Magnetic field anomaly separation using empirical mode decomposition

The geophysicsl potential field separation refers to separation of the regional and local anomalies from the superimposed anomaly. The Empirical Mode Decomposition (EMD) proposed by Norden E Huang is a kind of spatial and temporal filtering process in terms of the signal extremum characteristic scales. It is a new data analysis method suitable for processing non-stationary and non-linear data. Its power to filter and decompose the data has earned it a high reputation in signal processing.
Empirical mode decomposition is a time-frequency analysis method which can adaptively decompose complex signals. The decomposed component contains diffrent bands of frequencies from high to low, and the residual value is the signal trend component representing the signal averaged trend, which is similar to the regional anomalies in the geophysical field. The empirical mode decomposition (EMD) method is an algorithm for the analysis of multicomponent signals that breaks them down into a number of amplitude and frequency modulated zero-mean signals, termed intrinsic mode functions(IMFs). An IMF must fulfill two requirements: (1) the number of extrema and the number of the zero crossings are either equal or differ at most by one; (2) at any point, the mean value of the envelope defined by the local maxima and the envelope defined by the local minima is zero. Based on this theory, applying the EMD to separate the geophysical potential field is proposed in this article. When EMD is used for anomaly separation, the problem is to identify properly which IMFs contain residual characteristics. Certain modes will consist of mainly residual, whereas other modes will contain regional and noise characteristics.
Magnetic field anomalies are usually superposed large-scale structures and small-scale structure anomalies. Separation of these two categories of anomalies is the most important step in the data interpretation. Different methods have been introduced for this work, but most of them are the semi-automatic methods; it means that the interpretator’s opinion can directly affect the results. In this study, EMD method has been used to separate regional and residual magnetic anomalies. EMD decomposition results in what is “residual”, which is similar to the regional anomaly of a potential field data. This residual does not require any preset parameters unlike contemporary field separation methods. This automatic method is based on the extraction of the intrinsic oscillatory modes of data. Efficiency of this method has been investigated on both synthetic and real data acquired in North Mahalat area of Markazi Province to study the regional subsurface geology with the purpose of geothermal reserver explorations. Compared to the conventional method of trend analysis, the EMD method is affected by less artificial influence, and we did not need to set any parameters beforehand. Otherwise, it reflected the potential field intrinsic physical characteristics better. Separation results showed that this technique had higher accuracy than conventional methods such as polynomial fitting and had a good consistency with regions geology. Finally, the results of the new method were compared with resultsof the upward continuation filter and we observed that these results were matched with the upward continuation filter.