Geochemical processes change the microstructure of rocks and thereby affect their physical behaviour at the macro scale. A micro-computer tomography (micro-CT) scan of a typical reservoir sandstone is used to numerically examine the impact of three spatial alteration patterns on pore morphology, permeability and elastic moduli by correlating precipitation with the local flow velocity magnitude. The results demonstrate that the location of mineral growth strongly affects the permeability decrease with variations by up to four orders in magnitude. Precipitation in regions of high flow velocities is characterised by a predominant clogging of pore throats and a drastic permeability reduction, which can be roughly described by the power law relation with an exponent of 20. A continuous alteration of the pore structure by uniform mineral growth reduces the permeability comparable to the power law with an exponent of four or the Kozeny–Carman relation. Preferential precipitation in regions of low flow velocities predominantly affects smaller throats and pores with a minor impact on the flow regime, where the permeability decrease is considerably below that calculated by the power law with an exponent of two. Despite their complete distinctive impact on hydraulics, the spatial precipitation patterns only slightly affect the increase in elastic rock properties with differences by up to 6.3% between the investigated scenarios. Hence, an adequate characterisation of the spatial precipitation pattern is crucial to quantify changes in hydraulic rock properties, whereas the present study shows that its impact on elastic rock parameters is limited. The calculated relations between porosity and permeability, as well as elastic moduli can be applied for upscaling micro-scale findings to reservoir-scale models to improve their predictive capabilities, what is of paramount importance for a sustainable utilisation of the geological subsurface.

Deep un-mineable coal deposits are viable reservoirs for permanent and safe storage of carbon dioxide (CO₂) due to their ability to adsorb large amounts of CO₂ in the microporous coal structure. A reduced amount of CO₂ released into the atmosphere contributes in turn to the mitigation of climate change. However, there are a number of geomechanical risks associated with the commercial-scale storage of CO₂, such as potential fault or fracture reactivation, microseismic events, cap rock integrity or ground surface uplift. The present study assesses potential site-specific hydromechanical impacts for a coal deposit of the Upper Silesian Coal Basin by means of numerical simulations. For that purpose, a near-field model is developed to simulate the injection and migration of CO₂, as well as the coal-CO₂ interactions in the vicinity of horizontal wells along with the corresponding changes in permeability and stresses. The resulting effective stress changes are then integrated as boundary condition into a far-field numerical model to study the geomechanical response at site-scale. An extensive scenario analysis is carried out, consisting of 52 simulation runs, whereby the impacts of injection pressures, well arrangement within two target coal seams as well as the effect of different geological uncertainties (e.g., regional stress regime and rock properties) is examined for operational and post-operational scenarios. The injection-induced vertical displacements amount in maximum to 3.59 cm and 1.07 cm directly above the coal seam and at the ground surface, respectively. The results further demonstrate that neither fault slip nor dilation, as a potential consequence of slip, are to be expected during the investigated scenarios. Nevertheless, even if fault integrity is not compromised, dilation tendencies indicate that faults may be hydraulically conductive and could represent local pathways for upward fluid migration. Therefore, the site-specific stress regime has to be determined as accurately as possible by in-situ stress measurements, and also fault properties need to be accounted for an extensive risk assessment. The present study obtained a quantitative understanding of the geomechanical processes taking place at the operational and post-operational states, supporting the assessment and mitigation of environmental risks associated with CO${}_2$ storage in coal seams.

Reliable predictions of groundwater levels are crucial for sustainable groundwater resource management, which needs to balance diverse water needs and to address potential ecological consequences of groundwater depletion. Machine learning (ML) approaches for time series forecasting have shown promising accuracy for groundwater level prediction and, furthermore, offer scalability advantages over traditional numerical methods when sufficient data are available. Global ML architectures enable predictions across numerous monitoring wells concurrently using a single model, allowing predictions over a broad range of hydrogeological and meteorological conditions and simplifying model management. In this contribution, groundwater levels for 5288 monitoring wells across Germany were forecasted up to 12 weeks using two state-of-the-art ML approaches, the Temporal Fusion Transformer (TFT) and the Neural Hierarchical Forecasting for Time Series (N-HiTS) algorithm. The models were provided with historical groundwater levels, meteorological features, and a wide range of static features describing hydrogeological and soil properties at the monitoring sites. To determine the conditions under which the model achieves good performance and whether it aligns with hydrogeological system understanding, the model's performance was evaluated spatially and correlations with both static input features and time series features from hydrograph data were examined. The N-HiTS model outperformed the TFT model, achieving a median Nash–Sutcliffe efficiency (NSE) of 0.5 for the 12-week prediction over all 5288 monitoring wells. Performance varied widely: 25 % of wells achieved an NSE >0.68, while 15 % had an NSE <0 with the best N-HiTS model. A tendency for better predictions in areas with high data density was observed. Moreover, the models achieved higher performance in lowland areas with distinct seasonal groundwater dynamics, in monitoring wells located in porous aquifers, and at sites with moderate permeabilities, which aligns with theoretical expectations. Overall, the findings highlight that global ML models can facilitate accurate seasonal groundwater predictions over large, hydrogeological diverse areas, potentially informing future groundwater management practices at a national scale.

Quantifying interactions and dependencies among geometric, hydraulic and mechanical properties of reservoir sandstones is of particular importance for the exploration and utilisation of the geological subsurface and can be assessed by synthetic sandstones comprising the microstructural complexity of natural rocks. In the present study, three highly resolved samples of the Fontainebleau, Berea and Bentheim sandstones are generated by means of a process-based approach, which combines the gravity-driven deposition of irregularly shaped grains and their diagenetic cementation by three different schemes. The resulting evolution in porosity, permeability and rock stiffness is examined and compared to the respective micro-computer tomographic (micro-CT) scans. The grain contact-preferential scheme implies a progressive clogging of small throats and consequently produces considerably less connected and stiffer samples than the two other schemes. By contrast, uniform quartz overgrowth continuously alters the pore space and leads to the lowest elastic properties. The proposed stress-dependent cementation scheme combines both approaches of contact-cement and quartz overgrowth, resulting in granulometric, hydraulic and elastic properties equivalent to those of the respective micro-CT scans, where bulk moduli slightly deviate by 0.8%, 4.9% and 2.5% for the Fontainebleau, Berea and Bentheim sandstone, respectively. The synthetic samples can be further altered to examine the impact of mineral dissolution or precipitation as well as fracturing on various petrophysical correlations, which is of particular relevance for numerous aspects of a sustainable subsurface utilisation.

Mineral dissolution and precipitation can substantially affect rock permeability, which is a critical parameter for a broad range of geological subsurface applications. Virtual experiments on digital pore-scale samples represent a powerful and flexible approach to understand the impact of microstructural alterations on evolving hydraulic rock behaviour and quantify trends in permeability. In the present study, porosity-permeability relations are simulated for a precipitation-dissolution cycle within a typical reservoir sandstone. A hysteresis in permeability is observed depending on the geochemical process and dominating reaction regime, whereby permeability of the six investigated reaction paths varies by more than two orders of magnitude at a porosity of 17 %. Controlling parameters for this hysteresis phenomenon are the closure and re-opening of micro-scale flow channels, derived from changes in pore throat diameter and connectivity of the pore network. In general, a transport-limited regime exhibits a stronger impact on permeability than a reaction-limited regime, which uniformly alters the pore space. In case of mineral precipitation, higher permeability reduction results from successive clogging of pore throats, whereas in case of dissolution, permeability significantly increases due to a widening of existing flow paths. Both, the geochemical process and dominating reaction regime govern characteristic microstructural alterations, which cannot be simply reversed by the inversion of the geochemical processes itself. Hence, permeability evolution clearly depends on the hydrogeochemical history of the sample.