A STUDY OF METHODS FOR TOMOGRAPHIC VELOCITY ESTIMATION IN THE PRESENCE OF LOW-VELOCITY ZONES

This paper deals with the problem of seismic velocity estimation from first-arrival traveltimes in a two-dimensional (2-D) cross-hole geometry where explosions are detonated in one borehole while recordings are made in another borehole and on the surface. Standard tomographic procedures are based on decomposition of the cross-hole area into a number of cells and a simplifying assumption of straight raypaths. In the presence of significant low-velocity zones, the resulting images may be contaminated. Different ways of performing tomographic inversion are tested on a number of synthetic examples. Images obtained by direct, unrestricted least-squares inversion are often seriously distorted. However, methods using more cells and some kind of damping often give more satisfactory results. Because the risk of distorted images is always present in inversion procedures, comparison with synthetic data (forward modeling) is a valuable tool in the interpretation process.