Direct Geostatistical Seismic Amplitude Versus Angle Inversion for Shale Rock Properties

Abstract

The purpose of seismic reservoir characterization is to predict the spatial distribution of the subsurface rock properties from a set of direct and indirect measurements. Commonly, rock property volumes are obtained in a two-steps approach. First, the elastic properties are inverted from seismic reflection data and then used to compute rock properties volumes by applying a calibrated rock physics models. Such an approach is not only time consuming but may also lead to biased results as the uncertainties related to seismic inversion may not be propagated through the entire geomodeling workflow. Here, we propose an iterative geostatistical shale rock physics seismic AVA inversion method to invert seismic reflection data directly for shale rock properties. The workflow consists of three main steps starting from shale properties model generation of brittleness, total organic carbon (TOC), and porosity using stochastic sequential simulation and cosimulation and calculation of a volume of shale. In the following step, elastic property volumes are calculated based on a calibrated shale rock physics model using the self-consistent approximation. The elastic models are then used for the calculation of the synthetic seismic data. In the final step, the misfit between synthetic and real data is calculated and used as part of the stochastic update of the model parameter space. The whole process is repeated until a minimum misfit between the observed and synthetic seismic is achieved. The proposed method is successfully tested on an onshore Lower Paleozoic shale gas reservoir in Northern Poland, and the predicted shale rock properties agree with those observed at a blind-well location.

Hybrid-Driven High-Resolution Prestack Seismic Inversion

“…HE accuracy of reservoir characterization usually relies on the analysis of subsurface attributes (e.g., velocity, density and impedance). Amplitude-variation-with-offset/angle (AVO/AVA) inversion is a frequent methodology to obtain velocity and density from recorded seismogram [1][2][3][4]. Due to the advantage of low cost, the convolutional model is still the most widely used method in the oil or gas industry, i.e., the seismic record is represented by convolving the reflection coefficient with seismic wavelet [5].…”

“…Stochastic inversion and sparsely constrained deterministic inversion are the most commonly used methods to recover high-resolution results from seismic data. Stochastic inversion methods can produce high resolution inversion results, however, the expensive computational costs limit their use in practice [2,[10][11][12]. Deterministic inversion methods based on sparse constraints are widely adopted in practical applications as an alternative to provide high-resolution inversion results due to the low cost.…”