Sand thickness estimation using spectral decomposition

Abstract

Spectral decomposition of seismic data has been used to identify thin beds and other possible changes subsurface sand layers. Spectral decomposition can be used as a tool for estimating bed thickness. This thesis uses a 3-D Seismic data set in the Shelf Gulf of Mexico. Two Sands were identified and designated , sands A and C in a traditional 3-D seismic survey with the use of nine well logs. For each sand the frequency spectrum is extracted and the peak frequency found. Peak frequency is the frequency that has the highest amplitude value in the spectrum. The peak frequency at each location is then plotted against the thickness of the sand. The peak frequency response of the data is compared with a modeled case. Sand A shows some correlation to the modeled case; however , bimodal spectra confuse the relationship. Sand C shows no clear relationship and thickness appears to have no relation to peak frequency. Sands below the tuning frequency for the bed have clearer relationship to the modeled and theoretic curves. The method of peak frequency as a tool for bed thickness estimation shows no simple relationship. For unequal odd impulse pairs there are two thicknesses for one peak frequency value.