Introduction

In previous reports (Etgen, 1989, 1988), I detailed a method for estimating interval velocities using prestack depth migration and residual prestack migration. Amoco and Seismograph Service Limited (SSL) have provided me with a dataset to test this velocity estimation method. The dataset is from the North Sea and has a salt dome and other structure that causes lateral velocity variation. I wanted a dataset that had significant lateral velocity variation, but one where the lateral variation wasn't so severe that it was difficult to get any image at all or that I had to know anything about local geology or needed well logs to solve the problem. Furthermore, I wanted a dataset that wasn't dominated by deconvolution or noise problems. The dataset I received appears to meet these requirements.

This paper describes the data and my initial attempt at deriving an interval velocity model that gives a good prestack depth migration of the data. After migration with a velocity model v(z), the data had residual moveout indicating errors in the velocity model. Tomographic back-projection of those errors using the method I presented in previous reports gives an update to the velocity model in the form of a change in interval slowness.