Prediction of glutenite reservoirs using an adaptive Bayesian seismic inversion in the slope zone of Mahu Depression in Junggar Basin, NW China

Abstract

The slope zone of Mahu Depression in Junggar Basin contains the largest glutenite oilfield discovered over the world so far. However, because of the mechanism of near-source sedimentation and rapid facies transition, the glutenite reservoirs are strongly heterogeneous which makes it difficult to be discriminated directly from seismic data. To characterize the lateral variation and improve the vertical resolution of glutenite reservoirs, we develop an adaptive impedance inversion method based on the Bayesian theory. In this method, an automatically adjusted damping factor is derived to obtain the best balance between the vertical resolution and the inversion stability according to the noise level. And the trace-by-trace recursive inversion strategy, which uses the inversion result of the previous adjacent trace as the initial model for the next, is adopted to ensure the lateral variation. Synthetic data tests on 1-D and 2-D models verify the adaptive ability and the high resolution advantage of the introduced method. Real data application results in Mahu Oilfield shows that the prediction results of glutenite reservoir have distinct lateral variation and high vertical resolution, with good agreement with the actual drilling and sedimentary trend, which indicates a good application prospect in the heterogeneous glutenite reservoir exploration.