A Case For Storing LNG in Dead Wells
A jackup rig sits over four wells in the North Sea that used to produce gas and now inject
carbon dioxide instead. Deep plugs went in first to isolate the old reservoir, then the wells were cleaned, then ultrasonic inspection checked every metre of steel and cement before new tubing and valves went on top. Months of rig time, four wells. The same rig type, run by the same class of marine warranty surveyor who signs off a jackup move anywhere in the world, could sit over depleted wells off the Indian coast doing something almost nobody in India has done before: banking imported gas underground for the day the tap runs dry.
That day may be closer than the planning documents assume, and the number that should worry a reader is not on the table yet.
The physics nobody states plainly
LNG is not gas in a tank. It is methane cooled to roughly minus 162 degrees Celsius until it turns liquid at normal pressure, a state change that shrinks its volume by about six hundred times so a ship can carry enough of it to matter. None of that liquid ever goes underground. A depleted reservoir, a salt cavern, an aquifer, all of them store gas at reservoir pressure and reservoir temperature, which only works once the LNG has been warmed back into a gas at a regasification terminal. Any underground storage scheme built on imported LNG is really two projects stitched together: a terminal that turns liquid into gas, and a pipeline that carries that gas inland to an injection site. Skipping past that seam in a headline is where credibility gets spent for nothing.
Three options, three cost curves
India is weighing three ways to hold a buffer of gas for the day supply breaks. Engineers India Limited is running the feasibility work, and Indian Strategic Petroleum Reserves Limited, which has managed only crude oil reserves until now, has been floated as the eventual operator.
Depleted reservoirs sit at the bottom of the capital cost curve because the wells, the gathering lines and the reservoir characterisation already exist. What they cost instead is time and well integrity work, the kind that ate months of rig time on four wells at Porthos in the Dutch North Sea before a single molecule went in. Salt caverns sit above that on capex per unit of gas held, since the cavern has to be leached out of the salt formation before it holds anything, but they cycle gas faster than a depleted reservoir. The US Energy Information Administration is blunt about the trade: cavern construction costs more dollars per thousand cubic feet of working gas capacity than converting a depleted field, but the ability to inject and withdraw several times a year instead of once a season lowers the effective cost of every unit of gas actually moved through the cavern.
LNG terminal tankage sits at the top of the readiness curve. India already runs import terminals with an aggregate regasification capacity in the range of 48 to 53 million tonnes a year, a range rather than a confirmed single figure because different filings report it slightly differently. Expanding tank capacity at an existing terminal is fast because the jetty, the regasification train and the pipeline connection already exist. What that speed costs in dollar terms is worth putting a number on, because the assumption baked into most of the coverage is that tankage is the cheap option. It is not, on a per unit of storage capacity basis. It is only cheap relative to how fast it can be deployed.
Unit economics, layer by layer
Start with underground storage. A representative US facility, thirty billion cubic feet of working gas capacity, carries a capital cost of roughly nineteen dollars per thousand cubic feet of that capacity, based on cost modelling published by Thunder Said Energy. That prices a facility of that scale at somewhere around five hundred seventy million dollars, and it needs a storage spread of about a dollar fifty per thousand cubic feet between injection and withdrawal prices to clear a ten percent return. That is a US number, built on US labour and US regulatory timelines, and it should be read as a directional benchmark rather than an India specific cost, since no comparable India figure has been published for a project of this kind.
Now put an actual LNG tankage expansion next to it. Southern LNG’s Elba II project in the United States added three point three billion cubic feet of storage capacity to an existing terminal for approximately one hundred fifty seven million dollars, a real, disclosed, completed project rather than a model output. Divide that out and it comes to roughly forty seven dollars per thousand cubic feet of capacity added, more than double the per unit cost of the underground benchmark above. A single large LNG storage tank in the one hundred fifty thousand to two hundred seventy thousand cubic metre range, the size class terminal operators would likely build to meet a new buffer mandate, runs one hundred twenty to two hundred fifty million dollars on its own and takes thirty to thirty six months to build, according to current market cost tracking.
Lay those two numbers side by side and the story most coverage tells gets inverted. Underground storage is not more expensive than tankage. On a dollar per unit of gas capacity basis, tankage is the more expensive option. What tankage actually buys is speed and an existing regulatory shell, since expanding a facility that already holds an operating licence moves faster through approval than drilling and workover consent for a new injection site. India appears to be selecting tankage not because it is cheaper per unit of gas held, but because it is faster to build, and those are different virtues that get flattened into a single word, cheap, whenever this gets summarised in a headline.
The precedent that already exists inside ONGC
The reason the depleted reservoir option is not purely theoretical for an Indian operator is that ONGC is already running the conversion playbook, just for a different gas. The Gandhar CCS pilot in Gujarat will use two abandoned onshore wells to inject roughly one hundred tonnes of carbon dioxide a day into depleted subsurface reservoirs, sourced from industrial units in the Dahej area and from ONGC’s own Hazira gas processing plant, with an earlier phase having sourced CO2 from Indian Oil’s Koyali refinery about eighty kilometres away. The pilot carries a second objective beyond storage: testing whether the injected CO2 can also mobilise remaining oil in the reservoir, turning a disposal problem into a recovery tool. ONGC signed a memorandum of understanding with Shell in December 2022 specifically to study CO2 storage and enhanced recovery screening across Indian basins, and the company plans to bring in a dedicated CCUS consultant for technical design, regulatory guidance and cost assessment on the pilot. No public figure for the pilot’s own capital cost has been disclosed at the time of writing, and that gap should be stated plainly rather than papered over with a borrowed number from a different country’s project.
None of the capital or technical groundwork behind Gandhar was built for strategic gas storage. It was built for climate compliance and incremental oil recovery. But a converted well, a characterised reservoir and a trained workforce do not care what gas they are asked to hold next. The Gandhar work is the closest thing India has to a live rehearsal for exactly the well integrity and reservoir screening questions a gas storage conversion would raise, and the fact that ONGC is the operator running it means the institutional muscle for this kind of conversion already sits inside the company being asked, elsewhere in government, to think about strategic gas reserves.
What history actually shows about storing gas rather than moving it
There is a temptation to read India’s decades of associated gas flaring as evidence that underground storage has already been tried and abandoned. It has not. Bombay High in 1990 was producing more associated gas than the pipeline network could carry, and the response was to shut in high gas oil ratio wells and accept an oil production hit rather than build storage capacity to hold the surplus. The reinjection that Indian offshore fields do practice is overwhelmingly for pressure maintenance and enhanced recovery, pushing gas back into a reservoir to move oil, not banking it to sell later when the price or the emergency justifies withdrawal. Those are different economic instructions to the same wellbore. India has extensive experience disposing of unwanted gas and almost none banking wanted gas for a future date. Anyone citing decades of reinjection experience as proof the storage question is already solved is quietly swapping one problem for the other.
The fork nobody has picked yet
Reported policy direction, sourced from a single account and not yet corroborated against a Petroleum and Natural Gas Regulatory Board filing, points toward the faster option winning by default. Rather than commit to the multi year lead time of underground storage, the Ministry is said to be evaluating a rule that would require LNG terminal operators to hold a buffer equal to ten percent above normal commercial throughput, ring fenced for government requisition during a supply shock and funded through a toll style tariff passed through to customers rather than a budget line. That structure gets India a reserve without the capital outlay of drilling new injection wells or characterising a salt formation, and it gets there in a fraction of the time a depleted field conversion would take.
What none of the reporting on that mandate does is translate ten percent of throughput into days of national gas consumption, which is the only number that actually tells a reader whether the buffer means anything in an emergency. Compare that gap to what mature storage markets disclose without being asked. The United States holds roughly four trillion cubic feet of working gas across four hundred underground facilities, enough to cover about forty five days of national demand at typical draw rates. Europe holds around one hundred ten billion cubic metres of storage, close to one hundred days of demand. Both of those figures come from operators and regulators who treat duration as the headline metric, because duration is the entire point of a strategic reserve. India’s proposed cushion has been described only in percentage terms relative to a terminal’s normal throughput, a framing that makes the mandate sound substantial without letting anyone check it against an actual number of days.
Stress test
Assume the terminal tankage rule passes as reported and the underground storage feasibility work continues at its current administrative pace, meaning years rather than one budget cycle. A regional supply shock lasting beyond the buffer window forces India back onto the spot LNG market during exactly the period when spot prices spike hardest, which is the scenario a strategic reserve exists to avoid in the first place. The cheaper per unit of speed option optimises for the disruption that is short enough to survive without it. The counterintuitive finding, once the actual capex numbers are on the table, is not that underground storage is superior because it is cheaper. It is that tankage is not even the cheap option on a dollar per unit of gas basis, only the fast one, and India appears to be building toward days of cover while citing a supply security rationale that was triggered by a disruption risk measured in weeks, not days.
The close
A reserve that empties before the emergency does is not a reserve. It is a delay wearing a reserve’s name, and paying more per unit of gas held to get that delay faster is not the bargain it is being sold as.
