The role of capillary forces during buoyant migration of CO₂ is critical toward plume immobilization within the postinjection phase of a geological carbon sequestration operation. However, the inherent heterogeneity of the subsurface makes it very challenging to evaluate the effects of capillary forces on the storage capacity of these formations and to assess in situ plume evolution. To overcome the lack of accurate and continuous observations at the field scale and to mimic vertical migration and entrapment of realistic CO₂ plumes in the presence of a background hydraulic gradient, we conducted two unique long‐term experiments in a 2.44 m × 0.5 m tank. X‐ray attenuation allowed measuring the evolution of a CO₂‐surrogate fluid saturation, thus providing direct insight into capillarity‐dominated and buoyancy‐dominated flow processes occurring under successive drainage and imbibition conditions. The comparison of saturation distributions between two experimental campaigns suggests that layered‐type heterogeneity plays an important role on nonwetting phase (NWP) migration and trapping, because it leads to (i) longer displacement times (3.6 months versus 24 days) to reach stable trapping conditions, (ii) limited vertical migration of the plume (with center of mass at 39% versus 55% of aquifer thickness), and (iii) immobilization of a larger fraction of injected NWP mass (67.2% versus 51.5% of injected volume) as compared to the homogenous scenario. While these observations confirm once more the role of geological heterogeneity in controlling buoyant flows in the subsurface, they also highlight the importance of characterizing it at scales that are below seismic resolution (1–10 m).

An accurate assessment of the physical states of the Earth system is an essential component of many scientific, societal, and economical considerations. These assessments are becoming an increasingly challenging computational task since we aim to resolve models with high resolutions in space and time, to consider complex coupled partial differential equations, and to estimate uncertainties, which often requires many realizations. Machine learning methods are becoming a very popular method for the construction of surrogate models to address these computational issues. However, they also face major challenges in producing explainable, scalable, interpretable, and robust models. In this paper, we evaluate the perspectives of geoscience applications of physics-based machine learning, which combines physics-based and data-driven methods to overcome the limitations of each approach taken alone. Through three designated examples (from the fields of geothermal energy, geodynamics, and hydrology), we show that the non-intrusive reduced-basis method as a physics-based machine learning approach is able to produce highly precise surrogate models that are explainable, scalable, interpretable, and robust.

Benchmarks are essential in the design of modern HPC installations, as they define key aspects of system components. Beyond synthetic workloads, it is crucial to include real applications that represent user requirements into benchmark suites, to guarantee high usability and widespread adoption of a new system. Given the significant investments in leadership-class supercomputers of the exascale era, this is even more important and necessitates alignment with a vision of Open Science and reproducibility. In this work, we present the JUPITER Benchmark Suite, which incorporates 16 applications from various domains. It was designed for and used in the procurement of JUPITER, the first European exascale supercomputer. We identify requirements and challenges and outline the project and software infrastructure setup. We provide descriptions and scalability studies of selected applications and a set of key takeaways. The JUPITER Benchmark Suite is released as open source software with this work at github.com/FZJ-JSC/jubench