Claerbout (2000) proposes a test case for which the Gaussian curvature
of the model vanishes. In this paper, we present an even simpler test case. Given a
2-D random field, we deconvolve with a known dip (or steering) Clapp et al. (1997)
filter to obtain a ``plane wave'' model, as shown in Figure 1.
To simulate collected ``data'', we sampled the model of Figure 1 at
about 60 points randomly, and about two-thirds of the way down one trace in the center.
The results are shown in Figure 2.
Figure 1 True model - plane waves dipping at .
Figure 2 Left: Collected data - one known trace, about 60
randomly-selected known data points. Right: Known data selector.