Detection of the horizontal boundary of gravity anomalies using the hybrid positive and negative curvature (PNH) procedure

Determining the edge and horizontal position of geologic structures is one of the fundamental steps in interpreting potential field data. Several filters have been introduced that use the concept of curvature to determine the edge of potential field data. However, these filters have advantages and disadvantages in detecting causative sources. Therefore, it seems necessary to introduce more efficient approaches. In this work, the most positive and most negative curvatures of gravity field data were analyzed, and a more efficient filter was introduced and applied that uses the concept of curvature and its combination to delineate the edges of geological structures and buried sources. The proposed method, called the hybrid positive and negative curvature (PNH) approach, combines the most positive and most negative curvatures into one curvature by fitting the formula and weighted summation. The proposed strategy takes advantage of both positive and negative curvatures to improve the edge detection of gravity field data. To this end, the performance of the PNH procedure was investigated considering different density assumptions (positive, negative, and positive-negative) for the relatively imposed synthetic gravity model resulting from buried prisms. The results obtained on synthetic models with and without noise show that the PNH procedure can detect the horizontal boundaries of buried structures relatively well. Of course, due to the use of directional derivatives in the filter of the hybrid positive and negative curvature approach, it seems very necessary to use noise-reducing filters before applying edge detection methods. Moreover, conventional filters such as the second vertical derivative (SVD) and the tilt angle (TDR) were used to compare the performance of the hybrid positive and negative curvature filter on the synthetic model. However, the obtained results show that the second vertical derivative and the tilt angle do not have the required capability to determine the edge of the synthetic model. In the following, the quality of the most positive and most negative curvatures filter and the hybrid positive and negative curvature were investigated using real data from a gold mine in the Witwatersrand area (South Africa) and also gravity data from the Aji-chai salt dome, East Azerbaijan province (Iran) and then using WGM-2012 derived gravity data belonging to the Marian trench area. Due to the sensitivity of the filters to noise, the upward continuation filter was applied before determining the edge of the buried structures. The edge maps from the Witwatersrand area and the data from the Aji-chai salt dome obtained using the hybrid positive and negative curvature determination method, demonstrate acceptable accuracy of this filter in determining the edge and representing the horizontal position of various geological structures. By using the PNH filter, the lateral boundaries of the main structures and other subsurface sources are well detected. Of course, due to the noise sensitivity of this filter, which is due to the use of second-order gravity derivatives, good quality data without noise must be used. Therefore, it is suggested that noise attenuate filters, such as upward continuation method, must be used prior to creating the maps to determine the edge. Therefore, the PNH edge detection method can be reliably used for qualitative interpretation of gravity field data.