The velocity models are depicted in the same color scale in the upper panels of Figures (``WEMVA works'') and (``WEMVA breaks''). The background velocity is 2000 m/s for both. The peak anomalies, from left to right, as departures from the background, are (in m/s): -153, -188, +231 for the ``WEMVA works'' velocity model and -586, -766, +519 for the ``WEMVA breaks'' velocity model.
The ``WEMVA works'' velocity anomalies cause visible amplitude focusing, but minimal traveltime anomalies (middle panel in Figure ). The intense ``WEMVA breaks'' velocity lenses, on the other hand, cause visible triplications and departures from hyperbolicity (middle panel in Figure ). The specific signature of FEAVO before migration can be examined in both cases by squaring each of the values in the data cube, then summing along the time axis. The results of this operation is shown in the lower panels of Figures and respectively. The axes of the image are midpoint and offset, and specific ``V'' shapes are visible for each of the three velocity anomalies.
It is interesting to compare these FEAVO effects with those present in the only field dataset currently available to us that exhibits FEAVO anomalies. The dataset in question is the Grand Isle dataset analyzed by Kjartansson (1979) - the original dataset on which FEAVO was defined. The FEAVO effects in it are comparable with those in the lower panels of Figures and because the background velocities are approximately similar. Given that the intensity of the effects is linked to the magnitude of the velocity anomalies, we can empirically estimate that the deviation of the Grand Isle velocity anomalies from the background is around 200-300 m/s. Since the ``thickness'' of the V's is linked to the dimensions of the velocity lenses, we can also infer that they are approximately 20-30m in diameter.