今週のうなぎセミナーについてお知らせいたします。

Here is information of the Unagi-seminar(October, 19).

************** Seminar on Seismology IV B, D /地震学ゼミナールIV B, D (Unagi Seminar) **************

科目：地震学ゼミナールIV B, D / Seminar on Seismology IV B, D（修士・博士）
日時：2023年 10月 19日 (木) 13:30~
場所：京都大学 防災研究所 本館E-232D または オンライン(Zoom)

Date and Time：2023-10-19, 13:30~
Place：Uji Campus Main Building E232D or Zoom (Hybrid)

　　　Please join the seminar on-site, especially students who need credit.

---------- ---------- ---------- ----------

Speaker（発表者）:　Atikul Haque Farazi

Title（題目）:　
Sedimentary and shallow crustal structure from passive seismic methods: cases in onshore and offshore subduction zones

Abstract:
　Estimation of shear wave (S-wave) velocity (VS) from shallow to deep subsurface is crucial for engineering projects and seismic hazard study (e.g., site characterization, microzonation), industrial exploration, and the understanding of structural and tectonic control of a region. Seismic wavefields emerging from passive seismic sources, i.e., seismic ambient noise (SAN) and earthquakes, have been successfully used for decades to serve this purpose. In these kinds of studies, the observables are forward modeled and inverted for subsurface velocity versus depth profile under various theories of physics, which mainly involve reconstruction of the elastodynamic Green’s function (GF) to reconstruct the seismic wavefield in turn. The advent of the diffuse field theory of the seismic wavefield in the past decade led seismologists to reconstruct the GF by considering the full seismic wavefield or contribution from all kinds of seismic waves. The diffuse field assumption (DFA) has been successfully used so far, both onshore and offshore, for estimating the subsurface velocity profile by forward modeling of an observable and then inversion.
　This doctoral research could be divided into two domains: (i) subsurface investigation methodology and (ii) VS of two subduction zones onshore and offshore, i.e., the Bengal Basin (BB), Bangladesh, and off Fukushima in the southern Japan Trench, Japan, respectively. The methodology involved analysis of the horizontal-to-vertical spectral ratio (HVSR) of Fourier amplitude spectra and the Rayleigh wave fundamental mode group velocity dispersion curve (DC) from ambient noise tomography (ANT) for obtaining one-dimensional (1-D) subsurface VS profiles. We used algorithms based on the DFA for forward modeling and inversion of the DC and the HVSR. For the offshore case, the latest algorithm, hvgeneralized, was used, which could add the effect of the water layer to the inversion process.
　The 1-D VS profile was estimated to have an intermediate depth (maximum ~2500 m) in the BB by noise HVSR (HVSRN) analysis (from 0.2-10 Hz) at 19 seismic stations. Nominal engineering bedrock depth was estimated from the profiles regionally in the BB for the first time, along with an estimation of the response spectrum for an example site. Next, shallow sedimentary to deep crustal imaging was performed using ANT from 4-26 s using 13 seismic stations in the BB. Joint inversion of the DC and HVSRN was performed to obtain a 1-D VS profile, and a 2-D cross-section was obtained by interpolating such several profiles. Thus, we present a new high-resolution velocity and crustal model of BB from this study that conforms well with the existing models and hypotheses.
　In the offshore instance, using 19 ocean bottom seismometer (OBS) data, we set the first application of the HVSRN (from 0.1-10 Hz) to the sea floor to generate near-surface to deep subsurface VS profiles with higher resolution than the previous studies. Moreover, unstable HVSRN <0.1 Hz at the sea bottom was explained by the power spectral density of SAN energy. Additionally, we evaluated the contribution of the water layer to an observed HVSRN by forward modeling with energy propagation in both cases with and without the layer. In a later project, being motivated for HVSR analysis <0.1 Hz and eventual deeper investigation, late seismic coda (with 200 s window length) of regional and teleseismic earthquake records at 3 broadband OBSs were utilized to obtain HVSR (EHVSR). The EHVSR curves showed stability up to 0.05 Hz. They were then inverted for VS profiles up to 10 km depth from the sea bottom. At one station, we also successfully modeled the shallow plate interface (at ~7 km depth). Thus, this study enabled lower-frequency seismic observation at the sea bottom without any complex mathematical operation. Furthermore, in order to satisfy the inversion and forward modeling of the EHVSR under the DFA, we studied the diffusion in seismic coda waves of 9 regional and teleseismic earthquakes recorded at 3 OBSs with a nearly L-shaped configuration. Herein, the stability of P-to S-wave energy ratio was used as an indicator of the diffused wavefield. Along with demonstrating diffusion in the seismic coda at the ocean floor, we also found a correlation between the evolution of the <0.1 Hz HVSR peak and the diffuse seismic wavefield.
　This doctoral research presents several new observations in the ocean bottom seismology field along with offering new VS models for the off Fukushima. This study also shows the potential of the HVSR method in modeling shallow to deep subsurface. A new crustal model has also been proposed in the onshore BB from this study. Moreover, this research paves the way for a comparison of the passive seismic methodologies, HVSR results onshore and offshore, and the crustal structure of two subduction zones. Hopefully, the insights from this research will be very helpful for future seismological studies.

---------- ---------- ---------- ----------