In this section, I perform velocity analyses at the surface locations of the drill rig, the steam injection well, and three pumps. This is done for two different data records - one where the pumps were operating, and one where they have been temporarily turned off.
Figure shows the result of velocity analysis at the drill rig location. Again the largest semblance values are shown in white. The top 1% of semblance values are clipped by the plot program and shown in black. As in the case where a constant velocity was used to compute semblance in 3-D, here we get a clear picture of what I believe to be direct arrivals from the drill bit. The horizontal line on the plots indicate the depth of the drill bit. This line intersects the main feature of the plots (for the records from the ``quiet period'') near the constant velocity of 10000 feet/second that we found in the previous section to do a good job of stacking. As in the previous section, when the pumps are operating there is no good evidence of drill bit signal.
Figures and show the result for the locations of the steam injection well and the pumps. Figure is for a record where the pumps are operating, and Figure is for a record where the pumps have been turned off. Both sets of plots have been scaled uniformly. We see that there is little indication of coherent sources except for very shallow depths. This is somewhat surprising, since we might expect both sources to manifest themselves near the reservoir depth of around 2700 feet. Some of the strong semblance values at very shallow depths disappear when the pumps are turned off. Beyond that there is little of interest in these results. The drill bit seems to be a much stronger source than any of the pumps or the steam injector.
Semblance as a function of velocity and depth for a fixed surface location, the point above the drill bit. The depth of the drill bit is shown on each plot by the horizontal line. Larger semblance values are shaded white, with the largest 1% clipped and shown in black. The top plots are for records where pumps within the array have been turned off. In this case the direct arrivals from the drill bit are easily seen. Pumps were still operating during the recording of the two records shown at bottom. The result is that there is no clear drill-bit signal.
Semblance as a function of velocity and depth for four fixed surface locations. Top left, for the location of a steam injection well. The other three plots are for the locations of pumps. In this record pumps were operating. Larger semblances are shaded white, and the largest 1% are clipped and shown in black.
Semblance as a function of velocity and depth for four fixed surface locations. Top left, for the location of a steam injection well. The other three plots are for the locations of pumps. In this record, the pumps are not operating, and some of the stronger semblance values, at shallow depths, have disappeared.
The dominant features in most plots dip toward lower velocities at greater depths. This feature can easily be explained. Energy from a source at great depth will exhibit very little moveout across the array. The same moveout can be expected from shallower sources only if the medium velocity is correspondingly higher. There is a tradeoff between depth and stacking velocity, and this tradeoff explains the dipping features. This tradeoff is a necessary consequence of the fact that our source is operating continuously over time rather than being impulsive.