Improved total magnetization direction determination by correlation of the normalized source strength derivative and the reduced-to-pole fields

Abstract

The knowledge of total magnetization (magnitude and direction) makes it easier to interpret magnetic anomalies. We have developed a simple crosscorrelation-based method to determine the total magnetization direction of a magnetic source from the vertical derivative of normalized source strength (dNSS) and the reduced-to-pole (RTP) magnetic fields. For most source types, the spread of the dNSS field (or its half-width) is similar to that of the RTP field computed with the correct total magnetization direction, and, thus, the comparison results in a more meaningful correlation coefficient than other functions used in the literature. We have determined the utility of our method using several compact source types (i.e., sphere, dike, horizontal sheet, vertical and horizontal cylinders, and prism). Moreover, the existing methods for determining the direction can be unstable at low latitudes due to noise amplification. A filter that isolates the main features of the anomaly of interest, when applied to both the fields being correlated, improves the performance of the method. We also implement a stabilizing amplitude threshold filter that made the method stable at low latitudes. Model tests indicate that our method estimates the total magnetization directions accurately for low inclinations of total magnetization and inducing field directions. We applied the method to estimate the total magnetization direction of magnetic anomalies in the north and central part of the Goiás Alkaline Province in central Brazil. The RTP fields from the total magnetization directions derived from our method meet the expectations of anomaly symmetry and centering on the outcrops or the edges of the alkaline intrusive bodies. In addition, we found that the resulting magnetic and gravity models of the Goiás Alkaline intrusives were consistent with the geologic model of inverted conical diatremes.