Evaluating reservoir simulation models against 4D seismic data for open access data from the Sleipner field

The Sleipner CO 2 storage project in the North Sea has become a global reference for carbon capture, utilization, and storage (CCUS), offering a dataset for testing and validating geophysical methodologies. Among these, time-lapse (4D) seismic data stands out as a key tool for addressing the crucial challenge of verifying CO2 plume confinement and mitigating leakage risk, which are essential for ensuring the long-term safety and effectiveness of CCUS operations. This is because 4D seismic captures dynamic changes occurring in the reservoir over time across its entire extension, including areas where well data are limited. Integrating such data with dynamic reservoir models further enhances monitoring activities, enabling assessments of models’ accuracy and their ability to produce reliable forecasts. In this study, we assess the capacity of publicly available three-dimensional flow models from the British Geological Survey to replicate observed 4D seismic responses at the Sleipner East site. Real 4D seismic data provided by Equinor were compared against synthetic responses generated via petroelastic modeling and convolutional forward modeling of the simulation outputs. To address gaps in model documentation, we implemented a robust workflow that integrates literature-derived rock physics parameters with targeted calibration of petroelastic models. Model accuracy was quantified using the Structural Normalized Cross Correlation (NCC) and Root Mean Square Error metrics applied to attribute maps across three discrete reservoir intervals. Although none of the evaluated models capture every feature of the field data—reflecting their simplified geological frameworks—several closely reproduce the CO 2 plume morphology and signal polarity in specific zones. Our workflow enables reproducible seismic modeling for CCUS benchmarking, with Model 10 achieving the highest similarity (NCC = 0.754) in Zone 1. Moreover, this work provides a reproducible protocol for constructing petroelastic models and conducting seismic forward modeling using publicly accessible datasets for Sleipner, thereby informing best practices for future CCUS monitoring strategies and providing a quantitative basis for risk management. • Evaluation of ten public flow models against 4D seismic at Sleipner East. • Workflow integrates rock physics parameters with petroelastic calibration. • Model performance assessed using NCC and RMSE across reservoir intervals. • Several models reproduce CO 2 plume morphology and signal polarity. • Provides reproducible protocol for seismic modeling in CCUS benchmarking.