Ray Reflection Tomography: A Comparison Between Layer-Stripping and Simultaneous Updating Methods

In this work we compare two traveltime inversion methods in terms of robustness and stability. The first one, which was formulated in [Soares, 1994], is a layer-stripping procedure where the geological model reconstruction is carried out layer to layer, from above downward. The second method consists of an iterative process where all the parameters are updated simultaneously. We show that for suitable medium parametrizationa, large offsets aquisitlons, low to moderate lateral velocity variation and constraints imposed in the Tikhonov's sense [Tikhonov, 1963], both methods work very well. We show that the reconstruction of a geological medium described by sixty-four B-splines coefficients is carried out in only three iterations. Moreover, both methods show the superiority of the interfaces reconstruction compared to the velocity resolution, as it was analytically demonstrated by Bube et al. for an horizontal layer [Bube et al., 1991). Nevertheless, although the second method reduces the cumulative error resulting from the errors in the upper layers retrieva.ls, the layer stripping method is more robust and stable.