5. Seismic AVO for Gas-Hydrate-Related Reflections

Abstract

This paper examines the usefulness of amplitude versus offset (AVO) analysis for marine and permafrost gas-hydrate settings. In marine environments, AVO analyses have traditionally focused on bottom-simulating reflections (BSRs) for estimating associated marine gas-hydrate and free-gas concentrations. A nonlinear Bayesian inversion is applied to estimate marginal probability distributions (MPDs) of physical parameters at a BSR interface, which are related to overlying gas-hydrate and underlying free-gas concentrations via rock physics modeling. The problem is further constrained by prior information and reparameterization of inversion results. Inversion of BSR AVO data from offshore Vancouver Island, Canada, shows that gas-hydrate and free-gas concentrations are 0%–23% and 0%–2% of the pore volume, respectively, at a 90% credibility level. However, it should be noted that these two parameters are not independent in the analysis but instead are inversely related. Thus, 0% in gas-hydrate concentrations requires some percent of free gas within the range stated (and vice versa). This result indicates that the data do not provide sufficient information to independently resolve gas-hydrate and free-gas concentrations to useful accuracy. The same Bayesian inversion method is applied to synthetic AVO data generated from well-log data obtained at the Mallik 5L-38 well in the Mackenzie Delta, Northwest Territories (NWT). The model cases representing typical permafrost gas-hydrate occurrences are investigated and include shale over gas-hydrate-bearing sand, gas-hydrate-bearing sand over shale, and gas-hydrate-bearing sand over water-saturated sand scenarios. The AVO inversion sufficiently constrains the shear-wave velocity for reliable quantitative analysis only if the gas hydrate concentration exceeds ∼40%. The variable degree of model constraint obtained in this AVO study highlights the need to include rigorous quantitative uncertainty analysis in all AVO studies.