Laboratory-Scale 4D Seismic Monitoring of CO₂ Storage Under Sparse Acquisition

Reliable monitoring of subsurface CO₂ plumes is essential for ensuring the safety, efficiency, and regulatory compliance of carbon capture and storage (CCS) operations. While time-lapse seismic monitoring and numerical simulations are widely used, both approaches face limitations related to cost and uncertainty. Laboratory-scale experiments provide a valuable complementary pathway by enabling controlled, repeatable studies with real-world physics. Here, we present a new open-access laboratory facility designed to investigate seismic monitoring strategies for CO₂ storage using three-dimensional ultrasonic imaging.

The facility consists of a 1:2000 downscaled physical replica of the upper caprock of the Utsira Formation at the Sleipner CO₂ storage site, submerged in water and monitored by a dense array of ultrasonic transducers. A total of 128 high-frequency (1 MHz) and low-frequency (0.15 MHz) piezoelectric transducers are configured to act as both sources and receivers, enabling highly flexible acquisition geometries. Using air as a safe laboratory proxy for CO₂, controlled injection experiments were conducted to emulate plume migration beneath the caprock. Continuous scanning of the transducer array allows the acquisition of large 3D time-lapse (“4D”) datasets within minutes.

The resulting data demonstrates clear detection of gas accumulation and migration pathways beneath the model caprock, as well as successful imaging of the aquifer topography. Time-lapse amplitude changes correlate well with independently observed plume evolution, particularly when using the low-frequency, wide-beam transducers, which provide improved illumination of complex topography.

We further investigated the impact of data sparsity by systematically decimating the ultrasonic dataset. Both systematic and random receiver reductions were tested to emulate sparse and irregular monitoring geometries. Because the true plume extent is known from direct visual observations, the quality of the resulting seismic monitoring could be quantitatively evaluated as a function of sampling density and decimation strategy. The results demonstrate clear differences in detection performance between the two sparsity patterns. These findings provide important insights for the design of cost-efficient and reliable seismic monitoring programs for CO₂ storage.

The facility provides a versatile and scalable platform for testing seismic imaging techniques, acquisition strategies, and processing workflows under controlled conditions. Future developments will include an enhanced acquisition setup, for even higher flexibility, repeatability and quality, opening up new avenues of research with advanced processing techniques.