Carbon dioxide can be separated from an industrial gas stream, compressed, moved through a pipeline or ship—and still not deliver durable climate value until the molecule reaches a place designed to keep it away from the atmosphere for a very long time.
That place can be deep underground.
Geological carbon storage is the engineered injection of carbon dioxide into suitable deep subsurface rock formations where physical, chemical and hydrodynamic processes retain the CO₂ and where monitoring is used to verify containment over time.
The idea is not to create an underground gas balloon. The goal is to use pore spaces already present in rock, beneath sealing formations, under conditions where carbon dioxide can occupy the reservoir and become progressively trapped through several mechanisms.
Wait, what? The underground storage space is inside rock
A sandstone can look solid in your hand and still contain a large network of microscopic pores.
Those pores may contain salty water, oil, natural gas or other fluids. Geological storage replaces part of the existing pore fluid with injected CO₂ while managing pressure and fluid movement.
The most important reservoir properties are therefore not “Is there a cave?” but:
- how much pore space exists;
- how well pores connect;
- how fluids move through the rock;
- whether a competent sealing layer overlies the reservoir;
- whether wells or faults create leakage pathways;
- how much pressure the formation can accept safely.
The direct answer
A geological carbon-storage project works through this sequence:
- identify a potentially suitable deep formation;
- characterise geology, pressure, temperature, fluids, faults and existing wells;
- estimate storage capacity and injectivity;
- design injection and monitoring wells;
- establish a baseline before injection;
- inject conditioned CO₂ at controlled rates;
- monitor pressure and plume movement;
- manage wells, faults and reservoir pressure;
- verify that the CO₂ remains contained;
- close wells and continue post-injection monitoring until regulatory closure criteria are met.
The project therefore begins years before the first tonne is injected and can continue long after injection stops.
Storage versus disposal
Calling the process “disposal” can hide the engineering discipline.
A storage project needs:
- site selection;
- reservoir modelling;
- pressure management;
- well design;
- monitoring;
- accounting;
- closure planning;
- long-term responsibility.
The objective is controlled containment with evidence.
Deep saline formations
Deep saline formations are porous rock layers containing salty water unsuitable for ordinary drinking-water use.
They are important because they can be geographically widespread and can offer very large theoretical storage capacity.
The reservoir rock must provide sufficient pore space and flow capacity. A sealing formation—often a low-permeability rock such as shale or evaporite—must limit upward migration.
Saline storage is therefore a reservoir-plus-seal system, not just a porous layer.
Depleted oil and gas reservoirs
Depleted hydrocarbon reservoirs have already demonstrated an ability to hold fluids over geological time.
They often come with extensive geological data and existing wells.
Those advantages are balanced by complications:
- legacy wells can become leakage pathways;
- pressure history may be complex;
- existing infrastructure may not meet CO₂ service requirements;
- field boundaries may be smaller than major saline formations.
A reservoir that safely held oil does not automatically become a ready-made CO₂ store without new integrity analysis.
Unmineable coal seams and other formations
Other geological storage concepts include certain coal formations, basaltic rocks and reactive formations where carbon dioxide can interact strongly with minerals.
The suitability of each depends on geology, permeability, chemistry, access and regulatory framework.
There is no universal storage formation. Site-specific evidence decides.
Porosity
Porosity is the fraction of rock volume occupied by pore space.
Higher porosity generally provides more potential fluid storage per unit rock volume.
But porosity alone is not enough.
A highly porous rock can still have poorly connected pores, limiting injectivity.
Permeability
Permeability measures how readily fluids can move through connected pore space.
High reservoir permeability can support higher injection rates at lower pressure increase.
Very low permeability can limit commercial injection even when theoretical pore volume is large.
Capacity and injectivity are therefore different questions:
- capacity: how much CO₂ might the formation ultimately hold?
- injectivity: how quickly can CO₂ be put into it safely?
Caprock
Caprock is the sealing formation overlying the storage reservoir.
It needs very low permeability and sufficient mechanical integrity to resist fluid migration under storage conditions.
Site characterisation asks:
- how thick is the seal?
- is it laterally continuous?
- are faults present?
- what is its capillary entry pressure?
- has it been penetrated by old wells?
- how will it respond to pressure change?
The reservoir stores. The caprock contains.
Why depth matters
At sufficiently high pressure and appropriate temperature, carbon dioxide becomes dense.
Dense CO₂ stores much more mass in a given pore volume than low-pressure gas would.
This is one reason geological storage targets deep formations rather than shallow subsurface cavities.
Depth also provides thicker overburden and separation from shallow groundwater resources.
Supercritical and dense CO₂
Under many deep-reservoir conditions, CO₂ exists as a dense fluid with gas-like mobility and liquid-like density.
Its exact behaviour depends on pressure, temperature and impurities.
Reservoir simulators therefore use equations of state that represent CO₂-rich mixtures rather than assuming ordinary ideal-gas behaviour.
Injection wells
Injection wells provide the controlled path from surface facilities into the storage formation.
A well includes:
- steel casing;
- cement;
- tubing;
- packers and seals;
- wellhead valves;
- pressure and temperature instrumentation.
Materials must tolerate CO₂-rich environments, impurities and any acidic fluids that can form when CO₂ contacts water.
Well integrity
Wells are engineered penetrations through the caprock.
That makes them one of the most important potential leakage pathways.
Well integrity depends on:
- casing condition;
- cement bonding;
- corrosion control;
- seal performance;
- pressure limits;
- abandonment quality.
New wells can be designed specifically for CO₂ service. Old wells require investigation because historical construction standards may differ from modern storage requirements.
Legacy wells
Depleted oil and gas fields can contain dozens or hundreds of historical wells.
Some may have incomplete records. Some may have been abandoned using old practices.
A storage project therefore maps every known penetration through the seal, assesses integrity and remediates wells where necessary.
The highest-quality reservoir can still be weakened by a poor old well.
Injection pressure
CO₂ must enter the reservoir at sufficient pressure to overcome formation pressure and flow resistance.
But pressure cannot rise without limit.
Excessive pressure can:
- fracture the caprock;
- reactivate faults;
- reduce injectivity;
- push brine into unwanted areas;
- stress wells.
Pressure management is therefore one of the central constraints on storage capacity.
Pressure footprint can be larger than the CO₂ plume
CO₂ occupies a particular region of the reservoir, but injection pressure can propagate much farther through connected fluids.
This means site boundaries cannot be defined only by where the CO₂ molecule is expected to travel.
Pressure interactions with neighbouring wells, formations and permits can matter over a wider area.
Brine displacement
Injecting CO₂ into a saline formation displaces existing brine.
The brine moves through connected pore space according to pressure gradients and formation geometry.
Models therefore track both CO₂ and water.
Large-scale storage is a multiphase-fluid problem, not simply “put gas into empty rock”.
Pressure management through brine production
Some storage concepts produce formation brine from separate wells to relieve pressure and improve injectivity.
This can increase usable storage capacity but creates another system:
- brine handling;
- treatment;
- disposal;
- water chemistry;
- additional wells and pumps.
Pressure management therefore trades subsurface benefit against surface complexity.
The CO₂ plume
Injected CO₂ forms a plume within the pore network.
Because CO₂ can be less dense than formation brine, buoyancy tends to move it upward within the reservoir until the caprock or internal barriers restrict further rise.
Permeability variations, faults, layers and injection rate shape the plume geometry.
Reservoir simulation predicts where the plume should move; monitoring tests whether reality matches the model.
Structural trapping
Structural trapping occurs when buoyant CO₂ accumulates beneath an impermeable or very low-permeability seal.
This resembles the geological structures that trap natural gas and oil.
Structural trapping can retain large volumes early in the storage period.
Residual trapping
As CO₂ moves through pore space and brine later moves around it, small disconnected blobs of CO₂ can become immobilised by capillary forces.
This is residual trapping.
Once disconnected, the CO₂ cannot migrate as a continuous plume easily.
Solubility trapping
Carbon dioxide can dissolve into formation brine.
Dissolved CO₂ no longer exists as a separate buoyant free phase.
This generally strengthens containment over time, though it changes water chemistry and can drive mineral reactions.
Mineral trapping
Carbon dioxide dissolved in water can react with minerals in the rock.
Under favourable geochemistry, part of the carbon can eventually become solid carbonate minerals.
Mineral trapping can be extremely durable but may occur over longer timescales depending on rock type and fluid chemistry.
Trapping strengthens through time
Early storage may depend strongly on structural trapping beneath caprock.
Over time, residual, solubility and mineral trapping can immobilise increasing fractions of the stored CO₂.
This means storage security can improve as the plume equilibrates—provided the site was selected and operated correctly.
Site characterisation
A storage site begins as a geological hypothesis.
Characterisation turns that hypothesis into an evidence base using:
- seismic surveys;
- well logs;
- core samples;
- pressure tests;
- fluid samples;
- geomechanical analysis;
- fault mapping;
- historical production data where available.
The objective is to reduce uncertainty about capacity, injectivity, seal quality and leakage pathways before large-scale injection begins.
Seismic imaging
Seismic surveys use controlled energy sources and receivers to image subsurface rock structures.
They help identify layers, faults, structural traps and reservoir continuity.
Repeat seismic surveys can sometimes detect changes associated with injected CO₂, providing a way to track plume evolution over time.
Well logs and cores
Well logs measure physical properties along the borehole. Core samples provide actual rock for laboratory testing.
These data can reveal:
- porosity;
- permeability;
- mineralogy;
- capillary behaviour;
- mechanical strength;
- fluid chemistry.
A model becomes credible when it is anchored to measured rock.
Baseline monitoring
Before injection begins, operators measure the natural state of the site.
A baseline can include:
- groundwater chemistry;
- soil-gas composition;
- natural seismicity;
- reservoir pressure;
- surface deformation;
- background atmospheric CO₂;
- existing well conditions.
Without a baseline, future changes are harder to interpret.
Monitoring during injection
Injection monitoring can use:
- wellhead pressure and flow;
- downhole pressure and temperature;
- seismic surveys;
- microseismic arrays;
- gravity methods;
- electromagnetic methods;
- tracers;
- groundwater sampling;
- satellite or surface-deformation measurements.
No single monitoring technology answers every question. Strong programmes combine several independent lines of evidence.
Model versus measurement
Reservoir models predict pressure and plume movement.
Monitoring data then test the prediction.
If the plume moves differently from expected, the model is updated and operating conditions can change.
This is not model failure. It is the normal scientific loop:
predict → measure → compare → revise.
Measurement, monitoring and verification
Storage projects often describe monitoring under the broader idea of measurement, monitoring and verification.
The programme needs to show:
- how much CO₂ was injected;
- where it is moving;
- how reservoir pressure is evolving;
- whether wells remain intact;
- whether leakage indicators appear;
- whether the site behaves within the approved model envelope.
Verification converts a storage claim into an evidenced storage claim.
Leakage pathways
Potential leakage pathways include:
- poorly sealed wells;
- faults or fractures;
- caprock defects;
- unexpected permeable layers;
- surface equipment failures.
Site selection and pressure management aim to make these pathways unlikely. Monitoring aims to detect problems early if they occur.
Leakage is not one binary event
Leakage can mean different things:
- movement out of the intended reservoir but still trapped deeper underground;
- movement into an overlying formation;
- movement toward groundwater;
- release through a well;
- release to the atmosphere.
Risk assessment should distinguish these outcomes because consequence and remediation differ.
Groundwater protection
Storage formations are selected deep below potable groundwater systems.
Monitoring can track groundwater chemistry to detect unexpected changes.
The main concern is not that deep saline storage water itself becomes drinkable. It is preventing pressure or fluids from creating pathways into shallow protected aquifers.
Induced seismicity
Changing subsurface pressure can alter stresses on faults.
If a fault is near failure, pressure changes can contribute to small or, in adverse cases, larger seismic events.
Storage projects therefore map faults, limit pressure and monitor seismicity.
Injection rate is not chosen only for commercial throughput. It is constrained by geomechanics.
Traffic-light systems
Some injection operations use threshold systems that trigger additional review, reduced injection or shutdown if seismic or pressure indicators exceed predefined levels.
The general principle is adaptive control: monitor the reservoir and change operation before a small warning becomes a large problem.
Storage capacity is not one number
Storage estimates often move through several levels:
- theoretical pore volume;
- effective capacity after geological constraints;
- practical capacity after pressure and well constraints;
- commercial capacity after economics and permitting.
A basin can look enormous on a geological map while only part of that capacity is practically injectable on the timescale a project needs.
Injectivity can dominate economics
If one well can inject large volumes safely, storage cost per tonne can fall.
If the reservoir requires many wells, drilling and operating costs rise.
High theoretical capacity with poor injectivity can therefore be less valuable than a smaller but highly injectable reservoir.
Storage hubs
Large storage sites can serve many capture projects through shared transport infrastructure.
A hub can spread appraisal, wells, monitoring and offshore infrastructure across greater volume.
But the hub must reserve injection capacity contractually and physically.
Too many capture projects chasing insufficient injectivity creates a carbon-storage version of grid congestion.
Onshore versus offshore storage
Onshore storage can reduce transport distance where industrial sources and geology align.
Offshore storage can offer large sedimentary basins and can reduce direct interaction with populated land areas.
Offshore projects require platforms, subsea wells, marine pipelines, ships or combinations, increasing logistics and maintenance complexity.
Enhanced oil recovery is not identical to dedicated storage
Carbon dioxide can be injected into oil reservoirs to increase oil recovery.
Some injected CO₂ remains underground.
But a climate assessment must account for the oil produced, the CO₂ source, recycle, storage permanence and project boundary.
Dedicated geological storage has a clearer objective: maximise secure storage rather than hydrocarbon production.
Closure
Injection eventually ends.
The project then enters closure and post-closure phases.
Wells are plugged according to approved procedures. Monitoring continues. Reservoir pressure declines or redistributes. The operator demonstrates that the plume behaves predictably and that containment remains secure.
Post-injection monitoring
Storage responsibility does not stop when the compressor stops.
Post-injection monitoring provides evidence that:
- pressure is stabilising;
- the plume remains within expected boundaries;
- wells remain secure;
- no unacceptable leakage occurs;
- models remain consistent with observations.
The duration depends on regulatory requirements and site behaviour rather than one universal number.
Permanence
Permanence means the stored carbon is expected to remain out of the atmosphere for a duration appropriate to climate accounting.
Geological storage is attractive because suitable formations can retain fluids over geological timescales and because multiple trapping mechanisms can strengthen containment.
Permanence is therefore a probabilistic engineering claim supported by geology, design, monitoring and regulation—not a promise made without evidence.
Long-term liability
Who is responsible decades after injection?
Jurisdictions answer this differently. Some frameworks allow responsibility to transfer after closure criteria are met. Others retain operator obligations longer.
Liability design affects financing because investors need to know when and how long obligations persist.
Monitoring cost versus confidence
More monitoring provides more information but costs money.
The strongest monitoring programme is not the one with the greatest number of instruments. It is the one that can answer the important risk questions with independent evidence.
Monitoring should be risk-based and adaptive.
Storage economics
Storage cost includes more than drilling.
- geological appraisal;
- seismic surveys;
- injection wells;
- monitoring wells;
- surface facilities;
- compression;
- offshore platforms or subsea systems;
- monitoring;
- closure;
- financial assurance.
High-quality geology reduces cost because fewer wells and less pressure management may be required.
Storage resource development is infrastructure development
The International Energy Agency’s 2026 CCUS work emphasises that capture, transport and storage projects increasingly need coordinated business models and risk allocation.
A storage resource must therefore move through stages:
- regional screening;
- site appraisal;
- permit development;
- commercial commitment;
- well drilling;
- injection;
- monitoring;
- closure.
Calling a basin “large” does not make storage capacity available tomorrow.
Singapore and regional geological storage
Singapore’s dense industrial economy has limited domestic geological storage opportunity relative to large sedimentary countries.
That makes regional storage chains important.
Captured industrial CO₂ can be conditioned in Singapore, moved by ship or pipeline to suitable formations elsewhere in the region, then injected under the host jurisdiction’s storage framework.
The geological question therefore extends beyond national borders: which regional formations have the right reservoir, seal, injectivity, monitoring framework and legal structure?
This makes Singapore a useful example of the distinction between where CO₂ is produced and where geology allows it to be stored well.
Worked example 1: saline formation
A coastal industrial hub identifies a deep porous sandstone overlain by thick shale.
Appraisal wells confirm good porosity and permeability. Pressure tests show strong injectivity. Seismic data show no major transmissive fault intersecting the planned plume.
Injection begins conservatively. Monitoring confirms the plume remains within the model envelope. Rates increase gradually.
The storage project expands because evidence supports expansion.
Worked example 2: depleted gas field with legacy wells
A depleted gas field has excellent historical production data and proven caprock.
It also has forty old wells.
Several wells show uncertain cement records.
The storage programme remediates high-risk wells before injection and installs pressure monitoring near others.
Historical infrastructure becomes both an information advantage and a leakage-management liability.
Worked example 3: pressure-limited reservoir
A formation has enormous pore volume but pressure rises quickly during injection.
Theoretical capacity looks huge; practical injectivity is poor.
Engineers add brine-production wells to relieve pressure, but this creates water-treatment cost.
The example shows why pore volume alone does not determine commercial capacity.
Worked example 4: offshore storage hub
Several countries ship liquefied CO₂ to one offshore basin.
A receiving terminal meters each cargo, buffers CO₂ and sends it through offshore pipelines to multiple injection wells.
Storage capacity is allocated contractually. Monitoring data are shared with regulators and customers.
The storage formation becomes regional infrastructure much like a shared energy terminal.
Failure mode: capacity estimated from pore volume alone
A project assumes every pore can be filled with CO₂.
Pressure and displacement constraints make most of the theoretical volume unusable.
Repair: estimate effective and practical capacity with reservoir simulation and injectivity data.
Failure mode: excellent reservoir, poor caprock
A porous formation can accept CO₂ easily, but the overlying seal is fractured or discontinuous.
Repair: treat reservoir and seal as one storage system.
Failure mode: legacy wells ignored
A depleted field has dozens of old wells that were never included in the risk model.
Repair: map, inspect and remediate historical penetrations through the seal.
Failure mode: injection rate chosen by compressor capacity
The surface plant can deliver more CO₂ than the reservoir can accept safely.
Repair: let pressure and geomechanics set the injection envelope.
Failure mode: monitoring without baseline
Groundwater chemistry changes slightly after injection, but no pre-injection data exist.
Repair: establish baseline variability before operations.
Failure mode: one monitoring technology
A project relies only on seismic surveys to prove containment.
Repair: combine independent methods appropriate to plume, wells, pressure and environmental pathways.
Failure mode: injection stops, monitoring stops
The project treats the final injected tonne as the end of responsibility.
Repair: plan closure and post-injection verification before first injection.
Failure mode: EOR storage claim ignores produced oil
Injected CO₂ is counted as climate mitigation without accounting for additional hydrocarbons produced and used.
Repair: use a complete lifecycle boundary and distinguish storage from oil-production objectives.
Common misconceptions
- Geological storage does not require giant underground caves.
- CO₂ is stored in pore spaces inside rock.
- Reservoir porosity and permeability are different properties.
- A storage site needs both a suitable reservoir and a competent seal.
- Theoretical capacity is larger than practical commercial capacity.
- Pressure can limit storage before pore volume does.
- Wells can be more important leakage pathways than intact rock.
- CO₂ becomes progressively trapped through structural, residual, solubility and mineral mechanisms.
- Monitoring tests models; it does not replace site characterisation.
- Storage responsibility continues after injection ends.
- Depleted oil and gas fields are not automatically storage-ready.
- Geological storage is not the same as enhanced oil recovery.
A universal geological-storage audit
- Screen regional geology.
- Identify reservoir and caprock.
- Map faults and all known wells.
- Acquire seismic data.
- Drill appraisal wells where necessary.
- Measure porosity, permeability and fluid chemistry.
- Measure initial pressure and temperature.
- Estimate theoretical, effective and practical capacity separately.
- Test injectivity.
- Model pressure footprint.
- Model CO₂ plume movement.
- Assess caprock geomechanics.
- Assess legacy well integrity.
- Define maximum injection pressure.
- Establish environmental and seismic baseline.
- Design injection and monitoring wells.
- Define monitoring methods and thresholds.
- Inject gradually and compare observations with model predictions.
- Update models as data arrive.
- Plan pressure management if needed.
- Define closure criteria before operation.
- Plug wells using approved long-term methods.
- Continue post-injection monitoring.
- Demonstrate stable containment before responsibility transfer where applicable.
The deepest geological-storage principle
The best carbon store is not the formation with the biggest number on a map.
It is the formation that can accept CO₂ at useful rates, keep pressure within safe limits, retain the plume beneath a competent seal, avoid problematic wells and faults, and provide enough monitoring evidence to show that reality matches the model.
Storage is geology converted into verified infrastructure.
How Geological Carbon Storage fits the Energy series
How Carbon Capture Works owns source-side separation. How Carbon Dioxide Transport Works owns movement between source and destination. How Carbon Dioxide Removal Works owns removal accounting and durability across multiple methods. This article owns subsurface storage science: site characterisation, injection, trapping, pressure, wells, monitoring and permanence.
Current evidence and further reading
- International Energy Agency — Carbon Capture, Utilisation and Storage
- International Energy Agency — Financing CCUS at Scale, 2026
- U.S. Department of Energy — Carbon Capture, Utilization & Storage
The final lesson is simple: durable storage is not created by injection alone; it is created by choosing the right geology, controlling pressure, protecting wells and proving containment over time.