13:45 〜 15:15
[MIS04-P08] A development of noise reduction method for ULF band electromagnetic data using Multi-channel Singular Spectrum Analysis (MSSA) (2)
キーワード：MT法、MSSA（マルチチャンネル特異スペクトル解析）、信号弁別、房総半島、ULF帯
An electromagnetic phenomenon associated with crustal activity in the ULF band (f < 10 Hz) is expected as a method that can detect the signal directly from the intra-crust or plate boundary. However, the observed signal includes not only the earthquake-related signal but also the solar-origin signal and artificial noise, such as the leak current from the DC-driven train. Detection of the signal is difficult because the amplitude of the earthquake-related signal is tiny compared with the other noises, and the waveform of the signal is similar to the artificial noise. Recent analyses were performed in a noiseless environment, such as the region far from the city area, and at night-time to escape this problem. However, by this scheme, we miss detecting the signal that occurs in the daytime. The new signal discrimination method is essential to increase the chance of detecting the signal in the daytime. We are developing a new signal discrimination method based on Multi-channel Singular Spectrum Analysis (MSSA) to overcome this problem. MSSA is the extension of Singular Spectrum Analysis (SSA) that is applied to the Boso electromagnetic data.
MSSA can decompose from C-ch time series to WL×C principal components (PCs) in descending order of large amplitude. The summary of the scheme is as follows: Tapering the observed time series with window length WL, then creating a trajectory matrix with shifting 1 data point. Performing Singular Value Decomposition (SVD) to the covariance matrix of the trajectory matrix, then reconstruct PCs from the eigenvalue and eigenvector of SVD. By choosing the PCs based on the property, such as contribution and correlation coefficient, and summating the selected PCs, we can divide the observed time series into signal and noise components. Compared with other methods, such as wavelet transforms, it is suitable for processing non-stationary signals, such as natural electromagnetic fields, because it does not require basis functions.
In the MSSA-based signal discrimination method developing in this study, remove solar-origin signals from 7ch data (target site 5 components (horizontal magnetic field 2ch (Hx and Hy), vertical magnetic field 1ch (Hz), and horizontal electric field 2ch (Ex and Ey)) and reference site 2ch (magnetic field 2ch (Rx and Ry))), then, obtain a local signal.
The result of applying this method to actual data acquired on the Boso Peninsula for signals with a period of less than 400 s, the solar-origin signal was removed, and a local signal of about 1 nT for the magnetic field and about 2 mV/km for the electric field was detected. This presentation will focus on the performance and effectiveness of the method.
MSSA can decompose from C-ch time series to WL×C principal components (PCs) in descending order of large amplitude. The summary of the scheme is as follows: Tapering the observed time series with window length WL, then creating a trajectory matrix with shifting 1 data point. Performing Singular Value Decomposition (SVD) to the covariance matrix of the trajectory matrix, then reconstruct PCs from the eigenvalue and eigenvector of SVD. By choosing the PCs based on the property, such as contribution and correlation coefficient, and summating the selected PCs, we can divide the observed time series into signal and noise components. Compared with other methods, such as wavelet transforms, it is suitable for processing non-stationary signals, such as natural electromagnetic fields, because it does not require basis functions.
In the MSSA-based signal discrimination method developing in this study, remove solar-origin signals from 7ch data (target site 5 components (horizontal magnetic field 2ch (Hx and Hy), vertical magnetic field 1ch (Hz), and horizontal electric field 2ch (Ex and Ey)) and reference site 2ch (magnetic field 2ch (Rx and Ry))), then, obtain a local signal.
The result of applying this method to actual data acquired on the Boso Peninsula for signals with a period of less than 400 s, the solar-origin signal was removed, and a local signal of about 1 nT for the magnetic field and about 2 mV/km for the electric field was detected. This presentation will focus on the performance and effectiveness of the method.