Japan Geoscience Union Meeting 2024

Presentation information

[J] Oral

S (Solid Earth Sciences ) » S-SS Seismology

[S-SS07] Seismic wave propagation: Theory and Application

Tue. May 28, 2024 10:45 AM - 12:00 PM 304 (International Conference Hall, Makuhari Messe)

convener:Kaoru Sawazaki(National Research Institute for Earth Science and Disaster Resilience), Akiko Takeo(Earthquake Research Institutute, the University of Tokyo), Masafumi KATOU(JGI, Inc.), Kyosuke Okamoto(National Institute of Advanced Industrial Science and Technology), Chairperson:Kyosuke Okamoto(National Institute of Advanced Industrial Science and Technology), Ryo Kurihara(Hot Springs Research Institute of Kanagawa Prefecture)

11:30 AM - 12:00 PM

[SSS07-08] Efficient full waveform inversion using boundary integral representation for time-lapse crosswell monitoring without downhole sources

★Invited Papers

*Shohei Minato1 (1.National Institute of Advanced Industrial Science and Technology)

Keywords:Seismic imaging, Full waveform inversion, Boundary integral representation, Crosswell monitoring

Full waveform inversion (FWI) is a high-resolution imaging method that exploits amplitude information of observed data as well as travel times to estimate the spatial distribution of velocity. In this presentation, I focus on my research in seismic imaging, particularly on full waveform inversion. The field application of FWI to crosswell tomographic data demonstrates its capability to estimate a velocity structure with higher resolution than that obtained from travel time tomography, closely matching the logging data. I have also developed a new approach for accurate and efficient crosswell imaging for long-term monitoring, utilizing surface sources and borehole receivers. This new approach exploits a boundary integration representation of seismic wavefield, allowing for the extrapolation of wavefield response at one of the boreholes using that at the other borehole. Importantly, this method does not require a velocity model to describe the wave propagation from the surface sources to the borehole during forward and adjoint modelling of wavefields. This characteristic enables mitigating monitoring errors originating from poor knowledge of near-surface heterogeneity and source repeatability. The effectiveness of this proposed approach is demonstrated through field and synthetic experiments, showing that it is crucial for achieving efficient and accurate crosswell monitoring.