Coral bleaching

Why does a rise of barely a degree make a coral expel the very algae that feed it?

A reef-building coral gets a large share of its energy — in shallow water, commonly quoted as up to about ninety per cent — from single-celled algae living inside its own tissue. Losing them is close to starvation. Yet a summer a degree warmer than usual and the coral turns them out. Before assuming the animal is malfunctioning, notice how strange the number is. One degree, on an animal that tolerates a seasonal swing several times that. What could possibly be that sharply tuned?

The first move is to stop treating "one degree" as an absolute. Corals in the Persian Gulf live through summers above 35 degrees Celsius; corals on many Pacific reefs bleach at 30. There is no universal lethal temperature — what matters is how far above this reef's own warmest normal month the water goes. The measure used worldwide is Degree Heating Weeks: the accumulated excess above the local maximum monthly mean, counted only when it exceeds that mean by a degree or more. Around four degree-heating-weeks, significant bleaching is expected; around eight, widespread bleaching with mortality.

Read that definition again, because it changes the question. The threshold is not a temperature, it is a temperature multiplied by a duration. A degree for a week is survivable; a degree for two months is not. So whatever is failing is something that accumulates damage and can be repaired at some finite rate — overwhelmed when the damage rate exceeds the repair rate.

What fails? Photosynthesis. The algal symbionts absorb light and use it to fix carbon. Heat damages the photosynthetic machinery, particularly photosystem II, so absorbed light energy can no longer be routed into carbon fixation. But the light keeps arriving. Energy with nowhere useful to go ends up producing reactive oxygen species — superoxide, hydrogen peroxide — which leak into the host's tissue and damage it. The coral responds by getting rid of the source: expelling symbionts, digesting them, or shedding the host cells containing them. The white colour is bare coral skeleton showing through transparent tissue, because the pigment was never the animal's.

Now test that account, because it makes a prediction the pure-heat story does not. If the trouble is light energy that photosynthesis can no longer absorb, then reducing the light should reduce the bleaching at the same temperature. It does: shaded corals, turbid water, and the undersides of colonies consistently bleach less than sunlit surfaces in the same thermal event. A story about heat alone denaturing proteins cannot explain that. A second, stranger check points the same way — corals also bleach in unusual cold snaps, which again disrupts the photosystem's ability to consume absorbed light.

So why is the response so steep? Because expulsion is a switch, not a dial. Below the threshold the repair keeps up and nothing visible happens; above it the damage runs away, and once the host begins shedding symbionts it loses the very tissue-level machinery that was managing the oxidative load. A gradual stress produces an abrupt outcome — which is what a threshold is.

Here the field genuinely splits. One camp reads expulsion as adaptive: a bailout that stops the oxidative bleeding and lets the host repopulate with heat-tolerant symbiont types, notably Durusdinium, giving a reef a route to acclimate. The other reads it as pathology — damage control at ruinous cost, with the tolerant symbionts supporting markedly slower growth and the acclimation far too slow against the pace of warming. Both readings have evidence and both have interested advocates: reef-restoration ventures and the tourism economies dependent on named reefs have a stake in the adaptive reading, researchers arguing for emissions limits in the pathological one. The honest position is that symbiont shuffling demonstrably shifts thermal tolerance somewhat, and that nobody has shown it shifts it enough.

Where this sits next to its neighbours: the collection's piece on symbiosis argues that mutualisms persist only while each side nets a benefit, and are policed by the host's power to withdraw. Bleaching is that same power exercised at a different fixed point — not against a cheating partner, but against a partner that has become actively toxic through no fault of its own, and whose loss costs the host nearly everything.

Three refinements. First, bleaching is not death: a bleached coral is alive and can be repopulated with symbionts if the heat relents within weeks, which is why recovery between events matters as much as the events. Second, the widely quoted "one degree" is one degree above the local warmest month, sustained — an anomaly, not a temperature, and this is the single most-mangled fact about the phenomenon. Third, the white is not a pigment released but a pigment removed, revealing skeleton.

One popular framing is simply wrong and worth naming: that the coral "kills its algae in a panic" or that warming poisons the coral directly. The heat's first victim is the symbiont's photosystem, and the harm to the animal is chemical fallout from light energy the algae can no longer use. Sun and heat together do the damage; neither alone is sufficient.

• Why recovery interval, rather than peak temperature, may be the binding constraint on reef survival.
• How symbiont community shuffling is measured, and what growth costs the heat-tolerant types impose.

Every term the collection defines is gathered in the glossary.