Spectrum Analysis of Magnetotelluric Data Series Based on EMD-Teager Transform

Abstract

In order to meet the non-stationary characteristics of Magnetotelluric (MT) data series, a new method, which is called EMD-Teager, was proposed based on empirical mode decomposition (EMD) and Teager operator for the first time. With the EMD-Teager method, the amplitude of data series is expressed as a function of frequency and time, and then margin spectrum are calculated statistically from the time–frequency spectrum. This paper focuses on two aspects which are facilitated by applying the EMD-Teager. The first aspect is the time–frequency characteristics analysis of MT signal and using the time–frequency spectrum to do pretreatment for MT signal through piece-wise stationary analysis and MT data subsets selection; the other concerns the calculation of the Teager margin spectrum from the time–frequency spectrum. The conclusion reached through discussion of the first aspect is that EMD-Teager has a strong ability to describe the time–frequency characteristic of MT signal. Using the Teager spectrum, the better data section can be selected and the reliability of geological data is improved greatly. The conclusion draws from the second aspect is that Teager marginal spectrum, which is coming from the integral of Teager time–frequency spectrum along the time axis, overcomes the drawbacks of Fourier methods and minimizes the estimation bias brought about by the non-stationarity feature of MT signal. Therefore, EMD-Teager method is effective in analyzing the time–frequency characteristics and marginal spectrum estimation of MT signals, and it will have a wide application in processing of MT data.