We performed these computations for two frequency ranges, a low-frequency (10-60 Hz) band, within which we expected pump noise to dominate, and a higher-frequency band (40-120 Hz) in which we hoped that pump noise wouldn't dominate and we would see energy from other sources, perhaps energy generated at the depth of the reservoir by the extraction of oil.
Figures 6 and 8 show the results of the semblance computations for the low-frequency range for two different records. The different panels show the semblance as a function of x and y for various depths. The limits of each panel are the boundaries of the array. The main feature of these plots is high semblance values at shallow depths. These values correspond to pump noise.
Figures 7 and 9 show the maximum value of semblance (for all x and y) as a function of depth. These plots confirm the conclusion that the main sources of energy in the low frequency range are on the surface.
The results are quite different for the high-frequency range, however. Figures 10 and 12 show the semblance results for the two records. For depths less than approximately 350 meters, the patterns are rather random, but become more organized as depth increases. The graphs of maximum semblance versus depth (Figures 11 and 13) show that semblance increases at a depth of around 700 meters for record 1 and 400-600 meters for record 2. This may be evidence for sources of energy other than pump noise occurring at depth. Perhaps it is energy generated by the reservoir during the process of oil extraction.