Sea saltiness

Why is the ocean salty when nearly every river that feeds it runs fresh?

The schoolroom answer is a bathtub. Rivers carry a trace of dissolved rock; water leaves the ocean by evaporation and the dissolved matter does not; over billions of years the trace accumulates until the sea holds about 35 grams of salt per kilogram of water.

It is a good answer, and it makes two predictions we can test. First, the ocean should still be getting saltier. Second, since the salt came up the rivers, seawater should look like concentrated river water. Do both hold?

Take the second prediction first, because it is the easier one to check. What dominates the dissolved load of the world's rivers? Bicarbonate and calcium, mostly — the products of rainwater, made faintly acidic by carbon dioxide, dissolving limestone and silicate rock. Evaporate river water and you should end up with a calcium-bicarbonate brine.

Seawater is nothing of the kind. It is overwhelmingly chloride and sodium, with sulfate and magnesium behind them; calcium is barely over one per cent of the dissolved solids and bicarbonate is a fraction of a per cent. So the composition does not match the input. Something is removing the abundant river ions and leaving the scarce ones behind.

That reframes the whole problem. The bathtub has drains, and they are of different sizes for different ions. Where do they go?

Calcium and bicarbonate leave in shells. Marine organisms build calcium carbonate skeletons; those skeletons rain down and become limestone. Silicon leaves the same way, in the opal shells of diatoms and radiolarians. Magnesium and potassium have a stranger exit: seawater circulates down through hot young basalt at the mid-ocean ridges and comes back out chemically altered, with magnesium stripped out into the rock — a volume comparable to the whole ocean passes through that plumbing on timescales of order ten million years. Sulfate is reduced by bacteria in oxygen-starved sediments and buried as pyrite, and it is also laid down as gypsum wherever a sea basin dries.

And chloride? Sodium? Their sinks are weak. Sodium is exchanged onto clay minerals and buried with the seawater trapped in sediment pore spaces; chloride mostly leaves as sea spray blown inland, or in the rare, enormous evaporite deposits laid down when a basin is cut off and dries — the Mediterranean's Messinian crisis removed a few per cent of the ocean's entire salt inventory into one hole in the ground.

So here is the answer to why the sea is a sodium chloride solution: not because those ions arrive fastest, but because they leave slowest. Each ion has a residence time — how long an average atom stays dissolved before something removes it — and the ordering spans an enormous range. Aluminium is out within something like a century, silicon within thousands of years, calcium within roughly a million, magnesium in the tens of millions, sodium and chloride slower still. The ocean's composition is a ranking of removal rates, read off backwards.

There is one further wrinkle, and it undermines the river story more sharply. There is not enough chlorine in ordinary crustal rock to account for the ocean's chloride at all. William Rubey's argument in the 1950s — that chlorine, like water and carbon dioxide, is an "excess volatile" degassed from the mantle in volcanic emissions rather than weathered out of rocks — remains the standard account. The chloride did not come up the rivers. And a good deal of the chloride that does come down rivers today is sea salt blown ashore and returned, not new salt at all.

Which disposes of the first prediction too. Once the drains are in view, the ocean looks less like a filling tub and more like a basin at steady state, with input and removal roughly in balance and salinity held within limits for a very long time. That claim needs care, though: the evidence from evaporite beds and from brines trapped inside ancient salt crystals shows that major-ion ratios — magnesium against calcium especially — have swung back and forth over the Phanerozoic with the vigour of seafloor spreading, and total salinity has probably drifted somewhat. What the record excludes is the runaway rise that pure accumulation predicts.

The accumulation story is not wrong so much as radically incomplete. Evaporation genuinely concentrates: it is why a landlocked basin like the Dead Sea is far saltier than the ocean, and why salinity varies across the open ocean's surface, high under the dry subtropical belts and low under the rainy tropics and near river mouths and melting ice. But evaporation cannot explain which salts are present, and that is the question the composition mismatch forces on us.

Two honest limits on the account above. Residence times are computed as the reservoir divided by an estimated flux, and the fluxes — especially hydrothermal exchange and pore-water burial — carry substantial uncertainty, so the numbers are order-of-magnitude claims rather than measurements. And how nearly constant salinity has been across deep time is genuinely argued; the constraint is real, but it is inferential, and the Precambrian ocean may have been considerably saltier than today's.

• How trapped brines inside ancient salt crystals let geochemists sample seawater from hundreds of millions of years ago.
• Why sea-surface salinity is now used as a rain gauge for the parts of the ocean nobody measures.

Every term the collection defines is gathered in the glossary.