13:45 〜 14:00
[S01-08] Extended Model Space to Account for Effects of Subwavelength Heterogeneities in Full Waveform Inversion Based on Wavefield Gradients
Recent advancements in seismic sensors, such as distributed acoustic sensing (DAS) and rotational sensors, have greatly enhanced the significance of wavefield gradient measurements. Wavefield gradient measurements have long been recognized for their sensitivity to localized sub-wavelength structures compared to translational motions. Consequently, it is important to pay attention to the distinct nature of the wavefield gradients compared with translational motions when analysing wavefield gradient measurements.
Seismic full waveform inversion (FWI) utilizes mostly translational motions including displacement, velocity, and acceleration for numerical tests and real-field applications. Most FWI schemes implicitly rely on the insensitivity of translational motions to sub-wavelength medium heterogeneities. This allows for the design of model parametrization and spatial resolution based on the maximum frequency of the inverted filtered data. However, this principle cannot apply to FWI based on wavefield gradients, and the impacts of small-scale heterogeneities on wavefield gradient-based FWI need to be thoroughly investigated.
This study aims to investigate the influence of sub-wavelength heterogeneities on wavefield gradient-based FWI. Our findings underscore the strong influence of small-scale heterogeneities on FWI inversion results. Knowing the vulnerability of wavefield gradient-based FWI to small-scale heterogeneities, we propose a new FWI framework that extends the model space to effectively mitigate difficulties. Furthermore, we demonstrate that the introduction of new parameters in the FWI, termed "correctors," within this extended model space, holds the potential to reveal subsurface structural images with an unprecedented high spatial resolution.
This work was supported by the Japan Society for the Promotion of Science (JSPS) through a Grant-in-Aid for JSPS Research Fellowship (22KJ2397) and JSPS Overseas Challenge Program for Young Researchers.
Seismic full waveform inversion (FWI) utilizes mostly translational motions including displacement, velocity, and acceleration for numerical tests and real-field applications. Most FWI schemes implicitly rely on the insensitivity of translational motions to sub-wavelength medium heterogeneities. This allows for the design of model parametrization and spatial resolution based on the maximum frequency of the inverted filtered data. However, this principle cannot apply to FWI based on wavefield gradients, and the impacts of small-scale heterogeneities on wavefield gradient-based FWI need to be thoroughly investigated.
This study aims to investigate the influence of sub-wavelength heterogeneities on wavefield gradient-based FWI. Our findings underscore the strong influence of small-scale heterogeneities on FWI inversion results. Knowing the vulnerability of wavefield gradient-based FWI to small-scale heterogeneities, we propose a new FWI framework that extends the model space to effectively mitigate difficulties. Furthermore, we demonstrate that the introduction of new parameters in the FWI, termed "correctors," within this extended model space, holds the potential to reveal subsurface structural images with an unprecedented high spatial resolution.
This work was supported by the Japan Society for the Promotion of Science (JSPS) through a Grant-in-Aid for JSPS Research Fellowship (22KJ2397) and JSPS Overseas Challenge Program for Young Researchers.