Online Journal for E&P Geoscientists

Alshangiti, Abdullah M.^*1; Chapman, Mark ²; Wilson, Adam ²; Clark, Roger ³
(1) Geophysical Data Processing Division, Saudi ARAMCO, Dhahran, Saudi Arabia. (2) Edinburgh Anisotropy Project, British Geological Survey, Edinburgh, United Kingdom. (3) School of Earth and Environment, University of Leeds, Leeds, United Kingdom.

Analysis of seismic data suggests that hydrocarbon saturated and/or fractured zones are often associated with high values of attenuation; this attenuation causes strong velocity dispersion. Due to dispersion, the impedance will be frequency-dependent. The measurement of attenuation and frequency-dependent impedance from seismic data may help in distinguishing between reservoir and non-reservoir zones. They measurements can also be used as a tool to detect fractures.

A recently developed method was applied to measure the frequency-dependent impedance from synthetic and field data. For a three-layer model of elastic and dispersive interfaces, frequency-dependent impedances were measured, and the results show that the impedance varies strongly with frequency in the dispersive interface (i.e., fractured and gas-saturated sandstone) but not in the elastic one (i.e., homogenous and isotropic shale). The frequency-dependent inversion produced results with error between 7.5-10%. The frequency-dependent impedances were then measured from field CMPs that fall in the same Fresnel zone. The results are very consistent along these CMPs, where two events with strong variations of impedance with frequency may be related to the shallow multiple gas-reservoirs. The inversion of these CMPs produced results with an error of 30% for the full-bandwidth inversion and 40% for the frequency-dependent inversion.

As an independent supportive measurement, attenuation 1/Q was measured using the prestack Q inversion method. This method produced accurate 1/Q values from a synthetic CMP gather with error of only 4%. The measured 1/Q values from the field CMPs are very consistent with very small uncertainties.

The results from both methods agree with each other in the following manner: where the interfaces with large variations of impedance with frequency have small Q, and those with small variations have large Q. These results are in agreement with the relationship between the frequency-dependent impedance and Q.