Screening depleted reservoirs for CO₂ storage: injectivity, capacity and containment

Screening Depleted Reservoirs for CO₂ Storage: Injectivity, Capacity and Containment

California's latest offshore lease sale has put mature acreage back on the table — acreage that produced hydrocarbons for decades and now sits at or near abandonment. If you hold those assets, the old question was simple: plug and walk away. That's not the question anymore. Subsurface teams are being asked whether the reservoir can store CO₂ commercially, and how to answer that before capital goes into an appraisal programme. Get the screen wrong in either direction and it costs you. False positives burn engineering budget on sites that fail detailed assessment. False negatives strand infrastructure that could have underpinned a viable carbon storage project.

This article covers the three governing technical criteria — injectivity, storage capacity, and caprock containment — at the level of detail you need to run a first-pass screen on a candidate depleted reservoir.

Why Depleted Reservoirs Deserve Serious Consideration

Depleted oil and gas reservoirs beat deep saline aquifers on several structural counts. The trap geometry is already proven — a hydrocarbon column sat in it for geological time. You have petrophysical data, well logs, core, and production history. Surface facilities, wellbores, and pipeline rights-of-way are often still in place, which matters when you're looking at the full project cost picture and not just the subsurface.

Callas et al. (2022, Stanford/SCCS) lay out a workflow that starts with readily available data and narrows the candidate list step by step. The logic holds for any operator working legacy assets: spend the cheap data first.

The Three Technical Pillars

Injectivity

Injectivity tells you whether CO₂ can physically get into the formation at the rates a commercial storage project needs. The controls are reservoir permeability, net pay thickness, fluid viscosity contrasts, relative permeability effects, and near-wellbore skin.

In a depleted reservoir, start with the well tests and production data you already have. Permeability-thickness (kh) products from pressure build-up analysis during hydrocarbon production carry over directly to an injectivity estimate, with corrections for the different fluid system. Supercritical CO₂ has a viscosity well below reservoir brine, which usually helps injectivity. But CO₂–brine relative permeability, plus the risk of near-wellbore dry-out or salt precipitation, can pull effective injectivity down over time.

The MDPI Energies review (2024) on wellbore injectivity flags skin damage from scaling, fines migration, and geochemical reactions at the wellbore face as the main operational threats to sustained injection rates in repurposed wells. And existing wellbores in depleted reservoirs often carry legacy damage or cement integrity problems that have to be fixed before CO₂ injection.

A practical first-pass screen should ask:

• Is the kh product, derived from existing well tests, sufficient to sustain target injection rates without exceeding fracture pressure?
• Are legacy wellbores in a mechanical condition that supports re-perforation or re-completion for injection service?
• Does the reservoir pressure depletion provide sufficient injectivity headroom before pore pressure approaches the minimum horizontal stress?

Storage Capacity

Storage capacity sets the ceiling on how much CO₂ a reservoir can take over project life. For a depleted reservoir, the theoretical upper bound is the pore volume that hydrocarbons used to occupy, adjusted for the compressibility of the remaining fluids and rock, and for the fraction of that pore volume CO₂ will realistically contact given sweep efficiency.

Callas et al. (2022) split this into static capacity — the geometric pore volume available — and dynamic capacity, which injectivity, pressure management, and plume behaviour constrain. A reservoir with large static capacity but low permeability may end up with a dynamic capacity that is a small fraction of the theoretical figure.

The IEAGHG report notes that capacity assessment must account for:

• Residual hydrocarbon saturation, which reduces the pore volume available for CO₂
• Aquifer support, which affects pressure build-up and therefore the practical injection volume before pressure limits are reached
• The geometry of the structural closure, which bounds the free-phase CO₂ plume

Pressure management is often the binding constraint in depleted reservoirs. Where reservoir pressure has dropped well below original conditions, there is headroom to inject before reaching fracture pressure. Where a strong aquifer has partially re-pressurised the reservoir, that headroom is reduced. Neither condition can be assumed without checking the current pressure datum against the fracture gradient.

Caprock Containment

Containment is the non-negotiable criterion. A reservoir that cannot retain CO₂ for the regulatory storage period — typically measured in centuries — is not a viable site regardless of capacity or injectivity.

The primary containment mechanism in a depleted reservoir is the same caprock seal that retained hydrocarbons. That's a real advantage: if the seal held a buoyant hydrocarbon column for geological time, it has demonstrated integrity at the relevant scale. CO₂, though, brings geochemical challenges that hydrocarbons don't. Dissolved CO₂ forms carbonic acid, which can react with carbonate cements in the caprock. For a first-pass screen there's a cheaper mechanical check that needs only existing geomechanical data: unconfined compressive strength, cohesion, and friction angle from core or log-derived estimates, combined with the in-situ stress state.

Key containment screening questions are:

• Does the caprock have documented column height capacity exceeding the proposed CO₂ column?
• Are there mapped faults that intersect the caprock within the closure? If so, what is their orientation relative to the current stress field?
• What is the legacy well count penetrating the caprock, and what is the documented cement integrity of those wells?

That last one is usually the hardest in practice. A mature field can have dozens of legacy wells with cement quality records of variable quality. Every penetration is a potential leakage pathway. Callas et al. (2022) treat wellbore integrity as a containment sub-criterion — assessed well by well, not as a field-level average.

Comparison of Screening Criteria: Data Sources and Failure Modes

Illustrative Screening Scenario

The following is illustrative and does not represent a specific field.

Consider a depleted sandstone gas reservoir at depth sufficient for CO₂ to remain in a supercritical state under reservoir conditions. Pressure depletion has pulled reservoir pressure well below original conditions, so there is substantial injectivity headroom before fracture pressure is approached. Production data show a moderate kh product from historical well tests — enough to suggest that a small number of injection wells could sustain commercially relevant rates.

The structural closure is a four-way dip closure with a shale caprock of documented thickness. Column height capacity, derived from capillary entry pressure measurements on core, exceeds the proposed CO₂ column height. But the field has multiple legacy producer wells, and several have incomplete cement bond logs. Those wells are the primary containment risk. They would need pressure testing and remedial cementing before a regulatory application could go anywhere.

The screening outcome: the reservoir passes on injectivity and capacity; containment is conditionally acceptable pending well integrity remediation. That's a realistic result for a lot of mature assets — the subsurface works, but the wellbore stock needs money spent on it.

Practical Screening Checklist

Injectivity

• Pull all available well test data and derive kh for candidate injection intervals
• Compare estimated injection bottomhole pressure against the fracture gradient at the injection horizon
• Assess legacy wellbore mechanical condition for re-completion suitability

Storage Capacity

• Build or update the static volumetric model using existing log and core data
• Establish the current reservoir pressure datum and compare it to original pressure and the fracture gradient
• Identify aquifer connectivity and estimate aquifer support strength

Caprock Containment

• Confirm caprock lithology, thickness, and lateral continuity from existing well and seismic data
• Run a yield envelope or column height assessment using available geomechanical data
• Inventory all legacy wellbores penetrating the caprock; flag those lacking cement bond logs for remedial assessment
• Map all faults within the closure and assess orientation relative to the current stress field

Infrastructure and Regulatory Readiness

• Identify reusable surface facilities, pipelines, and injection wellbores
• Confirm regulatory jurisdiction and applicable storage permitting pathway

Conclusion and Next Steps

Screening a depleted reservoir for CO₂ storage is a desk exercise. You can finish it with data already in the files, before committing any new capital. The IEAGHG and Stanford/SCCS frameworks both point the same way: start with what's cheap and fast — volumetric capacity, existing well test injectivity, caprock column height — then work toward the data-heavy geomechanics and wellbore integrity work.

If you're a reservoir or CCUS engineer holding mature acreage, the first move is to pull the existing well test archive and the abandonment pressure datum, then run the injectivity headroom calculation against the fracture gradient. Pass that filter and the next spend is a caprock seal analysis using existing core and log data. Legacy wellbore integrity is almost always the longest lead-time item, so scope it early.

Sites that clear a rigorous three-pillar screen are genuine optionality for operators working through the energy transition. Not because the economics are guaranteed — they aren't — but because decades of production data bound the subsurface risk in a way a greenfield saline aquifer simply cannot.

Note: the approved draft as supplied contained no fenced json metadata block, so none is reproduced here.