IAP-26-125

Protecting Groundwater above CO₂ and Hydrogen Stores: Will Clay-Rich Caprocks Hold?

Underground CO₂ and hydrogen storage could support decarbonisation, but protecting groundwater requires understanding the geological barriers above storage formations. Clay-rich caprocks can resist gas entry because their small, water-filled pores create capillary barriers. Experiments demonstrate substantial sealing capacity for both gases [1,2]. However, changes in fluid pressure can alter the stresses acting on existing fractures, potentially changing their ability to transmit fluids.
This PhD project will investigate whether pressure cycling causes temporary or persistent changes in caprock sealing, and whether those changes differ between CO₂ and hydrogen. The central question is whether a caprock’s previous exposure to gas and pressure changes must be considered when predicting future fluid migration.
The research will combine high-pressure experiments and X-ray imaging with predictive modelling and uncertainty quantification. Experiments on two contrasting clay-rich caprocks will measure gas-entry thresholds, fluid transmission and recovery after pressure reduction. These observations will test competing models describing fixed sealing properties, reversible responses to stress, and persistent changes following gas passage.
Inverse modelling will use measured pressure and flow histories to estimate uncertain properties and evaluate which explanations are supported by the data. Predictions will be tested against experiments withheld from calibration. The resulting models will investigate potential gas and saline-water migration through representative geological sequences towards overlying aquifers.
Groundwater protection provides the environmental focus: CO₂ reaching an aquifer can change water chemistry [3], while saline-water migration presents a separate potential threat. The project will identify how uncertainty in caprock behaviour affects migration predictions and which observations would most improve confidence in environmental assessment.
Key Research Questions
• When does pressure cycling produce reversible or persistent changes in caprock sealing?
• Can pressure, flow and imaging observations distinguish competing explanations of pathway opening and closure?
• How does accounting for pressure history change predictions of CO₂, hydrogen and saline-water migration?
• Which measurements most effectively reduce uncertainty in predictions relevant to groundwater protection?
Expected Outcomes
• Experimentally tested models of gas entry, pathway persistence and seal recovery.
• Quantified uncertainty in inferred caprock properties and predicted fluid migration.
• Evidence identifying when pressure-history effects matter and which observations best constrain them.

1. Rock characterisation and experimental design
Select two caprocks with contrasting clay composition and characterise their mineralogy, texture, water permeability and resolvable fractures. Use initial numerical sensitivity studies to select pressure cycles, recovery periods and measurements that can distinguish alternative explanations of sealing behaviour. Establish a focused experimental programme using intact specimens and selected specimens containing existing fractures.
2. High-pressure experiments and imaging
Test brine-saturated, matched specimens separately with CO₂ or hydrogen under controlled temperature and confining stress. Measure pressure, gas transmission and water displacement during initial gas entry, repeated pressurisation and recovery. Repeat selected conditions to assess reproducibility. Micro-CT observations will quantify resolvable fracture changes; hydraulic measurements will constrain transport through finer pathways. Compare permeability and gas-entry behaviour before and after cycling.
3. Process modelling and inference from experiments
Develop a gas–brine flow model for matrix transport that includes dissolved-gas diffusion, capillary entry, and stress-sensitive flow through existing fractures. Compare models with fixed properties, reversible stress dependence, and persistent changes following gas passage. Use Bayesian or ensemble-based inversion to estimate uncertain entry thresholds, fracture-flow properties, and recovery times from experimental histories. Assess whether the parameters can be distinguished and test predictions against withheld cycles or specimens. Conduct selected additional experiments to address the largest remaining uncertainties.
4. Migration prediction and uncertainty quantification
Incorporate the supported relationships into a representative two-dimensional model of storage formation, caprock and overburden. Propagate uncertainty in inferred properties, fracture connectivity and geological structure into predictions of upward gas/brine flux and potential arrival times at the aquifer base. Evaluate how pressure-history effects alter these predictions compared with fixed-property assumptions. If repeated simulations become costly, use a simulator-trained surrogate, checked against independent simulations, to accelerate uncertainty analysis [4]. Assess which pressure or flow observations most reduce prediction uncertainty. Evaluate limitations in transferring laboratory relationships to geological scales.

Year 1

• Literature review, laboratory and modelling training.
• Sample characterisation, pilot experiments and initial process models.
• Sensitivity analysis to refine experimental design.

Year 2

• Main CO₂ and hydrogen experiments, including pressure cycling and recovery.
• Progressive model calibration and comparison of alternative process descriptions.
• Initial assessment of parameter uncertainty and predictive performance.

Year 3

• Targeted experiments and testing against withheld observations.
• Geological migration scenarios and uncertainty propagation.
• Surrogate development where needed; preparation of manuscripts and thesis chapters.

Year 3.5

• Integrate experimental and modelling findings.
• Complete publications, thesis writing and submission within 3.5 years.

Experimental methods: High-pressure flow testing, mineralogical characterisation, micro-CT imaging and quantitative fracture analysis.
Modelling and data interpretation: Multiphase flow, relevant rock-mechanics principles, inverse modelling, Bayesian/ensemble methods, surrogate modelling and uncertainty quantification.
Research and professional skills: Python-based analysis, reproducible computing, research data management, scientific writing, and communication with environmental audiences. Training will draw on complementary expertise at Glasgow and Heriot-Watt and the wider IAPETUS programme.