Joint inversion of seismic slopes, traveltimes and gravity anomaly data based on structural similarity

Abstract

SUMMARY Attention is paid to joint inversion of multiple geophysical data because of its advantages on weakening the non-uniqueness of inversion and further enhancing comprehensive interpretation. Due to the good correlation between rock velocity and density, seismic and gravity data have been widely used in joint inversion. However, the joint inversion of pre-stack seismic reflection and gravity data remains underdeveloped at the exploration scale. Without a quantitive relation between velocity and density, we develop a structure-based joint inversion using seismic reflection traveltimes, slopes and Bouguer gravity anomaly data simultaneously for building both velocity and density models. In our method, cubic B-spline interpolation is used to parametrize the common knots of velocity and density models. Incorporating seismic slopes into the joint inversion framework, we build a composite objective function which minimizes the weighted-sum of seismic/gravity data misfits, regularization and structural constraint terms. By subdividing the knot spacing, a multiscale strategy is alternative to increase the stability of inversion. First, we describe the methodology, followed by three synthetic examples to illustrate the feasibility and benefits of the method. Examination of the convergence curves via inversion suggests that the desired solution is more likely to be obtained with gentle convergence of each term, thus it can be used as an indicator for weight adjustment. Additionally, locations of scattering points and acoustic impedance can be obtained as by-products. Compared with the inversion of the respective data, the joint inversion exhibits the complementary characteristics of seismic and gravity data, improves the distribution and structural features of the resulting physical properties, especially in deep and complex tectonic situations.