Parametric Controls on Dynamic CO₂ Storage Capacity in Fault-Bounded Saline Aquifer: The Lopín Structure Case Study (Ebro Basin, NE Spain)

1. Introduction

Geological storage of CO₂ is widely recognized as a key component of carbon capture and storage (CCS) strategies for reducing emissions from industrial sectors that are difficult to decarbonize through electrification alone [1,2]. Among the available geological options, deep saline aquifers are particularly relevant because of their broad distribution in sedimentary basins and their high theoretical storage potential [1,3,4]. However, these systems present specific technical challenges, including limited prior subsurface information, buoyancy-driven CO₂ migration, brine displacement, lower storage efficiency than depleted hydrocarbon reservoirs, and pressure build-up during injection [3,4]. Therefore, reliable assessment of saline aquifer storage sites requires not only estimating available pore volume, but also dynamically analysing pressure evolution, injectivity, plume migration, and long-term containment.

The concept of CO₂ storage capacity has evolved from static volumetric estimation toward dynamic, pressure-constrained assessments. Static methods commonly estimate capacity as a fraction of the available pore volume using storage-efficiency coefficients [5]. These approaches are useful for regional screening, but they do not fully capture the hydraulic response of the reservoir during injection. Several studies have shown that pressure build-up may become a primary limitation on the effective capacity of deep saline aquifers, especially in closed or semi-closed systems where displaced brine cannot dissipate freely [6,7,8,9]. In such cases, the technically achievable capacity may be significantly lower than the theoretical volumetric capacity, and the maximum injectable mass is controlled by reservoir transmissibility, boundary conditions, well configuration, injection strategy, and admissible pressure thresholds.

Regarding pressure management is closely linked to storage safety. Quick pressure increase may sometimes compromise seal integrity, induce hydraulic fracturing, reactivate faults, or generate leakage pathways through pre-existing discontinuities [6,7,10]. For this reason, operational storage assessments commonly incorporate bottom-hole pressure (BHP) limits or other admissible pressure criteria as practical constraints on injection. These thresholds establish a link between multiphase-flow simulation and geomechanical safety, even when fully coupled hydro-mechanical modelling is not yet available. Consequently, pressure-limited dynamic capacity should be reported together with the adopted pressure criterion and its geological or geomechanical justification.

In this sense, numerical simulation provides the most appropriate framework for evaluating pressure-limited dynamic capacity because it can represent the coupled effects of multiphase flow, buoyancy, pressure propagation, fluid properties, and operational constraints. Three-dimensional models can capture geological heterogeneity in detail, but their computational cost increases considerably when long-term post-injection behaviour, uncertainty, or multiple sensitivity scenarios must be analysed. Vertical-equilibrium (VE) models offer an efficient alternative for laterally extensive saline aquifers, where vertical segregation of CO₂ by buoyancy occurs faster than large-scale lateral migration [11,12]. In VE formulations, the three-dimensional problem is reduced to a vertically integrated representation that preserves the dominant physics of buoyancy-driven migration, pressure dissipation, and regional plume evolution.

The MATLAB Reservoir Simulation Toolbox (MRST) [15] and its co2lab module [SINTEF] have played an important role in developing open-source workflows for geological CO₂ storage modelling. MRST-co2lab includes tools for structural trapping analysis, spill-point identification, VE simulation, long-term migration assessment, and basin-scale capacity estimation [12,13,14,15,16,34]. Previous MRST/co2lab studies have demonstrated the value of vertically integrated models for estimating trapping capacity, optimizing aquifer-wide storage, and simulating large-scale CO₂ migration over long time scales [13,14,15,16,17]. Recent work has also demonstrated three-dimensional geological CO₂ sequestration simulations using MRST [33], while other studies have expanded the use of VE and reduced-complexity models to evaluate storage efficiency, well placement, and computationally efficient sensitivity workflows [18,19,20]. These advances show that VE and MRST-based approaches are well established; however, their application to real cases of fault bounded saline aquifers, considering site specific pressure limits and systematic BHP constrained capacity ranking, remains poorly studied.

In Spain, CO₂ storage research has advanced through national screening initiatives, site specific characterization studies, and European collaborative projects. The ALGECO2 initiative provided an early framework for identifying suitable structures for geological storage. Subsequent studies focused on the Hontomín experimental plant and on the petrographic, hydraulic, geological, and geochemical characterization of potential reservoirs and seals [21,22,23,24,25]. In the Ebro Basin, recent work has highlighted the potential of several structures and has applied integrated geological and geophysical methodologies to the Lopín structure [26]. European initiatives including STRATEGY CCUS, PilotSTRATEGY, and GSEU have further reinforced interest in deep saline aquifers as a strategic resource for regional decarbonization [27,28,29,30,31,32,35].

Despite these advances, a specific gap of knowledge remains at the intersection of pressure-limited dynamic capacity, and fault-bounded saline aquifers. Previous studies have provided important advances in static capacity estimation, pressure-constrained capacity theory, VE model development, MRST/co2lab workflows, and characterization of Spanish storage sites [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35]. However, few published studies combine: (i) a real ECLIPSE/GRDECL geocelular model, where GRDECL refers to the grid and property-deck format commonly used by reservoir simulators, of an onshore saline aquifers in Spain; (ii) a VE workflow adapted to a fault-bounded horst structure; (iii) mixed hydraulic boundary conditions representing partial lateral connection with the regional aquifer; and (iv) a systematic one-at-a-time ranking of BHP-limited dynamic capacity.

The Lopín structure, located in the southern sector of the Ebro Basin (Spain), provides a suitable case study to address this gap. The target reservoir, confined by low-permeability sealing units and bounded by normal faults, makes this structural configuration in Lopín particularly relevant for analyzing how permeability, lateral pressure dissipation, hydraulic boundary conditions, and operational pressure limits control dynamic storage capacity. In this context, the present study develops and applies a MATLAB/MRST VE workflow to quantify the parametric controls on pressure-limited CO₂ storage capacity in the Lopín structure.

The central hypothesis of this work is that, in fault-bounded saline aquifers such as Lopín, operational CO₂ storage capacity is governed more strongly by pressure dissipation and hydraulic transmissibility than by static pore volume alone. To test this hypothesis, the study uses a probabilistic realization with two injection wells, mixed hydraulic boundary conditions, and a BHP-constrained capacity-search algorithm over a 30-year injection period. A one-at-a-time sensitivity analysis is then performed on seven parameter groups: porosity, horizontal permeability, injection rate, BHP limit, brine compressibility, residual CO₂ saturation, and CO₂ density. By identifying the parameters that most strongly control dynamic capacity, this work aims to support early-stage site characterization, pilot design, and preliminary safety assessment of structurally compartmentalized saline aquifers.

2. Geological setting

The Lopín structure is located in the southern sector of the Ebro Basin, northeastern Spain (Figure 1 and Figure 2). The Ebro Basin is one of the major sedimentary basins of the Iberian Peninsula and contains several deep geological formations that have been evaluated for subsurface energy applications and CO₂ storage [24,25,26,27,28]. Within this setting, Lopín is an onshore, fault-bounded structural high selected as a candidate site within the H2020 PilotSTRATEGY project as a candidate site for geological CO₂ storage [27,28].

Structurally, Lopín is interpreted as a NW-SE-trending horst bounded by normal faults (Figure 3). The main target reservoir is the Triassic Buntsandstein B1 unit, a sandstone formation located at approximately 1760 m depth. The reservoir is overlain by low-permeability Triassic clay-rich and evaporitic units that constitute the main sealing interval. The fault-bounded geometry is relevant for storage assessment because it may control both hydraulic compartmentalization and lateral pressure dissipation [26,27,28].

3. Dataset and Methods

3.1. Dataset

The geological and petrophysical data used in this study were provided by IGME-CSIC within the PilotSTRATEGY project in ECLIPSE/GRDECL format [27,28]. The files included the three-dimensional reservoir geometry, porosity distribution, horizontal permeability fields, geological units, shale volume, active-cell mask, and fault traces.

The original geocelular model contains 160,532 active cells. From this model, a VE representation was generated with 2820 columns, of which 2808 were active for simulation. The P10, P50, and P90 probabilistic realizations were supplied by IGME-CSIC as pre-built ECLIPSE/GRDECL models generated within the upstream PilotSTRATEGY geological-modelling workflow; their generation and calibration were not performed as part of this study. Accordingly, the present work treats them as externally provided geological uncertainty realizations. The P50 realization was selected as the reference case because it represents the central condition of the available dataset, whereas P10 and P90 define the broader uncertainty context.

The pressure-limited dynamic capacity, hereafter CapBHP, is defined as the maximum injectable CO₂ mass over 30 years that does not exceed the adopted BHP limit.

3.2. Methodology

3.2.1. Modelling Approach

CO₂ injection and migration were simulated using a VE approach in MATLAB/MRST. The formulation assumes that buoyant CO₂ segregates rapidly toward the top of the reservoir compared with large-scale lateral migration. Under these conditions, each vertical reservoir column can be represented through integrated variables, allowing the model to capture reservoir-scale pressure evolution and plume migration at a lower computational cost than a full three-dimensional multiphase model.

The main dynamic variables were pore pressure and accumulated CO₂ height in each column. From these variables, equivalent CO₂ saturation, plume extent, pressure evolution, and trapping indicators were estimated during the injection and post-injection periods. MRST/co2lab served as a conceptual and computational reference, but the workflow was adapted to the Lopín geometry to control directly the vertical integration of the original 3D geocelular grid into VE columns, the mixed boundary conditions, the well configuration, and the iterative BHP-limited capacity search.

The numerical workflow followed five main stages: 1) input-data preparation, 2) conversion from the 3D grid to the VE grid, 3) Application of the MRST’s fully implicit Newton-AD solver to calculate the pressure and CO₂-height variables, 4) BHP-based injection control, and 5) post-processing of plume geometry, pressure evolution, dynamic capacity, trapping inventory, and mass balance (Figure 4).

3.2.2. Boundary Conditions, Wells, and Capacity Criterion

The geological configuration of the Lopín horst was represented in the numerical model using mixed hydraulic boundary conditions to simulate a partially open system. The northern and southern boundaries, corresponding to the main faults that structurally bound the horst, were defined as no-flow boundaries. In contrast, the eastern and western boundaries were assigned hydrostatic pressure conditions to represent lateral hydraulic connection with the regional aquifer. This configuration represents an intermediate hydraulic behaviour between a closed structural compartment and a fully open saline aquifer.

Injection was simulated through two active wells, C1 and C2. C1 is located on the northwestern flank and C2 near the structural crest (Figure 6); therefore, this pair was selected as the reference configuration to distribute injection and reduce local pressure build-up. The nominal total injection rate for the reference scenario was 0.254 Mt/year over 30 years, distributed between both wells. The complete set of base-case parameters is summarized in Table 1.

Dynamic storage capacity was defined as the maximum mass of CO₂ that can be injected over the 30-year period without exceeding the selected operational BHP criterion (28.5 MPa). This definition distinguishes pressure-constrained dynamic capacity from static pore-volume capacity and explicitly links the capacity estimate to an operational pressure constraint.

The reference BHP limit was set to 28.5 MPa as a conservative operational constraint within the site-specific geomechanical pressure envelope reported for Lopín [28]. This threshold is 1.1 MPa below P90 scenario (P90 = 29.6 MPa), 2.0 MPa below the allowable pressure (pAllow = 30.5 MPa), and 4.4 MPa below the estimated fracture pressure (pFrac = 32.9 Mpa). The 28.5 MPa value is therefore not interpreted as the fracture pressure itself, but as a modelling threshold that preserves a pressure buffer to the upper geomechanical limits. The sensitivity analysis subsequently varies this operational threshold from 26.31 to 30.50 MPa to quantify how the selected pressure criterion affects dynamic capacity. The reference pressure evolution is contextualized against the site-specific p80, p90, pAllow, and pFrac levels in Figure 5; these contextual lines do not control the capacity-search algorithm.

3.2.3. Sensitivity Analysis Design

The central objective of the sensitivity analysis was to identify which parameters exert the strongest control on pressure-limited dynamic storage capacity. For this purpose, a one-at-a-time (OAT) analysis was applied, in which one parameter is modified while all others remain at their base-case values to evaluate the real impact of the variation of this parameter over the modelling.

Seven parameter groups were evaluated: porosity, permeability, injection rate, BHP limit, brine compressibility, residual CO₂ saturation, and CO₂ density. Each parameter group was evaluated using five prescribed parameter values, resulting in 35 sensitivity simulations plus one base case, for a total of 36 simulations. The tested parameters and base values are summarized in Table 2. The system response was evaluated using two indicators: pressure-limited dynamic capacity, expressed in Mt CO₂, and maximum pressure reached during the simulation, expressed in MPa.

4. Results

The results presented in this section correspond to the P50 reference case, using the C1–C2 well configuration and mixed hydraulic boundary conditions described in Section 3.2.2. This configuration was kept fixed throughout the sensitivity analysis to isolate the influence of petrophysical, fluid, and operational parameters on pressure-limited dynamic capacity.

Two different simulation outputs are reported. First, the P50 reference simulation follows the injection of 8.46 Mt of CO₂ over 30 years and evaluates pressure evolution, plume migration, and trapping behaviour during 1000 years of post-injection. Second, the BHP-limited capacity search estimates the maximum injectable mass over the same 30-year injection period without exceeding the adopted operational pressure limit. The reference injected mass and the maximum pressure-limited capacity are therefore not the same quantity.

4.1. Pressure Evolution, Plume Migration, and Dynamic Capacity

In the long-term P50 reference simulation, approximately 8.46 Mt CO₂ are injected over 30 years. Maximum pressure increases during injection, reaches its peak near the end of the injection phase, and then decreases rapidly once injection stops. This decline reflects pressure dissipation through the hydraulically open east-west boundaries, whereas the north and south boundaries, associated with faults, restrict transverse flow.

The maximum-pressure curve exceeds the conservative p80 threshold during injection and approaches the threshold corresponding to 90% of the fracture pressure but remains below the allowable-pressure and estimated fracture-pressure criteria. This behaviour indicates that the reference configuration preserves a safety margin with respect to the upper limits, although it temporarily operates within the conservative alert zone.

Plume migration is governed by buoyancy-driven upward segregation within each VE column and lateral spreading beneath the seal (Figure 6). At early injection times, the plume develops around C1 and C2 sites as two separate accumulations. By the end of injection, both accumulations expand laterally following the structural relief of the reservoir top. During post-injection, the plume continues to redistribute, increases its lateral extent, and preferentially concentrates in structurally shallower areas.

4.2. Storage Mechanisms

The long-term reference simulation indicates that storage security evolves after the end of injection as the CO₂ plume redistributes beneath the vertical seal and is progressively partitioned among mobile, residual, structural, and dissolved fractions. During the post-injection period, part of the mobile CO₂ becomes immobilized by residual trapping as the plume migrates, while another fraction is retained in local structural closures or dissolves into the brine.

At 1030 years, the diagnostic inventory indicates that approximately 77% of the CO₂ remains mobile, 18% is residually trapped, 4% is structurally trapped, and less than 1% is dissolved (Figure 7). These results show that, for the simulated configuration, storage security depends not only on structural retention but also on the progressive evolution of residual trapping. At the same time, the high proportion of mobile CO₂ underlines the importance of evaluating long-term plume migration and seal continuity.

Figure 6. Spatial evolution of integrated CO₂ plume height in the P50 reference scenario, arranged as a 2-column × 3-row sequence for improved readability. Panels show t = 0, 5.1, and 30 years during injection, followed by t = 30, 295, and 1030 years during post-injection. Stars indicate the C1 and C2 injection wells; coloured cells represent integrated CO₂ height h, and contours show reservoir-top depth and the structural framework controlling migration.

Figure 6. Spatial evolution of integrated CO₂ plume height in the P50 reference scenario, arranged as a 2-column × 3-row sequence for improved readability. Panels show t = 0, 5.1, and 30 years during injection, followed by t = 30, 295, and 1030 years during post-injection. Stars indicate the C1 and C2 injection wells; coloured cells represent integrated CO₂ height h, and contours show reservoir-top depth and the structural framework controlling migration.

Figure 7. CO₂ trapping inventory for the P50 reference scenario. Left: final partitioning of CO₂ among structural, dissolved, residual, and mobile fractions after 1030 years. Right: temporal evolution of the trapping inventory throughout the full simulation, including cumulative injected mass.

Figure 7. CO₂ trapping inventory for the P50 reference scenario. Left: final partitioning of CO₂ among structural, dissolved, residual, and mobile fractions after 1030 years. Right: temporal evolution of the trapping inventory throughout the full simulation, including cumulative injected mass.

4.3. Parametric Sensitivity Ranking of Dynamic Capacity

The P50 reference case, using the C1–C2 well configuration and mixed boundary conditions, fixed throughout the sensitivity analysis, allows the individual influence of petrophysical, fluid, and operational parameters on pressure-limited dynamic capacity to be isolated. The one-at-a-time (OAT) analysis described in Section 3.2.3 varies each parameter group independently while keeping the others at their base-case values.

Among the seven parameter groups, horizontal permeability appears to be the dominant controlling factor. When the permeability multiplier varies between ×0.25 and ×4.00, Cap_BHP ranges from 4.665 to 16.252 Mt, with a total range of 11.588 Mt. This range is approximately 107.7% of the base-case capacity (Table 3; Figure 8). Here, Cap_BHP denotes the maximum injectable CO₂ mass over the 30-year injection period without exceeding the adopted BHP limit.

The second most influential parameter is the BHP limit. When the admissible pressure threshold varies between 26.31 and 30.50 MPa, capacity changes from 7.931 to 13.257 Mt, corresponding to a range of 5.326 Mt. This confirms that the operational definition of the pressure limit has a direct effect on the maximum injectable mass.

The third most influential parameter is CO₂ density. Over the evaluated range between 550 and 850 kg/m³, capacity varies from 8.347 to 13.090 Mt, with a range of 4.743 Mt. Together, these three parameters permeability, BHP limit, and CO₂ density explain nearly all relevant variability in pressure-limited dynamic capacity. In contrast, porosity and brine compressibility show much smaller influence, while injection rate and residual CO₂ saturation do not change Cap_BHP under the adopted search formulation.

Permeability exhibits the strongest and non linear response (Figure 9). When the permeability multiplier is reduced to ×0.25, capacity decreases to 4.66 Mt, indicating a strong limitation in lateral pressure dissipation. With a multiplier of ×0.50, capacity increases to 6.60 Mt but remains clearly below the base case. In contrast, doubling permeability increases capacity to 16.25 Mt, the maximum robust value obtained across the sensitivity set. The ×4.00 case produced an anomalous numerical response with non-finite maximum pressure and should therefore be interpreted as solver instability under extreme conditions rather than as a physical capacity reduction.

The BHP limit acts as direct design control. When the conservative threshold of 26.31 MPa is used, maximum capacity decreases to 7.93 Mt. When the criterion is relaxed to 30.50 MPa, capacity increases to 13.25 Mt. This difference shows that the selected pressure limit can have an impact comparable to key physical properties of the storage system.

CO₂ density also significantly affects capacity. At the lowest evaluated density, 550 kg/m³, capacity decreases to 8.34 Mt. At the highest density, 850 kg/m³, it increases to 13.090 Mt. This response indicates that thermodynamic fluid properties influence pressure behaviour and, consequently, the maximum mass that can be injected under the BHP criterion.

4.4. Secondary Control Mechanisms

Porosity has a limited effect on dynamic capacity. Although changing porosity modifies available pore volume, the capacity range remains narrow, between 10.34 and 10.76 Mt. This indicates that, under the evaluated conditions, pore volume is not the primary operational limitation.

Brine compressibility also produces a minor variation, with capacities between 10.59 and 10.92 Mt. This effect is consistent with the role of compressibility in pressure propagation, although its magnitude is much smaller than that of permeability or the BHP limit.

Injection rate does not modify the maximum capacity estimated by the BHP-search algorithm, which remains 10.76 Mt for all evaluated levels. This does not mean that injection rate is operationally irrelevant. Maximum pressure increases from 23.49 MPa at 0.15 Mt/year to 28.60 MPa at 0.40 Mt/year, placing the highest-rate case close to the operational BHP limit. Therefore, injection rate remains important for operational design, even if it does not change the final BHP-limited capacity calculated by the search algorithm.

Residual CO₂ saturation also does not modify Cap_BHP over the evaluated range. Capacity remains constant at 10.760 Mt. This result reflects the fact that the capacity criterion is controlled by injection-period pressure, whereas residual saturation mainly affects post-injection redistribution and trapping.

4.5. Secondary Control Mechanisms

The heatmap provides a compact comparison of the 35 sensitivity cases relative to the base case (Figure 10). The largest relative changes are concentrated in permeability, BHP limit, and CO₂ density. Levels associated with porosity, brine compressibility, injection rate, and residual CO₂ saturation show small or negligible changes in dynamic capacity.

This visual synthesis reinforces the main result of the analysis: under the evaluated configuration, the operational capacity of the Lopín system is controlled primarily by the reservoir’s ability to dissipate pressure rather than by static pore volume alone.

5. Discussion

5.1. Pressure-Limited Dynamic Capacity and Dominant Controls

The results show that CO₂ storage capacity in the Lopín structure cannot be interpreted solely as a function of available pore volume. Although porosity defines the physical space potentially available for CO₂ storage, its influence on pressure-limited dynamic capacity was secondary within the tested range. In contrast, horizontal permeability, the BHP limit, and CO₂ density explained most of the observed variability in maximum injectable mass.

This behaviour supports the central hypothesis of the study: in fault-bounded saline aquifers, operational capacity depends primarily on the ability of the system to dissipate injection-induced overpressure. Therefore, the limiting factor is not simply how much CO₂ can be accommodated within the reservoir pore volume, but how much CO₂ can be injected without exceeding admissible pressure limits. This distinction is fundamental because static volumetric estimates may overestimate effective capacity when pressure evolution is not explicitly considered [3,5,12].

Horizontal permeability was the dominant parameter in the analysis. Its effect is explained by its control on reservoir transmissibility and, therefore, on the ability of the brine displaced by CO₂ injection to redistribute laterally [5,6]. Lower permeability promotes pressure build-up near the injection wells and reduces the maximum injectable mass. Conversely, higher permeability enhances pressure dissipation and increases capacity before the pressure limit is reached. In this sense, permeability is not only a petrophysical property of the reservoir, but also direct control on the operational feasibility of storage.

The BHP limit appears as the second main control. In the base case, 28.5 MPa is used as an operational modelling threshold rather than as a fracture-pressure estimate. Its position below p90, pAllow, and pFrac preserves an explicit pressure buffer to the site-specific geomechanical limits reported for Lopín [28]. The sensitivity range from 26.31 to 30.50 MPa therefore quantifies how different conservative-to-permissive operational criteria affect the inferred dynamic capacity; it does not imply that all tested thresholds are equally acceptable for field operation. A final injection design would require site-specific coupled geomechanical analysis and validation of fault and seal stability. CO₂ density also has a significant influence because it modifies the relationship among injected mass, occupied volume, buoyancy, and pressure response.

5.2. Implications for Site Characterization and Pilot Design

The results have direct implications for early-stage geological CO₂ storage assessment. For systems like Lopín, the priority should not be limited to refining the available pore volume but should also focus on reducing uncertainty in the parameters that control pressure dissipation. Characterization of reservoir permeability and lateral transmissibility should be considered a priority, as these factors condition operational capacity under pressure constraints.

The definition of the BHP limit is also a critical design decision. A more conservative threshold reduces the estimated capacity, whereas a more permissive limit increases the maximum injectable mass. However, such an increase is acceptable only when supported by site-specific geomechanical evidence. In the present model, the 28.5 MPa base-case threshold provides a conservative buffer to the p90, pAllow, and pFrac values reported for Lopín [28]. BHP therefore functions as an operational link between hydraulic simulation and mechanical safety, but it does not replace coupled geomechanical assessment of reservoir deformation, fault stability, or induced-fracturing risk [6,10].

The influence of CO₂ density indicates that fluid properties must also be treated carefully. This study used constant values to explore sensitivity trends, but a more advanced assessment should incorporate pressure-, temperature-, and salinity-dependent properties through equations of state. This would reduce uncertainty in absolute capacity estimation and improve the representation of CO₂ behaviour under reservoir conditions.

5.3. Model Limitations

The VE approach enabled multiple sensitivity scenarios to be explored efficiently, but it has limitations that must be acknowledged. The formulation assumes that CO₂ segregates rapidly toward the top of the reservoir and that the main dynamics can be represented through vertically integrated columns [11,12]. This approximation is suitable for preliminary prospect-scale assessment, providing a more detailed level of analysis than regional-scale screening. However, it may not capture local-scale processes such as vertical channeling, subvertical heterogeneity, complex capillary effects, or migration through discrete fractures.

Faults are represented in a simplified way through structural traces and boundary conditions, but multiphase flow through fault planes and the evolution of fault permeability are not explicitly modelled. This simplification is relevant because the hydraulic behaviour of faults can modify pressure evolution, plume migration, and storage security.

Another important limitation is the absence of explicit geomechanical coupling. Pressure limits allow safety criteria to be incorporated, but they do not replace a complete analysis of reservoir deformation, fault stability, or induced-fracturing risk, and induced seismicity issues [6,10]. Therefore, the results should be interpreted as a preliminary dynamic flow-based assessment, not as a final demonstration of geomechanical safety.

Finally, the extreme permeability case ×4 produced an anomalous numerical response with non-finite pressure. This result should not be interpreted as a robust physical response of the system, but as an indication of the numerical limits of the solver under conditions far from the base case. The main conclusion regarding the dominant role of permeability remains valid, but it should be supported by the overall behaviour of the tested range rather than by this extreme point.

6. Conclusions

This research developed and applied a numerical VE workflow in MATLAB/MRST to evaluate the key parametric controls on dynamic CO₂ storage capacity in the Lopín structure, Ebro Basin, NE Spain. Unlike static volumetric estimates, the approach defines capacity as the maximum mass of CO₂ that can be injected over a 30-year period without exceeding an explicit operational BHP limit, thereby linking the capacity estimate to pressure-constrained injectability.

For the P50 reference configuration, plume evolution is controlled by the C1-C2 well layout, buoyancy-driven migration, structural relief, and lateral pressure dissipation through the open east-west boundaries. The long-term simulation shows that, after 1030 years, most of the CO₂ remains as a mobile phase, whereas residual, structural, and dissolution trapping provide additional immobilization.

The sensitivity analysis shows that dynamic capacity is controlled primarily by horizontal permeability, followed by the BHP limit and CO₂ density. Permeability produced the largest capacity variation, confirming that reservoir transmissibility and lateral pressure dissipation are critical factors for storage feasibility. In contrast, porosity, brine compressibility, nominal injection rate, and residual CO₂ saturation had a secondary influence within the tested ranges.

These findings support the central hypothesis of this work: in fault-bounded saline aquifers such as Lopín, operational capacity is not governed only by available pore volume, but by the hydraulic ability of the system to dissipate injection-induced overpressure. Early-stage geological CO₂ storage assessments should therefore prioritize permeability characterization, lateral connectivity, transparent justification of operational pressure thresholds, and accurate representation of CO₂ thermodynamic properties. For Lopín, coupled geomechanical assessment and improved characterization of fault transmissibility remain essential before the sensitivity-based pressure limits can be translated into field-operational criteria