The Generation of Near-Surface-Layer Models From Seismic Reflection Traveltimes

Abstract

SUMMARY Exploration seismology reflection ray paths sample the near-surface layer near both shot and receiver locations and are consequently affected by its velocity and thickness variations. The effect of the near-surface layer is normally considered to be consistent at each shot or geophone station for all traveltimes arriving at that location (the surface-consistent approximation). This assumption allows timeshifts to be calculated and the traveltimes corrected to a chosen datum, giving static corrections. The single correction at each point is an averaged correction, so the assumption is particularly inaccurate in the event of lateral variations of velocity or thickness of the surface layer, in the presence of large surface-layer velocities and in the presence of a thick surface layer. In this paper, we calculate the velocity or thickness of the near-surface layer from 2-D reflection data by minimizing, with respect to a model of the surface layer, the difference between data predicted by the model and the actual data. The non-linear relationship between data and model must be considered explicitly, because linearization results in a surface-consistent approach. Once a model has been produced, corrections that are dependent on the ray-path geometry through the near-surface layer can be calculated. The method is tested on synthetic data and applied to a reverse vertical seismic profile (RVSP) data set. The extension of the method to 3-D data is also considered.