Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Makuuchi, Ayumu; Asamori, Koichi; Negi, Tateyuki*
no journal, ,
Even though applying the far remote reference magnetotelluric (MT) method, we need long recording period to obtain usable data from the contaminated data by strong and coherent noise in DC railway area. In this study, we consider the electric time series model including a trend component, natural magnetic signal response, correlated noise components, and white noise, then attempt to separate to each component with a Kalman filter algorithm. The method was applied to the magnetotelluric data observed near the DC railway and seems to work well in the time domain.
Hama, Yuki*; Makuuchi, Ayumu; Negi, Tateyuki*; Asamori, Koichi
no journal, ,
The remote reference technique is widely used in the magnetotelluric method to decrease local noise by acquiring data simultaneously at a remote site. This technique can reduce the local noise component and extract the magnetotelluric signal by stacking the cross power spectrum between the remote data and field data. Remote reference is especially efficient when applying ideal remote data with no noise component. Yet it is usually difficult in practical surveys. The best way to compensate in such a difficulty is to select multiple remote sites. With multiple remote sites we can expect the same effect obtainable from an ideal remote site without noise. In this paper we applied multiple remote reference method using robust weighted stacking to noisy survey data, in order to redact the noise component included in the remote site which causes the escalation of data scattering.