112-6 Integrating Seismic Inversion Products into Multi-Attribute Analysis and Unsupervised Machine Learning for Reservoir Characterization

Session: Geophysics in Investigating and Exploring for Mineral, Energy and Groundwater Resources (Posters)

Presenting Author:

Authors:

Abstract:

Summary

Integrating seismic inversion products with multi-attribute analysis significantly improves reservoir characterization. Using the Pohokura Field, located in the Taranaki Basin of New Zealand, as a case study, we demonstrate three key benefits: (1) enhanced clarity of structural discontinuity boundaries, (2) improved recognition of reservoir heterogeneity, and (3) better identification and characterization of reservoir zones. Comparative Principal Component Analysis (PCA) demonstrates that combined attribute-inversion approaches provide substantially more detailed insights into the complex Mangahewa Formation. This approach provides a powerful workflow that can be effectively applied to other complex reservoirs.

Introduction and Methods

In the Pohokura Field, the primary reservoir unit, the Mangahewa Formation, is a marginal marine substantial tight-sand reservoir with high heterogeneity. The main objective of this study is to improve the understanding of reservoir heterogeneity and distribution by identifying lithofacies variations, structural features and reservoir rock properties. For this purpose, we adopted a novel workflow that integrates pre-stack seismic inversion products into multi-attribute analysis, reduces attribute redundancy through dimensionality reduction techniques, and enhances reservoir characterization using unsupervised machine learning methods.

We first utilized the Pohokura Merge 3D seismic amplitude volume in the Attribute-Assisted Seismic Processing and Interpretation (AASPI) software and generated seismic attributes. Among the extracted attributes, GLCM (Gray Level Co-occurrence Matrix) homogeneity, coherent energy, curvedness, peak frequency and peak magnitude were selected as inputs for PCA. To compare the effects of including seismic inversion products, inverted P- and S-impedance volumes were included among the inputs and PCA was performed again.

Results and Conclusions

Integrating seismic inversion products into multi-attribute analysis significantly improves the characterization of the Pohokura hydrocarbon reservoir. Firstly, structural discontinuities (1) are resolved with improved clarity, allowing clearer delineation of continuous structural features and enabling higher-detail imaging. Secondly, cross plot comparisons reveal considerable improvements in the representation of geological structures, stratigraphic layers and reservoir boundaries (2). Areas that previously appeared as chaotic clusters of colors have been transformed into coherent, meaningful contours on the depth slice. Thirdly, inter- and intra-formational heterogeneities become distinguishable at a higher resolution. The uniformity in color distribution associated with lithological variations has decreased, leading to improved differentiation and more accurate identification of reservoir heterogeneity zones (3). These improvements enhanced the definition of structural features and heterogeneity zones, thereby improving reservoir characterization.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

doi: 10.1130/abs/2025AM-9676

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.