Spectral analysis of a ghost

Nasser S. Hamarbitan, Gary F. Margrave

The vertical component of the Blackfoot high-resolution of 3C-2D data from two different source patterns was processed and two different stacks were produced. One source pattern consisted of 4 kg of explosives in a single 18 m hole. The second pattern consisted of 2 kg of explosives in each of two 9 m holes. A dramatic difference between the migrated stacks of the two datasets was observed. An f-x analysis was run to display the amplitude and phase spectra of the unmigrated datasets. Similar analysis was run on representative shot gathers from the two datasets. The analysis shows that the 9 m data has a higher signal band than the 18 m data has. A spectral notch due to a surface source ghost reflection is suggested as the main cause of the lower signal band in the 18 m dataset. A frequency of 55 HZ for the spectral notch was interpreted from various spectra. Further processing and analysis were run to confirm the estimated frequency. Due to the spectral notch in the 18 m data, better resolution was observed in the 9 m stacks than in the 18 m stacks. A theoretical ghost operator which installs a notch at the mentioned frequency was convolved with a 9 m migrated trace and compared to corresponding migrated traces from the 18 m data. A match between the resulting traces was observed.