Rock formation characterization for carbon dioxide geosequestration: 3D seismic amplitude and coherency anomalies, and seismic petrophysical facies classification, Wellington and Anson-Bates Fields, Kansas, USA

• Coherency and amplitude anomalies agree with the Mid-Continent Rift fault trend.
• Petrophysical facies mapping based on application of a supervised artificial network
• Validation of petrophysical facies classes based on impedance-porosity cross-plot

Higher resolution rock formation characterization is of paramount priority, amid growing interest in injecting carbon dioxide, CO2, into subsurface rock formations of depeleting/depleted hydrocarbon reservoirs or saline aquifers in order to reduce emissions of greenhouse gases. In this paper, we present a case study for a Mississippian carbonate characterization integrating post-stack seismic attributes, well log porosities, and seismic petrophysical facies classification. We evaluated changes in petrophysical lithofacies and reveal structural facies-controls in the study area. Three cross-plot clusters in a plot of well log porosity and acoustic impedance corroborated a Neural Network petrophysical facies classification, which was based on training and validation utilizing three petrophysically-different wells and three volume seismic attributes, extracted from a time window including the wavelet of the reservoir-top reflection. Reworked lithofacies along small-throw faults has been revealed based on comparing coherency and seismic petrophysical facies. The main objective of this study is to put an emphasis on reservoir characterization that is both optimized for and subsequently benefiting from pilot tertiary CO2 carbon geosequestration in a depleting reservoir and also in the deeper saline aquifer of the Arbuckle Group, south central Kansas. The 3D seismic coherency attribute, we calculated from a window embracing the Mississippian top reflection event, indicated anomalous features that can be interpreted as a change in lithofacies or faulting effect. An Artificial Neural Network (ANN) lithofacies modeling has been used to better understand these subtle features, and also provide petrophysical classes, which will benefit flow-simulation modeling and/or time-lapse seismic monitoring feasibility analysis. This paper emphasizes the need of paying greater attention to small-scale features when embarking upon characterization of a reservoir or saline-aquifer for CO2 based carbon geosequestration.