Exploring Earth's Oceans Since 1999 · Founded by Blane Perun
3,072 Articles·80+ Countries·25 Years of Ocean Science·Site Map
◆ Trending
25 years exploring Earth's oceans,
coral reefs, and the creatures within.
3,072Articles
80+Countries
1999Founded

Are Atolls Floating? What Holds an Atoll Up

Are Atoll Reefs Floating? — No — They Sit on a Kilometre of Limestone and a Sunken Volcano

Atoll Reef ▸ Q&A
November 7, 2026
4 min read

The short answer

No, atolls do not float. An atoll is the top of a solid column: living coral sits on a cap of reef limestone that can be more than 1 km thick, which in turn rests on an extinct volcano rooted to the ocean floor thousands of metres down. At Enewetak, drills passed through roughly 1,270–1,400 m of limestone before hitting volcanic rock. Atolls only look like they float because the land is so low and the outer slope drops away so steeply into deep blue water.

What holds an atoll up
Floats?
No
Limestone cap
Up to ~1.4 km (Enewetak)
Foundation
Extinct volcano
Surrounding depth
Often 1,000–5,000 m
Land height
Usually < 5 m
Can it sink?
Slowly, over geological time

Why do atolls look like they float?

Low land, a sudden drop-off and that contrast of colours.

Atoll islets rise only a metre or few above high tide, and their outer reef slopes plunge almost vertically into ocean thousands of metres deep. From a plane you see a pale ring sitting on dark indigo water with no visible connection to the sea floor. Diving the outer wall of an atoll makes the truth obvious: the reef just keeps going down past where light runs out.

Solid limestone wall of an atoll dropping into deep water - are atolls floating
Are atolls floating

How deep does an atoll go?

All the way to the sea floor, through the volcano it grew on.

LayerApproximate thicknessMade of
Islets and reef flatA few metresCoral sand, rubble, living reef
Reef limestone capHundreds of m to ~1.4 kmCemented coral and algal skeletons
Volcanic coreSeveral kmBasalt of the drowned island
Oceanic crustBase, often 4–5 km deepSea-floor rock

Drilling on Enewetak in 1951–52 is the classic evidence; see how do coral atolls form.

Can an atoll sink?

Yes, slowly. Atolls continue to subside as their plate cools, and the reef must keep growing to stay at sea level.

Subsidence is measured in fractions of a millimetre a year, slower than a healthy reef grows. But atolls can also “drown” if reef growth stops: the drowned, flat-topped seamounts called guyots are former islands and atolls that fell behind. Today the bigger concern is sea-level rise of several millimetres a year combined with bleaching and acidification. See why do atolls form.

Do atoll islands move?

The islets can. They are piles of sand and rubble that waves and storms rearrange.

Studies of hundreds of Pacific atoll islands over recent decades have found that many islets shifted position or changed shape, and most did not shrink, because storms also deliver new sand from the reef. That dynamism is a sign of a living reef supplying sediment, not floating. Islands only stay healthy as long as the reef that feeds them stays healthy.

Solid
Volcanic foundation
Mobile
Sand islets on top
Living
Reef supplies new sand

Are there any floating islands in the ocean?

Not true islands. Pumice rafts from undersea eruptions and mats of sargassum or vegetation can float, but they drift and break up.

Pumice rafts, such as those from Home Reef in Tonga, can cover hundreds of square kilometres and carry corals and barnacles across oceans, but they are temporary. Every atoll, by contrast, is anchored in the Earth’s crust. See what is an atoll reef.

Why can’t a coral reef float?

The simplest answer is density. Something floats only if it is, on average, lighter than the water it pushes aside. Coral skeletons are made of aragonite, a form of calcium carbonate that is nearly three times as dense as water, and the older reef rock beneath them is mostly calcite and aragonite cemented together. Even with all its pores and cavities, a block of reef limestone sinks straight to the bottom if you drop it overboard. A whole atoll is billions of tonnes of that material stacked on a volcano, so there is nothing in it that could make it buoyant.

The idea of floating reefs probably comes from what an atoll looks like rather than what it is made of. The land is a thin ring of sand and rubble barely above the tide, there is no hill or mountain in sight, and the outer slope drops away so quickly that the water turns from turquoise to deep blue within a few tens of metres of the reef edge. From the air the ring seems to rest on the ocean surface. Underneath, the reef carries on down as a steep wall and then a long slope of limestone and volcanic rock that reaches the deep sea floor.

How did scientists prove what lies under an atoll?

For most of the nineteenth century nobody knew. Charles Darwin argued in 1842 that atolls were coral reefs that had kept growing upward on top of slowly sinking volcanic islands. If he was right, a drill pushed down through an atoll should pass through a great thickness of reef rock before reaching volcanic basement. In 1881 Darwin wrote that he hoped someone would bore into a Pacific or Indian Ocean atoll and bring up cores from 500 or 600 feet down.

The first serious attempt came at Funafuti, in what is now Tuvalu. The Royal Society set up a Coral Reef Committee in 1895, and the Admiralty lent HMS Penguin to carry the expedition. The 1896 party under William Sollas had to give up at 105 feet because loose sand and cavities kept collapsing the hole. Edgeworth David’s 1897 expedition did better, recovering cores from 698 feet despite repeated breakdowns. A third party reopened the borehole in 1898 and drilled until October, when it ran out of replacement diamonds at 1,114 feet, nearly twice the depth Darwin had asked for. The cores were reef rock all the way, with reef-building corals that appeared to be in place, but the drill never reached the bottom of the limestone. The question stayed open for another half century.

The answer finally came from Enewetak in the Marshall Islands, where the United States was testing nuclear weapons. In 1951 and 1952 geologists from the US Geological Survey drilled two deep holes on opposite sides of the atoll. Hole F-1 on the northwest side hit hard basement rock at 4,610 feet, about 1,405 m, though no sample came up. Hole E-1 on the southeast side brought up unweathered basalt cuttings from 4,154 feet, about 1,266 m, and then a solid core of olivine basalt a little deeper. Directly above the basalt was shallow-water limestone of Eocene age. In the words of the 1953 report by Harry Ladd and colleagues, the atoll is a thick cap of limestone resting on the summit of a volcano that rises 2 miles above the ocean floor. That one sentence settles the floating question: an atoll is the top of a mountain.

Is the rock under an atoll solid all the way down?

Not in the sense of one unbroken block, and this is worth knowing because it explains some of the confusion. The Enewetak holes passed through several hundred feet of soft, young Quaternary reef limestone and then a thick Tertiary section of similar rocks with some dolomite. The report describes most of the section as soft or weakly consolidated. Below about 1,400 feet the two holes differed sharply, with different amounts of hard rock and different numbers of cavities. The early Funafuti drillers ran into the same thing near the surface, where sand and holes in the rock made boring very difficult.

So the limestone cap is porous, layered and riddled with old solution cavities formed when sea level fell during ice ages and rainwater ate into exposed reef rock. Seawater moves through it, and in places the rock is more like a sponge than a slab of marble. None of that makes it float. A sponge made of stone is still stone, and it is sitting on a volcano that is itself welded to the oceanic crust. The holes simply mean that an atoll’s foundation is a natural archive of past sea levels and climates, which is why scientists keep drilling into them.

Why do atolls slowly sink?

Atolls sink because the sea floor under them sinks. New oceanic crust forms hot at mid-ocean ridges and then cools and contracts as plate motion carries it away. Cooler rock is denser, so the sea floor gets deeper with age. Barry Parsons and John Sclater’s 1977 analysis of ocean depth and heat flow showed that this pattern holds across the world’s oceans: young sea floor stands high and old sea floor lies deep. A volcano built on the plate rides down with it. On top of that, a large volcano is a heavy load that bends the plate beneath it, adding to the subsidence near the island.

The rates are very slow. A rough check using the Enewetak numbers shows why. Shallow-water limestone of Eocene age, laid down near sea level some tens of millions of years ago, now lies about 1,270 m down. Dividing one by the other gives an average sinking of a few tens of metres per million years, or a few hundredths of a millimetre a year. Healthy reefs can grow upward much faster than that, which is why most atolls have kept their heads at the surface for millions of years.

There is a limit, though. In 1982 Richard Grigg described what he called the Darwin Point, a threshold at about 29 degrees north at the far end of the Hawaiian chain, near Kure Atoll. Islands of the Hawaiian-Emperor chain are carried northwest by the Pacific Plate into cooler water. Grigg estimated that at the Darwin Point corals may supply only about 20% of the calcium carbonate needed to keep up with recent sea-level change, and that Hawaiian atolls and coral islands seem to have drowned near this line over roughly the last 20 million years. Beyond it lie the flat-topped seamounts of the Emperor chain. A drowned atoll does not sink because it was floating and lost buoyancy. It sinks because the reef on top stopped growing while the mountain underneath kept going down.

Can an atoll rise instead of sink?

Yes, and the examples are a useful reminder that atolls move with the Earth’s crust rather than bobbing on the water. Henderson Island in the Pitcairn group is an old atoll that now stands 33.5 m above sea level, rising from a sea floor about 3,500 m deep. Steve Blake’s 1995 study linked its emergence to lithospheric flexure after the nearby Pitcairn volcano was built. The weight of a new volcano bends the plate down around it and pushes up a low bulge further out, rather like a heavy person sitting on a mattress. Henderson sits on that bulge and is estimated to be rising at about 0.09 to 0.10 mm a year, with uplift above sea level beginning roughly 335,000 to 360,000 years ago.

Makatea in the northwest Tuamotus is even more striking, with an upper plateau 113 m above sea level. Its uplift was long blamed on the load of Tahiti and Moorea bending the plate in the same way. A 2023 review by Lucien Montaggioni and colleagues looked again at Makatea and six nearby uplifted atolls, including Mataiva, Tikehau, Rangiroa, Kaukura, Niau and Anaa, and concluded that flexure did not fit their history well. They proposed instead that the atolls were lifted as they drifted across a swell in the mantle linked to episodic heating of the Society hotspot, with Makatea rising at about 0.16 to 0.20 mm a year. Either way, the uplift comes from the solid Earth beneath the reef. Something that floated could not be lifted 113 m by the mantle.

Can an atoll crack or collapse?

An atoll cannot sink suddenly the way a holed boat would, but its steep outer slopes can fail. Underwater landslides happen naturally on the flanks of volcanic islands and atolls, and the clearest recorded example on an atoll was caused by people. France tested nuclear weapons at Moruroa in French Polynesia from 1966, first in the atmosphere and then, after 1974, in shafts drilled deep into the volcanic rock beneath the atoll. On 25 July 1979 a device got stuck halfway down an 800 m shaft and was detonated at that shallower depth. The explosion triggered a large submarine landslide on the southwest rim and a tsunami that struck the atoll and injured workers, and it opened a crack reported as 2 km long and 40 cm wide. A 2017 study by Poupardin and colleagues later modelled that landslide-generated tsunami in detail.

Cracking found on the atoll raised fears that radioactive material trapped in the rock could escape into the ocean, and reports in 2012 suggested that only 11 of 20 monitoring sensors were working. These worries are about a damaged rock structure fracturing and slumping, not about an island coming loose and drifting. They show the opposite of floating: Moruroa behaves like a cracked mountain, not a raft.

What did nuclear craters show about atoll rock?

The nuclear tests in the Marshall Islands left other evidence of how solid atolls are. The Castle Bravo test on 1 March 1954, the largest US explosion at about 15 megatons, was fired on an artificial island on the reef off Namu Island at Bikini Atoll. It blasted a crater about 2,000 m across and 76 m deep into the reef. The crater is still there more than seventy years later, a hole in rock rather than a gap that closed over like a dent in a floating mat.

At Enewetak, the US Geological Survey’s Pacific Enewetak Atoll Crater Exploration program in the 1980s drilled into and around the OAK and KOA craters to understand how the explosions had disturbed the reef limestone. That work, published in 1986, used fossils in the cores to date the layers and compare damaged rock with undisturbed reef, adding detail to the picture first sketched by the 1952 boreholes.

What in the ocean really does float?

Some rock does float, but only for a while and only when it is full of trapped gas. Pumice from undersea eruptions can drift for months. On 7 August 2019 an unnamed submarine volcano in Tonga erupted from about 200 m below sea level and produced a raft of pumice covering about 195 square kilometres. It drifted west and reached the Fiji Islands on 19 September, and Martin Jutzeler and colleagues used satellite images and yacht reports to track it. The 2012 Havre eruption, from a volcano about 900 m down in the southwest Pacific, made a raft estimated at around 1.2 cubic kilometres. A 2025 study of the 2019 pumice found that many clasts contain sealed bubbles that are not connected to the outside, which helps explain how this rock stays afloat. Pumice rafts can carry corals, algae and other animals to new reefs, but they eventually waterlog, break up and sink.

The one part of an atoll that does float, in a sense, is its fresh groundwater. Rain soaks into the islet sand and, because fresh water is slightly less dense than seawater, it collects as a thin lens sitting on top of the salty water that fills the rock below. People on atolls have relied on these lenses for drinking water for centuries. Real floating islands also exist on some lakes, where mats of reeds, peat or plants build up thick enough to walk on. Nothing like that exists on the scale of an atoll, and no coral reef anywhere drifts on the sea.

Frequently asked questions

Are atolls floating islands?

No. They rest on thick limestone over a sunken volcano attached to the sea floor.

What is under an atoll?

Reef limestone, then volcanic basalt, then oceanic crust.

How deep is the rock under an atoll?

At Enewetak, volcanic rock lies roughly 1,270–1,400 m below the surface.

Are atolls sinking?

Slowly over geological time; reef growth normally keeps pace, but sea-level rise is a modern threat.

Why do atolls look like they float?

Their land is very low and their outer slopes drop steeply into deep blue water.

Can an atoll island move?

The sandy islets can shift with storms and currents, but the reef base does not.

Sources & further reading
  1. Ladd, H. S., Ingerson, E., Townsend, R. C., Russell, M. & Stephenson, H. K. (1953). Drilling on Eniwetok Atoll, Marshall Islands. AAPG Bulletin 37: 2257–2280.
  2. Dickinson, W. R. (2009). Pacific atoll living: how long already and until when? GSA Today 19(3): 4–10.
  3. Kench, P. S. et al. (2018). Patterns of island change and persistence offer alternate adaptation pathways for atoll nations. Nature Communications 9: 605.
  4. Royal Society (2024). Boring old coral: the Funafuti boring expeditions, 1896–1898. Royal Society blog.
  5. Parsons, B. & Sclater, J. G. (1977). An analysis of the variation of ocean floor bathymetry and heat flow with age. Journal of Geophysical Research 82: 803–827.
  6. Grigg, R. W. (1982). Darwin Point: a threshold for atoll formation. Coral Reefs 1: 29–34.
  7. Blake, S. G. (1995). Late Quaternary history of Henderson Island, Pitcairn Group. Biological Journal of the Linnean Society 56: 43–62.
  8. Montaggioni, L. F. et al. (2023). Revisiting the geodynamical history of the so-called uplifted atolls, north-west Tuamotu, French Polynesia. Earth-Science Reviews 104532.
  9. Poupardin, A. et al. (2017). The 1979 submarine landslide-generated tsunami in Mururoa, French Polynesia. Pure and Applied Geophysics.
  10. Cronin, T. M. et al. (1986). Pacific Enewetak Atoll Crater Exploration (PEACE) Program, part 2: paleontology and biostratigraphy; application to OAK and KOA craters. USGS Open-File Report 86-159.
  11. Jutzeler, M. et al. (2020). Ongoing dispersal of the 7 August 2019 pumice raft from the Tonga Arc in the southwestern Pacific Ocean. Geophysical Research Letters.
BP
Last updated November 7, 2026
Written by
Blane Perun

Blane Perun has been exploring Earth’s oceans and marine life for over 25 years. Founder of TheSea.Org in 1999, underwater photographer, coral aquaculture pioneer, and explorer of 80+ countries.

About Blane Perun →

More on Atoll Reefs →

Coral Aquaculture · The Pioneer Story
Blane Perun, A coral aquaculture pioneer

Built one of the first home reef tanks designed as a multi-zone replica of natural reef conditions. Author of long-form profiles cited by Wikipedia, Quality Marine, and SeaScope. Independent — no advertisers, no sponsored placements, no editorial board outside the dive log.

Widely credited as the pioneer of named-coral culture in the reef-keeping hobby. The original — “Perun's Purple People Eater”, a Protopalythoa morph from his coral farm — became the first hobbyist coral to carry an aquaculturist's name in trade, sparking the now-standard practice of naming individual coral strains across the industry.

Founded1999 Articles3,072 Countries80+ Cited inWikipedia · Quality Marine · SeaScope
Read Blane's full bio →
The Dispatch · Monthly

Field notes from eighty countries worth of ocean.

One monthly email. New long-form reporting, species profiles, and dive reports — no filler, no tracking pixels, no sponsored placements. Ever.

Sign up and get my one-click underwater color-correction actions for Photoshop — the set I run on my own dive shots. Free.

Get the action now + monthly field notes
No spam · Unsubscribe anytime
Underwater reef photo before and after Blane Perun's Photoshop color-correction action — washed-out blue-green on the left, full natural color restored on the right
Same shot · one click