Shallow Prospect Reservoir Characterization Using Geomorphology and Rock Physics Guided Bayesian Classification

To improve the reservoir characterization in frontier exploration, we need to look beyond the conventional approaches and incorporate various spectra of geoscience. With the goal of Lithology and Fluid Prediction (LFP), we employ a variety of such techniques from survey matching, AVO modelling, elastic impedance (EI) inversion, spectral decomposition, edge attributes, followed by the rock physics guided Bayesian classification and bump map visualisation. Since the prospect is covered by a IsoMetrix broadband dataset acquired in 2015 without the well control and we have a slightly overlapping conventional data set acquired in 1998 with well control, we started with rigorous survey matching and then merged the data sets to utilize the available well as a control point. The well was used for signal estimation, the low frequency background model for inversion and to define the Bayesian classification framework.

We present a case study where the convincing results provide more understanding about the prospect, and acknowledging the uncertainties in the inference of LFP due to the prospect being covered only by far angle data.