Pumped hydro is the cheapest grid storage we have, but it needs mountains. What if we used the ocean as the lower reservoir — building a sea-walled lagoon that runs reversible turbines on the tide for baseload, and pumps water out (or in) to store surplus wind and solar? Call it a tidal lagoon battery.

Picture a 100 km² lagoon walled off from the sea — roughly the footprint of Swansea Bay's proposed scheme, scaled up 5×. Average tidal range on the UK coast: ~7 m. The trick is that we don't just harvest the tide; we overdrive the head difference with surplus renewable power.

The storage math

Energy from a hydraulic head is E = ρ · g · V · h. If we pump the lagoon down by an extra 5 m below low tide using surplus wind:

• Volume displaced: 100 km² × 5 m = 5 × 10⁸ m³
• Average head while refilling (to mean sea level): ~2.5 m
• Energy recovered: 1025 kg/m³ × 9.81 × 5×10⁸ × 2.5 ≈ 1.26 × 10¹³ J
• That's 3.5 GWh per pump-down cycle, at ~75% round-trip efficiency → ~2.6 GWh delivered.

Stack that on top of the natural tidal harvest (~400 MW average for a lagoon this size) and you get a hybrid that does both baseload and dispatchable storage. The UK currently has ~30 GWh of pumped hydro at Dinorwig and friends — three of these lagoons match the entire national fleet.

Why the sea wall is the hard part

A 30 km perimeter wall in 15 m of water needs to resist a 5 m differential head plus storm surge. Hydrostatic pressure at the base: P = ρgh ≈ 150 kPa over 15 m, giving a horizontal thrust of ~1.1 MN per meter of wall. That's manageable with a rubble-mound + concrete caisson design (the Dutch Delta Works hit similar loads), but the materials bill is brutal: roughly 40 million tonnes of rock and concrete — about 1% of global annual cement production for one site.

The reversible-bulb turbine bottleneck

You want bulb turbines (like La Rance, France, running since 1966) that pump and generate. To handle 2.6 GWh over a 6-hour discharge, you need ~430 MW of turbine capacity. La Rance's units are 10 MW each, so ~40 turbines per lagoon. Each one is a 7.5 m diameter steel beast costing ~$40M. Turbines alone: $1.6B.

The ecological gotcha

Here's the part the physics doesn't tell you: a 5 m artificial tidal range on top of the natural cycle destroys the intertidal mudflat ecosystem. The lagoon floor cycles between submerged and exposed on a 12-hour clock instead of a tidal one, and salinity stratifies. La Rance saw silt accumulation jump 5× and lost most of its native flatfish within a decade. Any real proposal needs a "minimum sympathetic cycle" mode — probably cutting usable storage by 30-40%.

Does it pencil out?

Capex: roughly $8-12B per lagoon (wall + turbines + grid tie). At 2.6 GWh × 300 cycles/year × $80/MWh arbitrage = $62M/year revenue, plus ~$280M from baseload tidal. Payback: ~25 years. Comparable to nuclear, half the carbon, and the wall lasts 120+ years.