., 2021). In order to do this, the elevation data are first extracted for a one-dimensional window and fitted to the diffusion equation to identify the height, diffusion age, and RMSE. A machine learning approach called linear discriminant analysis is then used to classify whether the detected scarp is a fault or not. Since continuous windows tend to detect the same fault scarp, kernel density estimation is applied to improve the detection results. Second, we define the amplitude of the fault as the roughness of the seafloor (Lallemand et al., 2018). For this purpose, we define the relatively short-wavelength roughness by performing a 2-D Fourier transform of the elevation data within a 2-D window and integrating the power spectral density for a specific wavelength band.
By comparing these results to various data in subduction zones, including seismic and geodetic couplings, we discuss the influence of outer-rise normal faults on dynamics of the plate.
Elevation data are obtained from ETOPO2022 (15 sec grid) and multibeam surveys. We use two methods to parameterize faults. First, we detect the fault scarp from the elevation data and parameterized it based on the individual fault geometry (Vega-Ramírez et al., 2021). In order to do this, the elevation data are first extracted for a one-dimensional window and fitted to the diffusion equation to identify the height, diffusion age, and RMSE. A machine learning approach called linear discriminant analysis is then used to classify whether the detected scarp is a fault or not. Since continuous windows tend to detect the same fault scarp, kernel density estimation is applied to improve the detection results. Second, we define the amplitude of the fault as the roughness of the seafloor (Lallemand et al., 2018). For this purpose, we define the relatively short-wavelength roughness by performing a 2-D Fourier transform of the elevation data within a 2-D window and integrating the power spectral density for a specific wavelength band.
By comparing these results to various data in subduction zones, including seismic and geodetic couplings, we discuss the influence of outer-rise normal faults on dynamics of the plate.