Rock Physics-Based Joint Assimilation of Seismic and Pressure Data for Monitoring and Predicting CO2 Plume Migration: a Cranfield Case Study

Abstract:

The sequestration of anthropogenic carbon emissions into geological formations has been identified as an innovative method to mitigate climate change as it ensures that CO₂ remains trapped in the subsurface while offsetting atmospheric CO₂ concentrations. However, geological carbon storage (GCS) carries the risk of CO₂ migration into nearby viable aquifers or the surface. This makes the monitoring, verification, and accounting (MVA) of injected CO₂ crucial to any GCS project. Furthermore, MVA plans are designed around the capabilities of the monitoring tools and risk-based mitigation strategies are significantly influenced by the quality of available data. Seismic surveys offer an attractive monitoring strategy due to their relative ease in interrogating a large volume of reservoir. As CO₂ is injected, visible changes in seismic attributes (most notably p-wave seismic velocity and attenuation) are recorded. Estimating CO₂ plume properties using seismic data is a challenging task and requires robust quantitative risk assessment tools to comply with the EPA guidelines.

We propose to explore the impacts and quantify the predictive accuracy of an integrated seismic-pressure-petrophysical characterization model using ensemble-based data assimilation. In the context of data assimilation, we test this approach in an observation system simulation experiment (OSSE) using the Cranfield site as our testbed to study the composite observation system. Additionally, by using a geologically realistic model that captures the inclined heterolithic stratification and accretion surfaces of the Cranfield reservoir, we are able to assess the impacts of heterogeneity on CO₂ plume evolution and migration. We simulate continuous temporal seismic observations (p-wave seismic velocity and attenuation) by utilizing pressure and saturation results at each assimilation cycle and passing them through White’s patchy gas saturation model. The risk-based value added of this data acquisition system is quantified in the form of improved gas saturation estimates and reduced uncertainty. With the objective to integrate real-time seismic and pressure monitoring data during GCS, our work provides a framework to test the value-added of a composite observation system that utilizes data assimilation and rock physics models.