CO₂ storage and subsurface monitoring: closing Part 8

What CO₂ does to a brine sand

Storage sites are chosen deeper than about 800 m. There the pore pressure, close to hydrostatic at about 10 MPa per km, is above CO₂’s critical pressure of 7.38 MPa, and the CO₂ is a dense liquid or supercritical fluid rather than a gas. Figure 8.6, from the Span and Wagner (1996) equation of state, gives 731 kg/m³ at the top of its storage sand, 800 m down on a geotherm of 31 °C per km: about 72 % of the brine’s density. Elastically the CO₂ is still very soft. Its bulk modulus there is 0.077 GPa, 32 times less than the brine’s 2.48 GPa (Batzle and Wang 1992). In shallow, cool basins the temperature sits near CO₂’s critical temperature of 31.0 °C, where a few degrees change its density a great deal.

Gassmann’s equation (Section 5.3) carries these fluid properties into the rock. For the unconsolidated sand of the figure, porosity 0.37 like the Utsira Sand at Sleipner:

• $V_P$ falls steeply, then flattens. Full CO₂ lowers it by 33.0 %. With the CO₂ mixed through every pore the first tenth already takes 25.0 %, 76 % of the whole fall, because the Wood average of the two fluids is ruled by the softer one. Held in separate patches, the same tenth costs only 5.9 %, and the fall grows almost in proportion to the CO₂.
• $V_S$ rises a little. The shear modulus is the frame’s whatever the fluid, and the CO₂ lightens the rock: +2.7 % at full CO₂.
• Density falls in proportion to the CO₂, by 5.3 % at full saturation.
• $Z_P$ falls by up to 36.6 %, most of it with the first tenth of CO₂ when the fluids are finely mixed.

The consequence runs through all of CO₂ monitoring: a plume’s outline is easy to see, because a little CO₂ makes most of the change, but the saturation inside it is hard to read, because more CO₂ changes little and the answer depends on how the fluids share the pores. At Sleipner, where about a million tonnes a year have gone into the Utsira Sand since 1996, the first repeat survey in 1999 already showed the plume as up to nine bright reflections from CO₂ trapped under thin mudstones within the sand, with the reflections beneath it pushed down in time (Arts et al. 2004; Chadwick et al. 2004).

Exercise: from the equation of state to the monitor survey

1. Start from the figure as it opens: the year-10 survey, the section showing the change $\Delta Z_P/Z_P$, the fluids mixed through every pore. The plume lowers $Z_P$ by up to 35.5 % and fills its outline almost evenly, edge included, and the dashed 4 % detection floor runs along that outline. The headline says why: a tenth of CO₂ already makes 76 % of the fall in $V_P$.
2. Switch the pores to Patchy. The core stays strong, 30.4 %, but the fringe fades, and 1.4 % of the CO₂ now lies below the floor, against none before. The same change in $Z_P$ can mean quite different amounts of CO₂.
3. Switch the survey to Leak. CO₂ in the overburden sands lowers $Z_P$ by up to 20.6 % even in patches, and 99.8 % of it clears the floor. Plate (a) shows where it boils: on this geotherm the column crosses the boiling curve near 590 m, and CO₂ that rises above it is a gas of about 200 kg/m³.
4. Back on Year 10, raise the pressure at the injector to 4 MPa. On the brine sand $V_P$ falls 1.9 % but $V_S$ falls 6.8 %: pressure softens the frame, which carries all of the shear stiffness, while CO₂ alone raises $V_S$. In the section a faint fall in $Z_P$ spreads through the sand beyond the plume.
5. With the top of the sand at 800 m, move the geothermal gradient from 25 to 40 °C per km. The CO₂ there goes from a liquid of 792 kg/m³ to a supercritical fluid of 434 kg/m³ as its temperature passes 31 °C: near the critical point, temperature decides how much CO₂ a pore holds.

What a monitor survey can and cannot prove

A time-lapse survey measures change, and only change larger than its repeatability noise: in Figure 8.6, the detection floor. With the fluids in patches the sand needs about 6 % CO₂ before its $Z_P$ falls by 4 %; finely mixed, 2 % CO₂ already lowers it by 11.5 %. A clean monitor above the caprock therefore proves that no accumulation able to beat the floor has formed there, at the survey’s resolution. It does not prove that no CO₂ has moved: a dilute patchy accumulation, a layer thinner than the survey resolves, a pathway a few metres wide, or CO₂ dissolved in the brine can all pass unseen.

Two ambiguities remain inside the reservoir too. Saturation cannot be read from amplitude without knowing how the fluids share the pores, and a fall in $Z_P$ can come from a pressure rise as well as from CO₂. Shear information from AVO or converted waves separates the two, because pressure lowers $V_S$ while CO₂ raises it (Landrø 2001). Monitoring programmes therefore pair seismic with wells, pressure gauges, gravity and surface deformation.

Integrated monitoring workflow

A full CO₂ storage monitoring program integrates ALL the Part 7 + Part 8 workflows:

• Pre-injection (Parts 7 + 8.1 + 8.3): (a) 3D baseline seismic with full QI workflow (inversion + RPT + facies); identify storage reservoir and caprock. (b) Geomechanical model: $\sigma_v$, $\sigma_H$ and $\sigma_h$ maps; fault locations; caprock integrity threshold. (c) Anisotropic velocity model (Section 8.3) for accurate depth imaging. (d) Baseline petrophysics and saturation from any analog wells.
• Injection commissioning: inject slowly under regulatory oversight. Downhole gauges track the pore pressure in real time. Microseismic network listens for induced events.
• Year-1 monitor survey (Section 8.2): first 4D acquisition. Verify that the plume has formed as expected (small, contained). Calibrate the rock-physics model against observed 4D magnitude.
• Year 3-5 monitors: repeat 4D every 2-3 years. Track plume growth. Each monitor updates: plume volume, saturation map, pressure buildup. Integrate with reservoir simulation to forecast future plume evolution.
• Alarm protocols: any 4D signal in the OVERBURDEN (above caprock) triggers investigation. Stop injection, increase seismic coverage, characterize leak pathway, plan remediation.
• Post-injection monitoring (regulatory requirement, 20-50 years): continue 4D surveys at declining frequency (every 5-10 years). Verify CO₂ stability after injection stops. Government transitions liability to state after verification period if no leaks.

What’s at stake: the business and regulatory picture

CO₂ storage is a regulated activity with multiple stakeholders:

• Regulators (the EPA in the United States, national authorities under the EU CCS Directive in Europe) require a monitoring plan as a permit condition: it must track the plume and the pressure front and show that the CO₂ stays in the storage complex. The regulations set the goals, not the method; offshore, repeated 3D seismic is the usual tool.
• Operators carry technical risk and economic liability. Proper monitoring = defensive documentation; leaks = remediation costs + reputational damage + possible criminal liability.
• Insurance carriers require 4D monitoring records to maintain coverage. An injection project without a monitoring plan is uninsurable.
• Carbon markets: stored CO₂ qualifies for carbon credits ($50-200 per tonne depending on market) only if it’s VERIFIED to stay stored. Monitoring, with repeated seismic at its core offshore, is how storage is verified. A leak = the stored carbon is uncredited and the credits must be returned.
• Public: local communities have concerns about groundwater contamination (if CO₂ leaks into drinking-water aquifers) and induced seismicity. Transparent monitoring builds trust.

A 1-million-tonne-per-year CO₂ storage site has carbon-credit revenue of $50M-$200M per year. The 4D monitoring program costs ∼$2-5M per year. The ratio alone justifies the investment. And the cost of a single documented leak (remediation, lost credits, penalties) typically exceeds 10 years of monitoring costs.

Real-world case studies

• Sleipner (Norway, since 1996): world’s first commercial-scale CO₂ storage. About a million tonnes a year into the Utsira Sand. A 1994 baseline, then repeat 3D surveys every two to three years from 1999. Plume clearly imaged from the first monitor as up to nine thin layers under intra-reservoir mudstones. No leakage detected. Reference project for the industry.
• Snøhvit (Norway): storage in the deeper Tubåen Formation. Unexpected pressure buildup after 2 years; monitoring detected this early; injection was redirected to preserve storage integrity. 4D worked.
• In Salah (Algeria, 2004-2011): storage monitored by InSAR surface heave, microseismic and a high-resolution 3D seismic survey in 2009. Detected fault-related pressure communication. Injection suspended in 2011 over caprock-integrity concerns flagged by InSAR and the 3D survey, not because a storage target was met.
• Weyburn-Midale (Canada): CO₂-EOR (enhanced oil recovery) combined with storage. Tens of millions of tonnes stored. 4D monitoring has tracked plume progression over 20 years of injection.
• North Sea Net Zero: the Endurance, Acorn, and several other projects under active development. Expected to inject several million tonnes a year each at full scale. 4D seismic monitoring plans in active permitting.

Common risks and their detection via 4D

• Caprock integrity failure: CO₂ breaches the caprock seal and enters overburden. DETECTION: a fall in $Z_P$ above the caprock larger than the survey’s detection floor (Figure 8.6, the leak).
• Wellbore integrity failure: CO₂ migrates UP an abandoned well (common pathway). DETECTION: localized 4D anomaly offset from injector, aligned with known or suspected well locations. Requires extensive pre-injection wellbore-integrity survey.
• Fault reactivation: pressure buildup from injection triggers slip on a pre-existing fault, opening a leak pathway. DETECTION: microseismic detection (separate monitoring technology) correlated with 4D amplitude changes along fault planes.
• Overpressuring: injection rate exceeds reservoir capacity; pressure buildup threatens caprock integrity. DETECTION: pressure gauges in monitoring wells first; in 4D, a fall in $V_S$ and a smaller fall in $V_P$ through the pressurised sand, reaching well beyond the CO₂, and surface uplift by InSAR, as at In Salah.
• CO₂ dissolving / fixing: over long timescales, CO₂ dissolves in brine and eventually mineralizes. This REDUCES the 4D signal (dissolved CO₂ lowers $Z_P$ far less than free CO₂). DETECTION: 4D amplitude decrease over long time, actually a positive sign (CO₂ is becoming more permanently stored).

The way forward: QI for the energy transition

Parts 0 to 8 have taken 3D seismic interpretation from fundamentals to the frontier. QI as a discipline is not just about finding oil and gas. Increasingly, it’s about:

• CO₂ storage: the largest growth area. CCS projects are being approved worldwide.
• Geothermal energy: reservoir characterization for engineered geothermal systems, fracture-controlled reservoirs.
• Hydrogen storage: salt caverns (simpler) and depleted gas reservoirs (requires QI verification of seal integrity).
• Groundwater management: aquifer characterization, saltwater intrusion monitoring, managed aquifer recharge.
• Critical minerals exploration: copper, lithium, nickel, many are found in deposits that seismic can help characterize.
• Construction-ready subsurface imaging: tunnels, underground storage, large foundations, nuclear waste repositories.

The QI workflow you’ve learned, data → inversion → rock-property transforms → probabilistic classification → volumetrics → monitoring, is REUSABLE across all these applications. The specific rock physics and mathematics change; the discipline and the respect for uncertainty carry over. Part 9 now runs the whole workflow on six real fields.