A comparison of competing amplitude variation with offset techniques applied to tight gas sand exploration in the Cooper Basin of Australia

Abstract

We have developed a tight gas amplitude variation with offset (AVO) case history from the Cooper Basin of Australia that addressed the exploration problem of mapping thin fluvial tight gas sand bodies. In the Cooper Basin, Permian Toolachee and Patchawarra sands are difficult to interpret on seismic data due to strong reflections from adjacent Permian coals. This is not the common AVO problem of distinguishing between coal and gas sand, but a more difficult class-I AVO problem of mapping fluvial sands beneath a sheet coal that varies in thickness. We have reviewed local rock properties and concluded that Poisson’s ratio is probably the most appropriate rock property to solve the above exploration problem. We have compared various seismic attributes made using the extended elastic impedance (EEI) technique and a rotation of near and far partial stacks. In a synthetic modeling study that included random noise and tuning, we compared the noise-discrimination abilities of three competing AVO crossplot techniques and “rotated” the attributes made from them. These three crossplots were as follows: intercept versus gradient (I-G), full-stack versus far-minus-near (Full-FmN), and near-stack versus far-stack (N-F). Previous papers on this subject have found that (I-G) crossplots had a spurious correlation in the presence of noise that did not occur with the (Full-FmN) and (N-F) crossplots. We found that for our class-I AVO case, (1) the advantage of the (Full-FmN) and (N-F) crossplots disappeared in the presence of tuning, (2) if tuning was present, the optimal rotation angle was determined by the “tuning angle,” not by the noise angle or some desired EEI angle, and (3) if the three different crossplots were rotated by their respective “tuning” angles, the results were identical.