Nonlinear Inversion of Full Tensor Gravity Gradiometry Data for Three-Dimensional Basement Modelling in the Gongola Basin, North East Nigeria

Full tensor gravity gradiometry provides a richer dataset for basement modelling than conventional scalar gravity measurements, enabling the resolution of structural details at scales relevant to petroleum trap and basement resource identification, yet the nonlinear inversion of FTG data for three-dimensional density distribution remains computationally challenging and methodologically underdeveloped for the geological contexts of Nigerian sedimentary basins. This research formulates a nonlinear inversion algorithm for full tensor gravity gradiometry data that incorporates geological constraints relevant to basement-sediment systems in the Gongola Basin of North East Nigeria. The algorithm implements a gradient-based nonlinear optimisation using a trust-region reflective method with adaptive regularisation, incorporating a geological compactness constraint that favours sharp density contrasts at the basement-sediment interface consistent with known geological contacts. A comprehensive synthetic benchmark study using 200 randomly generated basin models demonstrates that the FTG inversion recovers basement topography with a mean depth error of less than 3 percent, superior to both the conventional Euler deconvolution method (12 percent error) and to inversion of the vertical gravity component alone (7 percent error). Application of the algorithm to a 12,000-square-kilometre FTG dataset acquired over the northern Gongola Basin produces a three-dimensional basement topography model at 500-metre grid resolution. The model reveals a series of NE-SW trending grabens with maximum basement depths of 4.2 kilometres, hosting potential source and reservoir sequences. Keywords: full tensor gradiometry, nonlinear inversion, basement modelling, Gongola Basin, density