Separation of induced and remanent magnetization by using statistical susceptibility data

Abstract

Geological units typically contain both induced and remanent magnetization. The vector sumof the two components forms the total magnetization to which magnetic data is sensitive. The induced component depends on the susceptibility and current geomagnetic field, whereas the remanent magnetization depends upon the formation history of the geological units. Remanent magnetization, therefore, carries information about the formation process and subsequent structural movement of these units. For this reason, identifying remanent magnetization from geophysical magnetic data could yield additional information which would otherwise not be available for geological interpretation. I investigate an approach for separating remanent magnetization from the induced component in the total magnetization recovered from a 3D magnetic inversion. Assuming the known total magnetization, I allowed for a range of susceptibility values, and I tested the effects of the assumed probability distribution of the susceptibility onto the resulting distributions of the observed remanent magnetization. Given the importance of remanence in geophysical interpretation, I further examine the direction of the extracted remanent magnetization. This was accomplished by using Fisher statistics and the Parzen windows method to estimate the probability density distribution of the remanent magnetization direction. I build two synthetic examples that have two source bodies that went through a structural deformation and rotation to test our analyses. The two bodies are offset and rotated relative to each other, so the rotated object’s remanent magnetization will be different even though induced magnetization stays the same for each body. By using remanent magnetization extracted from the inverted total magnetization, I apply the methodology we developed for each of the blocks separately after taking the average to decrease the directional variability within the blocks. As a result, I estimate a remanent magnetization average direction for the unrotated and rotated objects in the model in order to recover the two units’ relative rotation angle. This study offers a new approach to understanding how to extract and quantify the remanent magnetization magnitudes and directions for revealing the rotation angle where the deformation is present in the study area. Having only statistical susceptibility information available from petrophysical data, I present a preferable alternative in real-world problems with a reliable angle recovery.