Japan Geoscience Union Meeting 2023

Presentation information

[E] Online Poster

S (Solid Earth Sciences ) » S-EM Earth's Electromagnetism

[S-EM14] Electric, magnetic and electromagnetic survey technologies and scientific achievements

Wed. May 24, 2023 1:45 PM - 3:15 PM Online Poster Zoom Room (4) (Online Poster)

convener:Kiyoshi Baba(Earthquake Research Institute, The University of Tokyo), Tada-nori Goto(Graduate School of Science, University of Hyogo), Yuguo Li(Ocean University of China), Wiebke Heise(GNS Science, PO Box 30368, Lower Hutt, New Zealand)

On-site poster schedule(2023/5/23 17:15-18:45)

1:45 PM - 3:15 PM

[SEM14-P05] Attempt of signal separation of magnetotelluric data based on singular value decomposition

*Rei Amano1, Tada-nori Goto1, Ryokei Yoshimura2 (1.University of Hyogo, 2.Disaster Prevention Research Institute , Kyoto University)

Keywords:singular value decomposition, signal separation, magnetotelluric

The magnetotelluric (MT) method is one of electromagnetic survey methods for estimation of deep subsurface structures by observing natural fluctuations of geomagnetic and electric field. However, it has not been fully discussed how the ratio of noise in the electromagnetic field data varies over time and space depending on the observation point, a period of time, and day of observation when the MT method is performed.
In this study, we applied singular value decomposition that is the matrix decomposition methods to the spectra of five observations obtained from MT surveys in the Noto Peninsula where cluster earthquakes have been occurring since 2020 to separate electromagnetic field variations into many principal components.
By sorting out which principal components are geomagnetic fluctuation components or artificial noise, the amount of noise for each observation condition was estimated. Separately, the errors of the MT response function were obtained and compared to the noise.
As a result, a correlation was observed between the amount of noise calculated by singular value decomposition and the error of the MT response function, indicating the usefulness of using singular value decomposition to estimate the amount of noise. We were also able to capture the characteri