Integrated gravity and wide-angle seismic inversion for two-dimensional crustal modelling

An inversion scheme for integrated wide-angle seismic and gravity modelling of 2-D crustal structure is presented. The inversion scheme uses a standard least-squares, linearized iterative approach, which relies on forward algorithms for calculating gravity responses and seismic traveltimes. The presented scheme has been integrated in an existing and widely used program package. Wide-angle seismic data have the potential to provide detailed knowledge about structures in the crust and upper mantle. The corresponding density distribution may, however, provide essential geodynamic information in relation to isostasy, which seismic data cannot provide. In order to describe the interrelation between velocity and density for crustal rock types, numerous empirical relations linking P-wave velocity and density have been estimated over the last 40 years. These experiments have been based on both laboratory experiments and well log data. We use such relations as prior constraints on how velocities and densities should be linked during the joint inversion. These constraints may be narrow in regions of the model where the chosen velocity-density relation is expected to be trustworthy, and broader in parts of the model where the velocity-density relation is believed to be less certain. The velocity-density model of the subsurface is described by a layered, blocky structure defined by velocity, density and depth nodes. During inversion, all three parameter types may vary. The inverse scheme is tested in three synthetic examples, which resemble important problems likely to occur in crustal modelling. In particular, we focus on how the inverse estimates are influenced by the interplay between seismic data and gravity data, and the uncertainties of the relation between velocity and density. Important advantages of the least-squares inverse modelling scheme, as opposed to manual trial-and-error modelling, are that data are assured to be fitted according to a pre-specified norm, the automated interpretation reduces the time used for modelling, and a more exhaustive analysis of the gravimetric and seismic data sets can be made.