A quarter of the reef surface is living coral

Across the world’s shallow reefs, 25.8% of the reef surface was covered by living hard coral in 2024, after the largest single-year fall in a global record that starts in 1980¹². Hard coral is the stony, reef-building kind, and that figure is easy to misread as the share of the world’s reefs still alive. It isn’t.

A healthy reef was never all coral. Between 1980 and 2009, living hard coral covered 30.2% of the reef floor on average¹, so 25.8% means the world’s reefs have thinned, not that three quarters of them are gone. The number at the top of this page has kept falling since 2024 because it follows a projection to 2050, explained further down.

Hard coral cover is something a diver can measure with a tape. You run a line across the reef and record what lies under fixed points along it, or photograph a square of reef floor and count what each dot lands on: living hard coral, turf algae, seaweed, sand, rubble, bare rock. The Global Coral Reef Monitoring Network (GCRMN) pooled 21.1 million such points, from 87,299 surveys at 36,886 reef sites, to get the 2024 figure².

What those points sit on is a thin sliver of the sea. The world’s shallow reefs cover roughly 249,677 square kilometres, a little more than the United Kingdom and under 0.2% of the ocean floor, and they support more than 30% of all known marine species². That density is why reefs get called the rainforests of the sea, and like the rainforests they are losing ground: the forests have their own count.

A tenth of the coral lost since 1980, not three quarters

Measured against GCRMN’s own baseline, the world has lost about a tenth of its coral cover, not the three quarters a quick reading of 25.8% suggests. GCRMN compares averages, 30.2% cover across 1980–2009 against 27.3% across 2020–2024, a relative fall of 9.5% or 2.9 percentage points¹. That average was set before most of 2024’s drop, so on the latest single year the loss is somewhat larger.

The same change can be stated two ways because coral cover is itself a percentage. GCRMN gives the problem a box of its own in the report, warning that relative change means “taking a percentage of a percentage” and that the two measures “describe different things and they should not be used interchangeably”². Its example: a reef going from 4% cover to 1% has lost three percentage points, or 75% of its coral, depending which you report².

Reach back further and the loss is much larger. Live coral cover has roughly halved since the 1950s, from an expert-estimated 58% to 70% of reef surface, according to a 2021 study in One Earth⁷. GCRMN agrees its own starting line is not a clean one: the 1980–2009 reference period “likely already includes reefs in a degraded state”². A 2003 study in Science that reconstructed reef condition back thousands of years put it more bluntly: “all reefs were substantially degraded long before outbreaks of coral disease and bleaching”⁹.

So the familiar line that half the world’s coral is gone and a cover figure of 25.8% don’t contradict each other. One is a loss measured from the 1950s, the other a share of the reef floor today, and neither means three quarters of the world’s reefs have vanished.

Heatwaves now come faster than reefs can recover

What is killing the world’s coral reefs is marine heat. When the sea crosses a coral’s thermal limit, the animal expels the microscopic algae living inside its tissues, the zooxanthellae that give it both its colour and most of its food, and the living tissue turns white². A bleached coral is not yet a dead one: if the water cools before the coral starves it can slowly take its algae back, though survivors grow and breed less and catch disease more easily². Extreme heat can also kill a coral outright, with no bleaching first².

Heat stress is counted in Degree Heating Weeks, which add up how far above the local summer maximum the water sits and for how long. Bleaching is expected above 4 °C-weeks and death above 8². Sea surface temperature over the world’s reefs has risen 0.81 °C since 1985², and four global bleaching events have followed, in 1998, 2010, 2016 and 2023 to 2025, costing a relative 6.5%, 9.9%, 6.6% and 8.9% of global coral cover¹.

Each event has been hotter than the last. In 1998, about one reef in forty took heat strong enough to kill coral; in 2024 it was more than two in five².

Counted inside a single 365-day window, the only way to compare the four events fairly, bleaching-level heat reached 77% of reefs in 2024 and 2025, against 55% in the third event and 20% in the first². The 84% you may have seen is NOAA’s count of every reef touched at any point across the whole event, from early 2023 to mid-2025⁸. On either count it was the worst of the four.

The trend comes from the gaps between the shocks. Coral cover can partly recover over five to ten undisturbed years, and before 2010 reefs got roughly a decade between major disturbances. Since 2010 they have been hit every four to six². Reefs now get less time than recovery takes, so the cover ratchets down a step at a time rather than crashing.

Nets, sewage and the shore

Heat sets the trend, and local pressure decides how deep each step cuts. GCRMN names three kinds. Nutrient run-off and sediment from poor land and coastal management degrade water quality and feed the turf and seaweed that compete with coral for space. Overfishing of plant-eating fish and invertebrates strips out the grazers that keep those algae down, while removing large predators thins the reef’s fish community. And destructive fishing breaks up the reef structure itself².

None of that substitutes for cutting emissions, but it measurably changes the odds. A 20-year study of Hawai’i’s reefs through an unprecedented marine heatwave found that reefs with more plant-eating fish and less wastewater and urban run-off had been gaining coral before the heat arrived. Easing both pressures at once gave a reef a three- to sixfold better chance of still being well covered by reef-building coral four years afterwards than easing either one alone¹³.

The Caribbean: from 19.5% coral cover to 11%

Globally the decline is about a tenth. Regionally it runs from a gain to a collapse, and the Caribbean is the worst case: median coral cover there fell from 19.5% across 1980 to 2009 to 11.0% across 2020 to 2024¹².

Heat is the newest cause there, not the first. The region’s branching Acropora corals, the elkhorn and staghorn that built its shallow reef crests, were already declining in the 1960s, which is why GCRMN warns that “1980 should not be interpreted as a pristine reference point”². White-band disease devastated them from the late 1970s. Mass die-offs of the long-spined sea urchin Diadema antillarum in 1983 and again in 2022 removed a main grazer and let algae take over. Stony coral tissue loss disease, a fast-spreading infection first seen in 2014, had reached nearly every Caribbean country by 2023².

Cover figures also miss what that sequence did to the shape of the reefs. A study of nearly 500 surveys across 200 Caribbean reefs between 1969 and 2008 found the region’s reefs had flattened, with “the near disappearance of the most complex reefs over the last 40 years”¹⁰. That complex architecture, more than the cover, is what houses fish and breaks waves.

What coral cover doesn’t tell you

Hard coral cover is one number, and GCRMN is unusually blunt about what it misses. The network calls its own figures “a floor, not a ceiling”, because the data thin out before the mid-1990s and the reference period already contains damaged reefs, so “the actual decline in hard coral cover is likely steeper than the current evidence can measure”². Community composition, recovery rate and the density of young corals are all indicators it lists as things it does not report². Soft corals bleach too, and are not counted at all².

The biggest gap is that cover doesn’t say which corals are growing, and a reef can hold its cover while changing species entirely. South Asia’s coral cover rose a relative 16.2% between its 1997–2009 baseline and 2020–2024, and over the same stretch the 2015–2016 heat stress changed the mix of species: branching and finger-shaped Acropora went down, and boulder-shaped corals, called massive corals, went up².

Regional gains can mislead in a second way, too. The Western Indian Ocean appears to have gained 43%, but GCRMN’s own chapter notes that over 90% of the sites in that analysis had a single year of monitoring, and that 2024 returned the region to roughly its 1999 level anyway².

Which corals matters because the reef’s structure does the work people depend on. Three-dimensional coral architecture dissipates up to 97% of the wave energy hitting a coast², and 104 million people now live within five kilometres of a reef, 46% more than in 2000¹. A 2018 study in Nature calculated vertical growth for more than 200 Atlantic and Indian Ocean reefs and found that, under a moderate emissions scenario, few could keep up with rising seas without sustained ecological recovery first¹¹. That suggests the coastal protection in the 97% figure weakens just as more people come to depend on it.

Coral cover in 2050: a projection, not a deadline

The UN’s climate science panel, the IPCC, gives no year for the last reef. It gives a fall: holding warming to 1.5 °C still leaves “a downward trend in average coral cover (70–90% decline by mid-century)”, measured from the late 2010s, and at 2 °C more than 99% of reef-building corals are lost⁵⁶. The number at the top of this page takes the middle of that range, an 80% further decline, reads “mid-century” as 2050, and runs in a straight line from 25.8% in 2024 to 5.2% in 2050.

Two older forecasts still circulate, and neither is a forecast in its own right. “All the coral reefs in the world could be gone by 2070” comes from a 2019 comment piece in Nature that cites earlier work rather than presenting a projection of its own¹⁸. And the claim that all corals will be threatened by mid-century, with 75% facing high to critical threat, is from the World Resources Institute’s Reefs at Risk Revisited, a modelled threat classification published in 2011 and not updated since¹⁹.

The IPCC line is a projection, and a steep one. The measured trend since 1980 is far gentler, about a tenth of a percentage point a year, which would leave cover near 23% in 2050³. The IPCC’s pace is roughly six and a half times that. Two things argue for the steeper line anyway: a single year, 2024, cost 8.9% of the world’s coral², and the IPCC calls the assumptions behind its own coral assessment “highly conservative”⁵.

Other work points the same way without naming a date for zero. A 2022 analysis in Earth’s Future projected that at 1.5 °C of warming only 2 to 5% of reef locations stay below a severe-bleaching threshold, and that at 2 °C none do¹². Where coral cover ends up depends mostly on how much the planet warms.

Reefs still come back

Not globally, and not right now: the fourth bleaching event erased the last rebound, and 2024 brought the largest single-year fall on record². But reefs do come back when the heat gives them time. Global coral cover rose about 6% in relative terms between 2017 and 2019, in the gap after the third bleaching event¹. In GCRMN’s Australia region, cover went from 19.8% in 2018 back to 29.0% in 2023, before a 2024–2025 event expected to be the region’s most severe mass bleaching on record². South Asia, with its change of species, is the one region with a solid long-run increase. GCRMN’s verdict on the whole picture is that “coral reefs have not crossed a point of no return, and global coral reef collapse is not inevitable”².

What is being tried on the reefs themselves is still small. A 2026 study in Science Advances bred corals for heat tolerance: selectively bred young corals survived and grew better at 32 °C from one reef but not from another, and algae partners evolved to handle heat improved survival at 32 °C but slowed growth at 28 °C¹⁴. A global review of restoration projects found a median restored area of 100 square metres and 60% of projects monitored for under 18 months, and warned against treating restoration “as a replacement for meaningful action on climate change”¹⁵.

The pledges are far larger than the evidence behind them. The Coral Reef Breakthrough, launched in 2023 by the International Coral Reef Initiative and the Global Fund for Coral Reefs, aims to secure 125,000 square kilometres of reef and double the reef area under effective protection by 2030, with at least US$12 billion invested¹⁷². We found no published measurement of progress against it, so when the target is quoted, it is still only a target.

Modelling of 3,800 reefs on Australia’s Great Barrier Reef projects rapid decline by mid-century under every emissions scenario, then recovery this century if warming stays below 2 °C, because the corals’ heat tolerance can adapt fast enough at that level¹⁶. If the Great Barrier Reef is a guide, the reefs’ future is decided by how far warming goes, and they have already shown they recover when they get a long enough gap between heatwaves.