Cave formation

How does ordinary water hollow a cave out of solid limestone?

Drop a chip of limestone into a glass of distilled water and wait. Nothing happens worth noticing; calcium carbonate is very nearly insoluble in pure water. Yet whole cave systems — kilometres of passage, chambers the size of cathedrals — are cut into limestone by rainwater, which is mostly pure water. Something must be turning a solvent that does not work into one that does. What is it, and where does it come from?

Rain is not quite pure. Falling through the air it dissolves a little carbon dioxide, and dissolved carbon dioxide makes carbonic acid. Weak is enough: acid attacks carbonate, converting insoluble calcium carbonate into calcium and bicarbonate ions, which simply travel away in the water.

So how much acid does the rain carry? Atmospheric carbon dioxide is a little over four hundred parts per million — a fraction of a percent. Run the numbers on that and you can dissolve some limestone, but nowhere near enough to explain a cave. Where does the rest of the dissolving power come from?

Here is the step that surprises most people. The water does not go straight from sky to rock; it soaks through soil first, and soil air is nothing like the atmosphere. Roots respire, microbes decompose litter, and carbon dioxide accumulates in the pore spaces to concentrations typically measured in whole percent — often tens of times atmospheric, sometimes a hundred times. Water equilibrating with that emerges enormously more aggressive than rain ever was. The cave is being carved by a chemical the biosphere manufactured, which is one reason limestone landscapes in cold or barren regions dissolve so much more slowly than tropical ones.

Now the physical question. Where does the water go? Limestone is not porous like sandstone; it is dense rock cut by joints, fractures and bedding planes, and water uses those. Here a feedback appears: a fracture carrying slightly more flow is dissolved slightly wider, which lets it carry more flow still. Of thousands of hairline fissures, a few come to dominate and the rest are abandoned.

But why does the water not exhaust itself in the first centimetres, leaving a corroded rind and no cave? Because the dissolution rate collapses as the water approaches saturation — the last few percent of capacity is spent extraordinarily slowly. That near-saturation sluggishness is precisely what lets water still capable of a little work travel hundreds of metres into the rock before giving it up, and it explains the long slow initial phase, fractures widening imperceptibly for thousands of years, followed by a comparatively abrupt breakthrough once a continuous route is open.

Two shapes of passage follow from where the water table sits. Below it the rock is fully flooded and dissolution proceeds on all surfaces, producing rounded tubes; once a valley cuts down and the water table falls, the passage drains and a stream along its floor incises a narrow canyon instead. Read a passage's cross-section and you are reading the history of the water table.

And then the reverse. Water still charged with soil carbon dioxide and saturated with calcium seeps into an air-filled chamber whose air holds far less of that gas. The dissolved gas escapes, the acid weakens, the water is suddenly holding more calcium than it can, and calcite comes out of solution — at the drip point on the ceiling, and where the drop lands. Stalactites and stalagmites are the same equation running backwards.

Three refinements are worth holding onto.

First, the acid is not incidental — it is the entire mechanism. Nothing here is water wearing rock away mechanically. Streams in caves do abrade their floors, but the passage they run in was opened chemically, and abrasion could not have got started without it.

Second, the dominant carbon dioxide source being biological rather than atmospheric ties cave development to climate and vegetation, and is why dissolution rates differ so sharply between regions with the same rock.

Third, an honest exception: not all caves form this way. Some of the largest — Lechuguilla and parts of the Carlsbad system in New Mexico — were dissolved from below by sulfuric acid, generated when hydrogen sulfide rising from petroleum reservoirs oxidised at the water table. Different acid, different direction of flow, different mineral leftovers; specialists call it hypogene to distinguish it from the rainwater-driven case described here.

• How the layering in a stalagmite records past climate, and what makes it datable.
• Why sinkholes, dry valleys and disappearing streams tend to occur together in limestone country.

Every term the collection defines is gathered in the glossary.