TY - JOUR
T1 - Locating Spatial Changes of Seismic Scattering Property by Sparse Modeling of Seismic Ambient Noise Cross-Correlation Functions
T2 - Application to the 2008 Iwate-Miyagi Nairiku (Mw 6.9), Japan, Earthquake
AU - Hirose, Takashi
AU - Nakahara, Hisashi
AU - Nishimura, Takeshi
AU - Campillo, Michel
N1 - Funding Information:
We thank the Editor, Yehuda Ben‐Zion, and two anonymous reviewers for constructive comments. T. H. was partially supported by the International Joint Graduate Program in Earth and Environmental Science (GP‐EES) of Tohoku University and the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant 18J10232. We also acknowledge the support from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (Grant Agreement 742335, F‐IMAGE). Further support was received from JSPS KAKENHI Grant JP6K05528 and also from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan under its Earthquake and Volcano Hazards Observation and Research Program.
Publisher Copyright:
©2020. American Geophysical Union. All Rights Reserved.
PY - 2020/6/1
Y1 - 2020/6/1
N2 - Locating change regions of seismic velocities and seismic scattering properties associated with volcanic activities and earthquakes is important for structural monitoring. To increase such applications, we propose to use sparse modeling to estimate spatial distributions of seismic scattering property changes. The sparse modeling is an inversion technique that enables us to estimate model parameters from a small data set with sparsity condition such as ℓ1 norm regularization. We apply this technique to seismic ambient noise cross-correlation functions from 17 Hi-net stations around the epicenter of the 2008 Iwate-Miyagi Nairiku, Japan, earthquake (Mw=6.9). We compute waveform decoherences at the 0.5–1 Hz band and invert the waveform decoherences for the spatial distributions of seismic scattering property changes. Just after the main shock, the largest change occurred at the south of the epicenter, and the maximum change of the scattering coefficient in this region is estimated to be 0.032 km−1. The result from an ordinary linear least squares inversion with the ℓ2 norm regularization is almost consistent with that from the sparse modeling. Moreover, we confirm the superiority of sparse modeling in imaging with smaller data sets. Only five seismic stations that are deployed near the epicenter so as to surround the change regions are necessary to retrieve the result from 17 stations. On the other hand, in the case of the ℓ2 norm regularization, we need at least 15 stations. The sparse modeling will be helpful to estimate the spatial distribution of seismic scattering property changes from a small data set.
AB - Locating change regions of seismic velocities and seismic scattering properties associated with volcanic activities and earthquakes is important for structural monitoring. To increase such applications, we propose to use sparse modeling to estimate spatial distributions of seismic scattering property changes. The sparse modeling is an inversion technique that enables us to estimate model parameters from a small data set with sparsity condition such as ℓ1 norm regularization. We apply this technique to seismic ambient noise cross-correlation functions from 17 Hi-net stations around the epicenter of the 2008 Iwate-Miyagi Nairiku, Japan, earthquake (Mw=6.9). We compute waveform decoherences at the 0.5–1 Hz band and invert the waveform decoherences for the spatial distributions of seismic scattering property changes. Just after the main shock, the largest change occurred at the south of the epicenter, and the maximum change of the scattering coefficient in this region is estimated to be 0.032 km−1. The result from an ordinary linear least squares inversion with the ℓ2 norm regularization is almost consistent with that from the sparse modeling. Moreover, we confirm the superiority of sparse modeling in imaging with smaller data sets. Only five seismic stations that are deployed near the epicenter so as to surround the change regions are necessary to retrieve the result from 17 stations. On the other hand, in the case of the ℓ2 norm regularization, we need at least 15 stations. The sparse modeling will be helpful to estimate the spatial distribution of seismic scattering property changes from a small data set.
UR - http://www.scopus.com/inward/record.url?scp=85086991973&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85086991973&partnerID=8YFLogxK
U2 - 10.1029/2019JB019307
DO - 10.1029/2019JB019307
M3 - Article
AN - SCOPUS:85086991973
VL - 125
JO - Journal of Geophysical Research: Solid Earth
JF - Journal of Geophysical Research: Solid Earth
SN - 2169-9313
IS - 6
M1 - e2019JB019307
ER -