Learning Objectives
By the end of this knowledge development session, student divers will be able to:
- Explain why a coral colony's thermal environment is not the same as the sea surface temperature reported for the region.
- Describe two limitations of satellite sea surface temperature measurement in coastal water.
- State why recreational divers are positioned to observe a part of the ocean that the global observing system samples poorly.
- Explain why a single diver's measurement can be imprecise while a population of such measurements is useful.
- Describe one way in which diver-collected data is systematically biased, and explain why collecting more of it does not remove that bias.
Presentation Notes
a. The coral does not experience an average. A reef temperature reported in the news, or on a monitoring product, is a value averaged over an area — often several kilometres square — and over a day or a week. A coral colony experiences none of that. It experiences the water in contact with its tissue, at that moment.
A talking point. Think of a shower. If it runs scalding for a minute and then freezing for the next, the average over those two minutes is perfectly comfortable — and you have been burned and then chilled. The average describes nothing that happened to you. The same holds across space: a shower that is scalding at your head and freezing at your feet averages out to nothing you would call pleasant. A coral colony is in both positions at once, and unlike you it cannot step out.
b. The layer that matters is thinner than a millimetre. Recent work on coral physiology has made this sharper than it used to be. Corals are not passive in still water: they beat microscopic hairs called cilia to stir the water at their own surface, generating small vortices that sweep oxygenated water across the tissue and carry waste away. They do this because diffusion alone is far too slow — oxygen takes minutes to cross a single millimetre by diffusion, and the coral cannot wait. The consequence for us is that a coral's environment is not the reef, and not even the water a metre away. It is the thin film of water the colony is actively working on, and that film is renewed from the water immediately around it.
Where this comes from. The idea that coral cilia actively pump rather than merely stir is not new — Shapiro and colleagues showed in 2014 that ciliary flows enhance mass transport in reef corals, and Pacherres and colleagues mapped the resulting vortices and their oxygen dynamics in 2020 and 2022. What is recent is the three-dimensional picture. Selvan, Pacherres and colleagues (PRX Life, May 2026) combined measurements of how cilia are distributed and oriented across several reef-building species with a mathematical model of the flow, and found that counter-rotating vortices emerge robustly in three dimensions from the way the cilia are arranged — and that the variation in their orientation, which might look like untidiness, increases transport of slow-diffusing material by more than half. The full citation is in the Technical Supplement. A reader who wants the mechanism in detail should go to the paper; it is open access.
c. Warming turns the coral's own defence against it. This is the part worth understanding properly, because the mechanism is not the obvious one.
Warmer water holds less dissolved oxygen, and a warmer coral needs more of it. The colony responds by beating its cilia faster — in the laboratory, from around twenty-one beats per second at 27 °C to more than thirty at 37 °C. It is, in effect, breathing harder.
That compensation works, up to a point. Past it, the faster beating becomes self-defeating: the intensified motion traps oxygen-depleted water in a thin layer against the tissue, so the colony starves even while the surrounding sea remains well oxygenated. Push further and the ciliary coordination collapses altogether, the vortices dissipate, and oxygen transport falls back to plain diffusion — which, as note (b) said, is far too slow.
The relevant lesson for a diver is not the specific temperatures. Those are laboratory values, far above what a reef normally sees, and the researchers themselves caution against treating them as a standard threshold: each species is adapted to its own range of daily variation. What matters is the shape of the thing. Heat and oxygen supply are coupled, the coupling has a tipping point, and it happens in a layer of water the diver is swimming through.
Where this comes from. Pacherres and colleagues (Science Advances, May 2026) combined laboratory heating experiments with mathematical modelling to measure ciliary beat frequency and the oxygen field at the coral surface as temperature rose. The full citation is in the Technical Supplement.
d. Two colonies side by side can have different fates. Divers see this on a bleaching reef: patches bleach and neighbouring patches do not. Some of that difference is species and history. Some of it is local — and "local" here means the water immediately around the colony, at a scale of metres or less.
Be careful about assuming which way the flow relationship runs. It is tempting to conclude that sheltered colonies must fare worse because less water reaches them, but the published work does not straightforwardly say that. The modelling in note (b) found that externally imposed flow reduces the transport the cilia themselves achieve, and that ciliary transport matters most in low-flow conditions. The heating work in note (c) found the failure arising from the colony's own response rather than from a lack of ambient water. The honest position is that ambient flow and ciliary pumping interact, that the interaction is not yet settled, and that a diver measuring temperature and flow at reef scale is measuring the inputs to a question the field is still working on.
What is not in doubt is the scale. A regional average cannot explain a difference measured in metres, because it does not resolve metres — and nothing in the observing system measures at metres.
e. The instruments that watch the ocean do not watch here well. Satellites measure the temperature of a thin surface skin, from orbit, averaged over a pixel. In the open ocean this works well. Near a coastline it works less well: the pixel may contain land as well as water, the water is shallow and stratified differently from the open sea, and the reported value carries a systematic offset from what is actually in the water below. Autonomous profiling floats — the backbone of the global temperature record — are designed to drift in deep water and cannot operate over a shallow reef. The upper tens of metres of the world's coastal reefs is one of the least-sampled parts of the ocean, and it is the part where coral lives.
f. And that gap is not a local problem, because there is only one ocean. We speak of five, but the boundaries are administrative. The water is continuous and it moves — surface currents on a scale of weeks and months, the deep circulation on a scale of centuries — so water leaving one basin arrives in another, carrying its heat, its salt, its dissolved gas and everything suspended in it.
That is not a poetic observation, and it is not the butterfly effect: the mechanism is transport and mixing, which spreads and dilutes rather than amplifies. But it does mean the coastal zone is not a backwater. It is where the ocean meets the land, where heat is gained and lost fastest, where runoff and sediment enter, and where much of the water that eventually circulates everywhere is conditioned. The least-observed part of the ocean is one of the parts that does the most to it.
It also means this course is relevant anywhere there is ocean. The physics is the same in the Caribbean, the Red Sea, the Coral Triangle and a temperate kelp coast. The protocol does not know where it is.
g. Divers are already there. Every diver carries a pressure sensor, a temperature sensor, and a clock, and every dive produces a record of depth and temperature against time. A study that put twenty-eight dive computers through hyperbaric chamber and open sea trials against calibrated reference instruments found an overall bias of about −0.2 °C, with a spread of about ±1.1 °C. Hold on to both of those numbers, because they say two different things.
h. One reading is poor; many readings are not. The spread of 1.1 °C is several times worse than an ocean observing system asks for. So a single dive computer is not a scientific instrument. But most of that spread is scatter — instrument to instrument, dive to dive — and scatter shrinks when readings are averaged, in proportion to the square root of how many there are. A hundred divers measuring the same reef produce a mean far better than any one of them can. This is the reason a diver's crude measurement is worth making at all.
What does not shrink is the −0.2 °C. Nearly every model in that study read low against the reference, which means part of the error is common to the instruments rather than random between them — and no amount of averaging removes it. The saving grace is its size: 0.2 °C sits at the edge of what a global observing system requires for in-situ temperature, and the study's authors concluded that with enough data points the result is comparable to coastal temperature datasets already in use.
i. Averaging does not fix everything, and you must know which is which. There are two kinds of error. Random error scatters either side of the truth, and more measurements shrink it. Systematic error — bias — pushes every measurement the same way, and more measurements do nothing to it. They simply produce a tighter, more confident answer that is still wrong.
j. Diver data is biased, and here is how. Divers dive in daylight, so a diver record of reef temperature is a record of daytime temperature, and daytime is the warm half of the daily cycle. Divers dive in good weather and in season. Divers dive where there are boats, moorings and dive shops, which is not a random selection of the world's reefs. None of this is fixed by more divers. It is fixed by saying so — by stating what your sample represents and what it does not.
k. What this course adds is not the sensor; it is the protocol. The temperature record in an ordinary dive log is a by-product of a dive planned for something else. It has no stated depth accuracy, no calibration, no replicate, and no record of whether the instrument had settled. The same sensor, used to a protocol — with the instrument's error measured before the dive, with depth held and recorded, with a repeat observation to establish how much the reading wobbles when nothing has changed — produces something a scientist can use. That is the difference between a dive log and a measurement, and it is what the remainder of this course teaches.