The Summer Japan’s
Coral Escape Route
Overheated

At Tateyama, near the mouth of Tokyo Bay, the mathematical warning was unmistakable. Accumulated heat stress reached 12.9°C-weeks in 2024—well above the level commonly associated with widespread bleaching and mortality. Yet when researchers examined their permanent plots, every coral was healthy.

More than 1,500 kilometers southwest, the outcome was reversed. South of Sesoko Island in Okinawa, 38% of coral cover was completely bleached in September. When the team returned in February 2025, living coral cover had fallen from 22% to 6%. The white colonies had not merely lost color. Much of the community had died.

Between those two ends of Japan, corals in Tatsukushi, Kushimoto, Tsushima, Tagojima and Numazu produced five more versions of the same summer. Some died quickly. Some paled and recovered. At Tsushima, researchers documented heat-driven bleaching for the first time. At two sites in Shizuoka, it happened for the second year in a row.

Together, the observations form the first coast-spanning account of how the extreme heat of 2024 affected both subtropical reefs and non-reef temperate coral communities along the Kuroshio and Tsushima currents. The study, led by University of the Ryukyus professor Haruko Kurihara, was published in Scientific Reports on August 25, 2026.

A white coral is not yet a dead coral

Corals are animals living in partnership with photosynthetic algae in their tissues. Those algae—members of the family Symbiodiniaceae, often called zooxanthellae—supply much of the energy that allows corals to grow and build calcium-carbonate skeletons. Under sustained heat and light stress, that relationship breaks down. The algae or their pigments are lost, exposing the white skeleton through nearly transparent coral tissue.

Bleaching is therefore a stress response, not a death certificate. If conditions improve soon enough, algae can repopulate the tissue and the coral may recover. If heat persists, the energy deficit raises the risk of starvation, disease and death. Timing matters: a pale colony, a fully white colony and a recently dead skeleton are points on a possible trajectory, not interchangeable categories.

The 2024 study preserved those distinctions. In temperate permanent plots, researchers classified coral area as healthy, pale, fully bleached or recently dead. At Okinawa, they surveyed again after winter. That follow-up converted a September picture of bleaching into evidence of mortality. At Kushimoto, by contrast, no coral was dead during the survey and the authors report that many bleached colonies subsequently recovered.

The heat dose behind the warning

The study compared sites with Degree Heating Weeks, or DHW, a measure of accumulated thermal stress developed for coral-bleaching surveillance. Kurihara’s team used NOAA’s quarter-degree Optimum Interpolation Sea Surface Temperature record. For each location, it calculated a local maximum-monthly-mean baseline from 1982 through 2024, then accumulated daily heat above that baseline plus 1°C over the previous 12 weeks.

A simplified example makes the unit intuitive: 1°C above the threshold for one week contributes about 1°C-week; 2°C above it for four weeks contributes about 8°C-weeks. Bleaching risk commonly rises at 4°C-weeks. At 8, reef-wide bleaching and mortality of heat-sensitive corals become likely. NOAA’s current operational scale adds still higher mortality warnings at 12, 16 and 20.

All eight Japanese sites exceeded 8 in 2024. Tsushima reached 19.10, Tagojima 18.73 and Numazu 17.24—far above Ishigaki’s 9.24 and Okinawa’s 8.67. Every location recorded its highest DHW in the 1982–2024 series. Seven of eight also recorded their highest maximum monthly mean temperature; Ishigaki’s maximum had occurred in 2022.

The northern values are especially striking. In 2024, maximum monthly mean temperatures at Tsushima, Tagojima and Numazu were about 3°C above their respective 1982–2024 averages. Across the full record, the rate of summer warming increased significantly with latitude. Numazu warmed by 0.34°C per decade and Tateyama by 0.39, compared with 0.18 at Ishigaki and 0.19 at Okinawa.

Eight places, seven bleaching stories

These numbers are descriptive, not a league table. The Ishigaki observation was a one-hour swim. Okinawa was mapped by manta tow. At Tatsukushi, Kushimoto, Tsushima, Tagojima and Tateyama, the team used four permanent 3-by-3-meter photographic quadrats established in 2015. Numazu used a 10-by-5-meter permanent plot for bleaching status and a different belt-transect design for baseline cover.

The authors explicitly state that the nonuniform protocols prevent a formal statistical comparison of bleaching prevalence among sites. They did test whether heat stress tracked the combined proportion of pale, bleached and recently dead coral. It did not show a clear association across the eight locations (p=0.42). DHW correctly signaled widespread danger; it did not explain the pattern by itself.

An escape route drawn by warm currents

Japan occupies a rare ecological hinge. The Kuroshio carries warm water and coral larvae north along the Pacific side of the archipelago. The Tsushima Current sends a branch into the Sea of Japan. South of the paper’s “Kuroshio Barrier” near Tanegashima and Yakushima are subtropical reef systems. Farther north, reef-building corals settle on rocky coasts, often without constructing the large geological reef frameworks familiar in Okinawa.

That distinction matters. Calling every northern site a “reef” overstates what is present. Tatsukushi, Kushimoto, Tsushima, Tagojima, Numazu and Tateyama are treated in the paper as non-reef temperate coral communities. They can still be ecologically rich, locally dense and important to fisheries and tourism, but their structure and species mix differ from southern reefs.

In 2011, Hiroya Yamano and colleagues assembled eight decades of records and found that several tropical coral species had expanded their northern range in Japan, in one case at up to 14 kilometers per year. It was the first nationwide detection of rapid poleward coral expansion linked to rising sea-surface temperatures. The change made Japan a leading example of how mobile species might track climate.

It also supported a hopeful possibility: temperate coasts might become climate refugia, places where corals displaced from overheated tropical waters could establish populations and persist. A refugium is not simply somewhere cooler on a map. To matter over decades, it must remain tolerable through extremes, allow reproduction and recruitment, preserve enough diversity and, ideally, seed damaged areas elsewhere.

The Kuroshio complicates that hope. The current that transports larvae also transports heat. A coral can extend its range only so quickly, and a newly established marginal population may carry less genetic diversity than one near the center of the range. If the destination warms as fast as—or faster than—the species moves and adapts, the route narrows.

Half a century of protection, four decades of warning

Tatsukushi and Kushimoto were among Japan’s first marine-park areas, designated in 1970. Their temperate coral communities became enduring examples of the biological reach of the Kuroshio. Japan’s Ministry of the Environment later built one of the country’s most important long-term records through Monitoring Sites 1000, which has tracked coral cover, bleaching, crown-of-thorns starfish and other disturbances since the early 2000s.

The monitoring record shows that bleaching is not new. At Sesoko, more than 40% of corals reportedly bleached and 10% died in 1980. The global event of 1998 struck Okinawa severely; coral cover at Sesoko fell by 85%. High-temperature bleaching also reached temperate Kochi and Kushimoto that year, though early records there were limited.

Further bleaching followed around Okinawa in 2001, 2003 and 2007. Shallow coral cover around the island fell from 24.4% in 1995 to 7.5% in 2009. Yet the system retained its capacity to rebound. Around Sesoko, cover increased from 13.8% in 2017 to 28.7% in 2023, driven largely by fast-growing Acropora and Montipora.

That recovery contains a trap. Fast-growing branching corals can rebuild cover quickly, but many are especially vulnerable to heat. They can be both the engine of recovery and the largest component of the next collapse. In 2022, about 63% of corals bleached and 15% died in Sekisei Lagoon near Ishigaki. In 2023, divers recorded bleaching at Tagojima and Numazu, where it had not previously been documented. The 2024 event then raised DHW above 17 at both sites.

Why Tateyama refused the pattern

Tateyama is not proof that DHW is meaningless. It is evidence that a relative heat anomaly is not the same thing as an absolute biological threshold. Its maximum monthly mean reached a record level in 2024, but at 27.86°C it remained lower than every other study site. The local corals may simply not have crossed the temperature at which their particular host-symbiont partnerships break down.

The paper cites unpublished experiments in which two Acropora taxa from Tateyama did not bleach at 28°C. Because those data are unpublished, they should be treated as a clue, not independent confirmation. The dominant species at Tateyama may also be relatively heat tolerant. Light, water flow and access to zooplankton—an alternative energy source during bleaching—could have mattered.

Kushimoto offers a different clue. Its community, like Tatsukushi’s, was dominated by the A. hyacinthus complex, yet mortality was absent during the survey and many corals apparently recovered. At Tatsukushi, DHW remained above 9 for 47 days beginning September 10. Kushimoto crossed 9 about two weeks later and remained there for 36 days. A typhoon passed along Japan from September 18 to 21, lowering sea-surface temperatures and possibly interrupting the stress at Kushimoto.

Ishigaki adds another version. Bleaching was widespread in 2024, but post-event Ministry of the Environment surveys found little decline in cover. The authors suggest two nonexclusive explanations: typhoons cooled and shaded the reef in July and August, and the 2022 event had already removed many heat-sensitive colonies, leaving a different community in 2024. The latter is ecological filtering, not cost-free adaptation; survival can come through loss.

Northward pioneers were not heatproof

The most direct challenge to the refuge hypothesis came from the corals that had expanded poleward. Tropical-origin members of the Acropora hyacinthus complex, A. solitaryensis and A. muricata were highly susceptible. So were temperate A. cf. glauca and the temperate-endemic A. pruinosa. For several taxa, nearly all observed cover was fully bleached.

At Tatsukushi, the A. hyacinthus complex now accounts for roughly 80% of coral cover, but surveys from the 1930s did not record it. That history makes it a compelling example of recent establishment at higher latitude—and of the limits of establishment as protection. Reaching a cooler coast did not make it resistant to an extreme heatwave.

Other species remained healthy. Lithophyllon undulatum at Tsushima and Leptastrea aff. pruinosa at Tagojima appeared tolerant in this event. But tolerance was not a fixed property across geography. Some taxa that resisted a milder 2016 bleaching event in eastern Australia bleached severely in Japan in 2024. The heat dose, local population, hidden species boundaries and symbiotic algae can all change the response.

The authors discuss reduced host and symbiont genetic diversity at poleward range edges as one possible constraint. They did not measure genomes or symbiont communities at the eight sites in this study. It remains a biologically plausible explanation to test, not an established cause of the observed differences.

What the study can—and cannot—settle

The paper’s strongest result is geographic. It connects a record marine heat event to bleaching from tropical reef systems to some of Japan’s northernmost coral communities. It documents a first known thermal-bleaching event at Tsushima and consecutive events in Shizuoka. It shows that endemic temperate corals and poleward-expanding tropical corals can both fail under extreme heat.

Its limits are equally consequential. Every biological observation came from water shallower than five meters. The study does not evaluate mesophotic habitats, deep channels, groundwater-influenced coves or small upwelling areas that might remain cooler. Surveys occurred at different times, so one site may have been photographed near the peak of bleaching and another after partial recovery.

The temperature data also operate at a different scale from the coral plots. NOAA OISST blends satellite, ship, buoy and Argo observations into quarter-degree grid cells, providing a consistent 42-year history but not the temperature directly touching each colony. A sheltered cove and an exposed headland can occupy the same cell while experiencing different currents, light and daily temperature swings.

Finally, the study is not a formal event-attribution analysis. It places the 2024 Japanese heat within the well-established context of human-driven ocean warming and the 2023–24 El Niño, but it does not calculate how much anthropogenic climate change increased the probability or intensity of heat at each site. “Climate change caused every local outcome” would go beyond the analysis.

From a refuge line to a refuge mosaic

Japan’s event did not occur in isolation. NOAA and the International Coral Reef Initiative confirmed the fourth global coral bleaching event in April 2024. NOAA’s June 2026 assessment concluded that the episode likely ended in 2025 after bleaching-level heat had affected about 84% of the world’s reef area and mass bleaching had been documented in at least 83 countries and territories. It was the largest global event on record.

Emissions reduction is the only intervention that can address the underlying rise in ocean heat. Local management cannot cool an ocean basin. It can, however, affect whether a heat-stressed community has enough resilience left to recover. Reducing sediment and nutrient runoff, physical damage, destructive use and outbreaks of coral predators can remove additional burdens from survivors.

Japan’s Coral Reef Ecosystem Conservation Action Plan 2022–2030 already identifies continuous monitoring, stronger management and community-linked action as priorities. The 2024 evidence makes the geographic design of that work more urgent. Monitoring should use comparable methods, extend across depths and measure in-water temperature, light, flow, food availability, host genetics and symbionts—not coral color alone.

Most of all, conservation needs the exceptions. Tateyama’s healthy plots, Kushimoto’s recovery and the resistant species inside otherwise bleached communities may reveal mechanisms that averages erase. Some will be temporary accidents. Others may identify small, durable refuges worth protecting as sources of larvae and diversity.

The simple refuge map placed safety to the north. The 2024 heatwave replaced it with a mosaic: one bay cools during a typhoon, another stays hot for 47 days; one species bleaches everywhere, another survives in a neighboring quadrat; a shallow reef dies while a deeper slope remains unknown. Japan’s temperate waters are no longer credible as an automatic sanctuary. They may still contain sanctuaries—but science will have to find them at the scale on which corals actually live.