The Number That Was Mistaken for a Forecast

AThe chemistry has never been the contested part. Since 1850 the ocean has taken up something close to a quarter of the carbon dioxide released by burning fossil fuels, and the average pH of surface water has fallen by roughly 0.1 unit — a small-looking number that, on a logarithmic scale, amounts to a rise of about thirty per cent in hydrogen ion concentration. For animals that build shells and skeletons out of calcium carbonate, the quantity that matters is the aragonite saturation state, written as omega, which falls as the water acidifies; below one, the mineral tends to dissolve rather than accumulate. The first laboratory studies, run in the mid-2000s by transferring animals abruptly into acidified tanks, returned thinner shells, slower larval development and higher mortality, and it was those results that supplied the vocabulary of the earliest policy documents. A second generation of experiments has since made the picture less obliging, and the field has spent fifteen years arguing about what the difference means.

BIlse Brandt, who directs the Southern Shelf Carbon Programme in Hobart, holds that the water sets the limit and that everything else is detail. Her group has sampled three moorings on the Tasmanian shelf every fortnight since 2004, recovering pteropods — small swimming snails whose shells are built from the most soluble form of aragonite — and scoring each shell for dissolution. Over nineteen years the proportion of individuals classed as severely dissolved rose from nine per cent to thirty-one. The rise tracked the number of weeks each winter during which upwelled water drove omega below one, and it did not track the abundance of the phytoplankton the animals eat: in 2016 and again in 2019, both years of exceptional food supply, the dissolution scores were among the worst in the whole series. Brandt is careful about what the index does and does not show. It records the condition of shells at the moment of capture; it says nothing about how many animals were there to be captured.

CThe claim yields a prediction, and Brandt has tested it against a second dataset. At eleven sites elsewhere in the Southern Ocean where continuous sensors show omega has never fallen below one, severe dissolution stayed under four per cent of individuals in every year sampled, whether food was plentiful or scarce. The effect, on her reading, is confined to water that crosses the threshold: above it the animals appear indifferent to the chemistry, below it they are not. She is willing to grant that this leaves the mechanism underdetermined, since a shell may thin because the mineral is dissolving from the outside or because the animal has stopped depositing new material, and her images cannot tell the two apart. What she will not grant is that the threshold is an artefact of the tank, given that no animal in the series was handled before it was measured.

DBenedikt Kirchner, a physiologist at the Gulf of Guinea Marine Institute in Lagos, replies that a limit in the water becomes a limit in an animal only when the animal cannot pay for the repair. Calcification, she points out, is work, and work has to be funded from the same energy budget as growth and reproduction. In ninety-day trials her group held mussels at an omega of 0.9 and varied only the ration: well-fed animals kept calcification within four per cent of controls held at present-day chemistry, while animals on a restricted ration lost twenty-two per cent of shell mass. She adds that the coastal sites her animals come from already swing through 0.4 of a pH unit in a single day, a range wider than the change projected for the open ocean by 2100. Two limits she states herself. Ninety days is a fraction of a mussel's life, and the mussel is among the hardiest calcifiers anyone has tested.

EThe dispute stopped being academic around 2012, when hatcheries on the Pacific coast of North America began losing oyster larvae in whole batches and traced the failures to intake water drawn during upwelling. Several now treat that water with sodium carbonate before it reaches the tanks, and monitor omega as routinely as they monitor temperature. Brandt observes, awkwardly for her own case, that a hatchery can dose a tank while nobody can dose a shelf, so the remedy her chemistry implies exists only where the volume is small enough to treat. Kirchner answers that this is exactly why the energy side deserves the funding, since a fishery can be managed for food supply and for the pressures that compete with it, and cannot be managed for pH at all. Both object to the same habit in the literature: the reporting of pH change as though it were itself the outcome, when it is at most the beginning of one.

FWhat is emerging is less a verdict than a division of territory. In high-latitude and upwelling waters, where omega already falls below one for weeks at a time, the chemistry does the deciding and no ration will compensate for a mineral that is dissolving; in warmer coastal systems that stay saturated, the outcome turns on whether animals can afford the extra work, which makes food supply and the pressures that reduce it the variables worth managing. The question neither line of work has settled is what happens across many generations, since selection may do quietly what neither a ninety-day trial nor a nineteen-year mooring record was built to detect. Both researchers reject the assumption that governed the first round of studies, namely that a single number describing the water could be read straight off as a forecast for the animals living in it.