Seismic Attribute Analyses and Attenuation Applications for Detecting Gas Hydrate Presence

Identifying gas hydrates in the oceanic subsurface using seismic reflection data supported by the presence of a bottom simulating reflector (BSR) is not an easy task, given the wide range of geophysical methods that have been applied to do so. Though the presence of the BSR is attributed to the attenuation response, as seismic waves transition from hydrate-filled sediment within the gas hydrate stability zone (GHSZ) to free gas-bearing sediment below, few studies have applied a direct attenuation measurement. To improve the detection of gas hydrates and associated features, including the BSR and free gas accumulation beneath the gas hydrates, we apply a recently developed method known as Sparse-Spike Decomposition (SSD) that directly measures attenuation from estimating the quality factor (Q) parameter. In addition to performing attribute analyses using frequency attributes and a spectral decomposition method to improve BSR imaging, using a comprehensive analysis of the three methods, we make several key observations. These include the following: (1) low-frequency shadow zones seem to correlate with large values of attenuation; (2) there is a strong relationship between the amplitude strength of the BSR and the increase of the attenuation response; (3) the resulting interpretation of migration pathways of the free gas using the direct attenuation measurement method; and (4) for the data analyzed, the gas hydrates themselves do not give rise to either impedance or attenuation anomalies that fully differentiate them from nearby non-hydrate zones. From this last observation, we find that, although the SSD method may not directly detect in situ gas hydrates, the same gas hydrates often form an effective seal trapping and deeper free gas accumulation, which can exhibit a large attenuation response, allowing us to infer the likely presence of the overlying hydrates themselves.

A NEW LOOK AT THE DUAL DEPTH OF INVESTIGATION OF LWD PROPAGATION RESISTIVITY LOGGING

It is well established that phase shift and attenuation measurements acquired by an electromagnetic propagation tool come with different depths of investigation (DOI). The attenuation measurement sees deeper into the formation than the phase shift measurement. This difference has been reported not only for the 2 MHz propagation resistivity tool, but also for the deep propagation tool that operates at 25 MHz. Although the difference has been demonstrated with modeling, test tank experiments and logs, a complete physical explanation has been notably absent since the introduction of the MHz-frequency propagation logging in 1980s. The question is so intriguing that it has been raised repeatedly over the past decades: what drives the difference of DOI for the two measurements that are acquired with the same electromagnetic field? In this paper, we revisit this problem with an aim of providing a physical insight to bridge the gap between theory and application. This is an extension of our recent work on the theory of apparent conductivity for propagation measurements. We address the problem by applying high-order geometric theory for low-frequency electromagnetic problems in lossy media in conjunction with the Taylor series expansion for the voltage ratio measured by a propagation tool. In so doing, we find that in a resistive formation where the dielectric effect is small: 1) the phase shift measurement is primarily due to the first-order eddy current induced in the formation; 2) in contrast, the leading source of the attenuation measurement is the second-order eddy current. Since the second-order eddy current is more spread out than the first-order eddy current, this explains why the DOI of attenuation resistivity is larger than that of phase shift resistivity. The difference in spatial distribution of two eddy currents is also the reason for the difference of vertical resolution between the two. The same root cause for the difference of DOI and vertical resolution also holds when comparing R-signal and X-signal from induction resistivity logging. Other properties shared by propagation and induction resistivity logging will be discussed, such as skin effect and dielectric effect, as well as their asymptotic properties in high-resistivity formations. We conclude that propagation and induction resistivity logging are essentially similar, even though the two measurement principles may seem rather different.