Soil carbon loss under ploughing

Why does turning the soil every year, which looks like feeding it, steadily empty it of carbon?

Ploughing looks generous. You bury the stubble, break the pan, fold last year's residue into the ground — and the following crop responds, visibly, with a flush of nitrogen behind it. Every appearance is of feeding the soil. Yet a field ploughed annually for a few decades holds far less carbon than the grassland it came from, and the drift is one way. So either the appearance is wrong, or "feeding" means something odd here.

Treat the stock as a stock. Soil carbon is not a substance you add or subtract; it is the standing balance of two flows — what plants put in as litter, dead roots and exudates, and what microbes take out by respiring it away as carbon dioxide. If the store falls, either the input dropped or the output rose.

The input does drop, because a crop removed at harvest returns less than a perennial sward dying back in place with living roots all year. But that alone gives a slow decline, and the observed loss is fastest in the first years after breaking ground. So the output surged.

Why would turning the soil make microbes respire faster? The obvious answer is oxygen: a dense soil is patchily anaerobic, and ploughing makes a loose, well-aired tilth. That is real — but hold it lightly, because if it were the whole story, aerating a soil without breaking it would produce the same flush, and it does not.

Ask instead what had been keeping the carbon alive so long. Here the field changed its mind over the last two decades, and the change is the interesting part. The old picture was chemical: humus persisted because it was intrinsically recalcitrant, too gnarled for an enzyme to bite. That has largely given way to a view in which persistence is a property of the situation, not the molecule. A perfectly edible sugar can sit in soil for decades if no microbe can physically reach it.

There are two ways of being out of reach. Occlusion: soil aggregates into crumbs, glued by fungal hyphae, root mucilage and microbial gums, and organic fragments end up sealed inside them, in pores too fine and too poorly connected for bacteria and their extracellular enzymes to work. And mineral association: fine fragments bond onto clay and silt surfaces, chemically available but held.

Now the mechanism assembles itself. A plough shears aggregates apart. Everything occluded is abruptly on the outside, in a well-aerated, freshly mixed medium, among decomposers that were never short of appetite — only of access. The respiration flush follows, and with it the mineral nitrogen the farmer reads as fertility. The plough is not feeding the soil. It is opening the safe.

Worse, it does both things at once: the same pass destroys the fungal networks and crumb structure that would rebuild the protection, and leaves the surface bare to raindrop impact, which crusts and disaggregates it further. Output up, and the machinery for restoring the store degraded.

A third channel, erosion, moves carbon-rich topsoil off the field. Whether it is a net global source is genuinely contested — some of that carbon is buried at the deposition site and preserved — and the loss occurs on flat fields that erode negligibly, so erosion cannot be the engine.

Three folk accounts are worth ruling out properly rather than waving away.

"Ploughing exposes the carbon to air and it oxidises." Not chemically. Soil organic matter does not burn spontaneously at ambient temperature; the oxidation is done by organisms, and what limits them is access and moisture, not the mere presence of oxygen.

"Ploughing kills the soil life." It does harm earthworms and fungal networks — but notice the direction. Killing decomposers would reduce respiration, and tillage's immediate measured effect is an increase. The account has to explain a surge, not a deficit.

"It's the erosion." Secondary, as above, and absent on level ground where the loss still occurs.

The discriminating test, and the reason to prefer physical protection over the oxygen story: take a soil sample already fully aerated, so oxygen supply cannot change, and crush its aggregates. If protection is chemical, little should happen; if physical, a respiration flush should follow the crushing alone — and that is what such experiments show. The same logic explains why loosening a soil without shattering its crumbs does far less damage than inverting it.

Now to falsify my own account. The mechanism above is a story about loss, and it is tempting to run it backwards into "stop ploughing and the carbon comes back". That inference is much weaker. Much of the apparent gain under no-till turns out to be redistribution — carbon concentrated in the top few centimetres where everyone samples, with little or no gain when the profile is measured to depth. Whether no-till is a genuine sink, and under which climates and rotations, remains argued. What survives is the asymmetry: destroying physical protection is fast and rebuilding it slow, so the loss is far better evidenced than the recovery.

On numbers, sparingly. Long-cultivated soils are commonly reported to have lost something like a quarter to a half of their original organic carbon in the first decades of cultivation, but the scatter across soils, climates and sampling depths is very wide, and that range is recalled rather than derived. I decline a per-pass loss figure: it depends on aggregate strength, moisture at tillage and depth of inversion, and no clean general value exists.

• How deep sampling changed the no-till sequestration debate, and what would settle it.

Every term the collection defines is gathered in the glossary.