2.5D Regularized Gravity Data Inversion for the Detection of Faults in Basement Rocks

• Published in: Pure and applied geophysics Vol. 180; no. 9; pp. 3319 - 3338
• Main Authors: Feng, Xuliang, Yang, Liu, Ma, Jiayue, Wu, Chuanbo, Liu, Kaixuan
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2023; Springer Nature B.V
• Subjects: Basement rock; Basements; Computation; Depth; Detection; Direction; Earth and Environmental Science; Earth Sciences; Fault detection; Fault lines; Faults; Geological faults; Geophysics/Geodesy; Gravity anomalies; Gravity data; Inversion; Inversions; Methods; Parallel processing; Regularization; Three dimensional models; regularized inversion; Fault; edge detection; gravity anomaly
• ISSN: 0033-4553; 1420-9136
• DOI: 10.1007/s00024-023-03326-7

We have developed a basement fault detection technique (the so-called NVDTH) enhanced by regularized inversion of gravity data to improve the detectability of deep small-scale faults. The basement depth can be obtained by two-and-a-half-dimensional (2.5D) regularization inversion assuming that the density contrast is known, and the area of significant variation in basement depth is extracted to get the planar position of the fault in the target region. Equidistant gravity profiles are extracted from measured areal gravity anomalies along x - and y -directions, and the 2.5D regularized inversion results of each profile are merged to form the basement depths BSAX (the inverted basement depth along x direction) and BSAY (the inverted basement depth along y direction) for the entire region. The total horizontal gradient (THG) technique is used to detect fault locations from both BSAX and BSAY. The vertical derivative technique is applied to the THG, and then the negative values are returned to zero to finally obtain the NVDTH whose maximum value can characterize the planar position of faults. We use OpenMP for parallel computation, which can use all of the CPU cores to perform multiple profile inversions at the same time, greatly reducing time consumption. The inversion is not accurate when we focus on the basement depth because 2.5D profiles are used to simulate the 3D model. However, the main aim is to improve the capacity to detect the planar position of basement faults, and the inverted basement depth is just an intermediary processing datum whose inaccuracy has no bearing on fault identification. Numerical examples resembling a two-dimensional (2D) rifted basin and a three-dimensional (3D) step-faulted basin and the real data from the Weihe Basin show that our proposed method is more accurate in identifying basin basement faults than some edge identification techniques that act directly on gravity anomalies.