Putting The Carbon Back

AThe idea of putting carbon dioxide back underground is older than the anxiety that made it urgent. Engineers in the American oil industry were pumping the gas into ageing reservoirs in the 1970s in order to loosen the crude that remained, and only in the following decade did anyone propose doing the same thing for the sake of the atmosphere rather than the barrel. Carbon capture and storage, as the practice is now called, separates the gas from an industrial exhaust stream, compresses it until it behaves more like a liquid than a vapour, and drives it a kilometre or more below the surface into porous rock sealed by an impermeable layer above. The first project designed for storage alone began injecting beneath the Vestoy field in the North Sea in 1998, and has taken about 1.1 million tonnes a year ever since. Whether a record of that kind licenses the far larger enterprise now being planned is what divides the field.

BZara Haugland, a geochemist at the Nordvik Institute for Subsurface Studies, holds that the geological half of the problem has been settled by measurement rather than by argument. Her group re-shot seismic surveys over Vestoy nine times between 2001 and 2022, tracing the injected plume as it spread sideways beneath the caprock, and set each survey against what the reservoir models had predicted a decade earlier. The plume has stayed within the storage formation in every survey; the models, however, put it in the wrong place, having underestimated its lateral spread by an average of 21 per cent. Haugland treats the two findings as separate. The seal, she argues, has done what a seal is asked to do, and it did so in spite of a modelling error large enough to matter, which is the strongest evidence available that containment does not depend on forecasting the plume correctly.

CThe claim comes with a boundary that Haugland states more often than her critics repeat it. Containment of this quality, she notes, has been demonstrated only in formations overlain by thick marine shale of the kind that lies above Vestoy, and it is precisely where the geology is less obliging that storage will have to expand if the technique is to matter at all. In her 2021 review of injection sites worldwide, the eleven that recorded measurable leakage were, without exception, places where an old exploration well had been driven through the seal and plugged afterwards with cement. The rock itself failed nowhere. Haugland therefore describes the hazard as an inheritance from the drilling era rather than a property of storage, and recommends that regulators require the position of every legacy well to be established before an injection licence is granted.

DPriya Venkataraman, an energy economist at the Calder School of Public Policy, does not dispute the seismic record and says so explicitly. Her objection is that a technology is not defined by its best-behaved installation. Capture equipment, not storage, consumes between fifteen and twenty-five per cent of the output of the plant it is fitted to, a penalty that falls due in every hour the plant runs, and world capture capacity in 2023 stood at 49 million tonnes a year against the 1.2 billion that the mid-century scenarios assume. A method that has grown at this rate for twenty-five years, she argues, is being credited with a future its own history does not support. She concedes two points. Her cost figures come from power stations rather than from cement and steel, where no substitute process exists, and her scenarios assume no policy change of the kind that transformed the price of solar generation.

EThe dispute is sharpened by what governments have made of it. Tax credits introduced in 2022 pay a fixed sum for each tonne stored, and the sum is paid whether the captured gas comes from a cement kiln or from a gas turbine built in the same year. Venkataraman's complaint is not that the credits are wasteful but that they are read as a promise: an operator who can point to a capture plan is able to defer the harder decision about whether the plant should be built at all. Haugland, who might be expected to resist the argument, accepts it in a narrower form, remarking that the case for storage is strongest exactly where emissions cannot be avoided and weakest where they can. Both regard the habit of quoting a project's design capacity rather than the tonnage it has actually injected as the most misleading practice in the literature.

FWhat is emerging is a division of labour rather than a verdict. Where a process releases carbon dioxide for chemical reasons that no change of fuel can remove, and the roasting of limestone in a cement kiln is the standard example, storage has no competitor, and the energy penalty is accepted as part of the price of the product. Where the gas comes instead from burning something to make electricity, the comparison is with generation that never produces it at all, and there the arithmetic has moved steadily against capture since 2015. The question neither researcher can answer is what becomes of a site once the operator has gone: liability passes to the state after a fixed interval in most jurisdictions, and no storage project anywhere has yet reached the end of that interval.