Structural joint inversion coupled with Euler deconvolution of isolated gravity and magnetic anomalies

Abstract

We generalized the Euler deconvolution method to a joint scheme, which consists of locating the horizontal and vertical positions of the top of potential-field 3D sources. These results were then used to constrain the depth to the top of the models obtained by cross-gradient joint 3D inversions, imposing fixed known values in the a priori models. The coupling of both methods produced more realistic density and magnetization models for separate and joint inversions, relative to those obtained by applying cross-gradient joint inversion only. This strategy was tested on a 3D synthetic experiment, and on a real field data set from the northwest region of the Baja California Peninsula, Mexico. After locating the vertical position of the source, the algorithm uses this information to obtain density and magnetization models that enhanced their structural compatibility and reduces the ambiguity on the interpretation of their structural characteristics laterally and at surface.

Generalized L$_p$-norm joint inversion of gravity and magnetic data using cross-gradient constraint

“…where ∇ indicates the gradient operator (Gallardo & Meju 2003) and structural similarity is achieved when t(m) = 0, see Appendix A for details. As noted already in Section 1, this corresponds to the case in which the gradient vectors are in the same or opposite direction, or, alternatively, one of them is zero (Gallardo & Meju 2003;Gallardo & Meju 2004;Tryggvason & Linde 2006;Gallardo 2007;Fregoso & Gallardo 2009;Haber & Gazit 2013;Fregoso et al 2015). From a geological viewpoint this means that if a boundary exists then it must be sensed by both methods in a common orientation regardless of how the amplitude of the physical property changes (Gallardo & Meju 2003).…”

“…Mathematically, this may be achieved by forcing the cross product of the gradient of the different model parameters to be zero everywhere (Gallardo & Meju 2003;Gallardo & Meju 2004;Tryggvason & Linde 2006). Indeed, many successful results for simultaneous joint inversion with the inclusion of the cross-gradient constraint have been reported, (Gallardo & Meju 2003;Gallardo & Meju 2004;Tryggvason & Linde 2006;Gallardo 2007;Fregoso & Gallardo 2009;Haber & Gazit 2013;Fregoso et al 2015;Gross 2019;Zhang & Wang 2019). On the other hand, Zhdanov et al (2012) observed that the Gramian constraint can be used to enhance the correlation between different physical properties and/or their attributes.…”

Ratio‐Euler deconvolution and its applications

“…Euler deconvolution is a popular method to automatically and quickly estimate the locations of sources from potential field data. The estimated structural index (SI) of the source from the Euler method is useful to determine an optimal depth‐weighting function rate decay for 3D inversion of potential data (Cella & Fedi, 2012; Vatile & Fedi, 2020), and the estimated top depths of source can be used to constrain a more realistic inversion model (Fregoso et al ., 2015). Euler deconvolution has been widely used to interpret potential field data in target detection (Davis et al ., 2010; Mu et al ., 2020), joint inversion (Fregoso et al ., 2015), the interpretation of linear structures, such as faults and geological boundaries (Wang et al ., 2017; Castro et al ., 2020; Njeudjang et al ., 2020; Kumar et al ., 2020; Nuñez Demarco et al ., 2020), and locating crustal anomalous bodies (Ebbing et al ., 2007; Kumar et al ., 2020).…”

A comparison of the joint and independent inversions for magnetic and gravity data over kimberlites in Botswana

“…In this case, a direct or indirect linkage between the multiple model parameters is imposed (Leliévre et al., 2012), and the joint inversion restricts the models to only those that predict the observed datasets and satisfy the linkage constraint. There have been many successful applications of joint inversion of multiple geophysical datasets (Fregoso & Gallardo, 2009; Fregoso et al., 2015; Gallardo & Meju, 2003, 2004; Haber & Oldenburg, 1997; Haber & Holtzman Gazit, 2013; Lin & Zhdanov, 2018; Leliévre et al., 2012; Moorkamp et al., 2011; Nielsen & Jacobsen, 2000; Tryggvason & Linde, 2006; Zhdanov et al., 2012). In particular, when different datasets are sensitive to variations within specific sections of the subsurface target(s), a joint inversion process may be able to reduce, and sometimes significantly, the ambiguity of the reconstructed model.…”