John Wang walks the Strategic Petroleum Reserve as a design problem: store hundreds of millions of barrels for decades, keep it safe from attack, release it fast during a supply disruption, and keep maintenance cheap. The interesting part is that both obvious answers fail on the requirements rather than on engineering competence — and the solution that works gets three of those requirements out of one material.

External floating roof tanks, the commercial standard, cap out near 300 feet across and 50 feet tall — about 630,000 barrels each. Full SPR capacity would need roughly 1,130 of them on ~45,000 acres, about the size of Washington, DC. And the seal between a floating roof and the tank shell is a documented lightning-fire point, which makes it an obvious target for shrapnel or an incendiary.

Going underground didn’t solve it either. The Navy’s Red Hill facility at Pearl Harbor tunnelled 20 steel-lined concrete tanks into volcanic rock for 6 million barrels at $42.2M in 1943 dollars (~$820M today). Groundwater contamination — 27,000 gallons in 2014, then releases in 2021 that hit the drinking-water system — forced it permanently closed. Scaling that to SPR size means ~119 Red Hills and hundreds of billions of dollars.

The actual solution uses salt domes in Texas and Louisiana, 60 caverns at about 10 million barrels each:

  • Caverns are mined with water — drill into the dome, inject fresh water, dissolve the salt, pump out the brine
  • Rock salt has extremely low permeability, doesn’t react with petroleum, and slowly deforms under pressure, closing small fractures — so the salt is the container, with no steel-and-concrete liner
  • Oil floats, so pumping fresh water into the bottom displaces crude upward into the delivery system
  • Sitting on the Gulf Coast puts the reserve next to refineries, pipelines, and marine terminals — which is the point when the goal is reaching the market fast

Cost is about $3.50 per barrel of capacity versus $15–18 for aboveground tanks, and depth protects against aerial attack.

The tradeoffs are real and named. Building caverns needs a water supply and somewhere to dispose of brine. Every withdrawal injects fresh water, which dissolves more salt and enlarges the cavern — so cycling is limited and monitoring matters. Wells, pumps, and pipelines still need upkeep.

What makes the design worth reading is that containment, blast protection, and extraction are all consequences of the site rather than hardware bolted onto it, while the failure mode that closed Red Hill — groundwater — is the one the substrate sidesteps. The operating cost it does carry is drift: a hole that keeps growing every time you use it.