Curvature analysis to differentiate magnetic sources for geologic mapping
M. Lee, W. Morris, G. Leblanc and J. Harris
Journal name: Geophysical Prospecting
Issue: Vol 61, No sup1, June 2013 pp. 572 - 587
Info: Article, PDF ( 1.42Mb )
Curvature of a surface is typically applied in seismic data interpretation; however this work outlines its application to a potential field, specifically aeromagnetic data. The curvature of a magnetic grid (from point data) is calculated by fitting a quadratic surface within a moving window at each grid node. The overall and directional curvatures calculated within this window provide insight into the geometry of the magnetic grid surface and causative sources. Curvature analysis is an in-depth study of both qualitative (graphically) and quantitative (statistically) approaches. This analysis involved the calculation of full, profile and plan curvatures. The magnitude, sign and relative ratios enable the user to define source location and geometry and also discriminate source type; for example, differentiation between a fault and normal polarity dyke. The reliability of the analysis is refined when a priori geological knowledge is available and basic statistics are considered. By allotting a weighting scheme to various statistical populations (e.g., standard deviation), increased detail is extracted on the different lithologies and structures represented by the data set. Furthermore, the curvature’s behaviour is analogous to derivative calculation (vertical, horizontal and tilt) by producing a zero value at the source edge and either a local maxima or minima over the source. Application prior to semi-automated methods may help identify correct indices necessary for identification of magnetic sources. Curvature analysis is successfully applied to an aeromagnetic data set over the 2.6–1.85 Ga PaleoproterozoicWopmay orogen, Northwest Territories, Canada. This area has undergone regional and local-scale faulting and is host to multiple generations of dyke swarms. As the area has been extensively mapped, this data set proved to be an ideal test site.