Oceans absorb another record year of heat
An international team led by the Chinese Academy of Sciences reported in January 2026 that the upper 2,000 metres of the ocean absorbed about 23 additional zetajoules of heat in 2025, the ninth consecutive annual record. The warming rate has doubled since 2005, and the heat is not sitting evenly across basins.
About 90 percent of the excess heat trapped by greenhouse gases does not warm the air. It goes into the ocean, and the ocean has been accumulating it year after year, quietly, without the daily temperature maps that make land heatwaves visible. The peer-reviewed paper accepted in January 2026 by Advances in Atmospheric Sciences, led by Yuying Pan and Lijing Cheng of the Institute of Atmospheric Physics at the Chinese Academy of Sciences, with an international co-author team including John Abraham, Kevin Trenberth and Karina von Schuckmann, reports that 2025 was the hottest year the upper ocean has experienced in the instrumental record.
The headline number is an increase of roughly 23 plus or minus 8 zetajoules in the upper 2,000 metres of the ocean relative to 2024. A zetajoule is ten to the twenty-first joules, and 23 of them is on the order of twenty thousand times the annual electricity generation of the United States. This is the ninth consecutive year in which the upper ocean heat content has set a new record. The 2024 record was already exceptional, and 2025 exceeded it despite a transition toward La Nina conditions during the year, which normally suppress ocean heat uptake rather than accelerate it.
The warming is not uniform, and the regional pattern is where the paper earns its keep. About a third of the global ocean surface ranked among its top three warmest years on record back to 1958, and about 57 percent ranked in the top five. The tropical and South Atlantic, the Mediterranean Sea, the North Indian Ocean and the Southern Ocean are consistently the hot spots. The acceleration is also visible in the trend: the decadal warming rate of upper-ocean heat content was about 0.14 watts per square metre per decade between 1960 and 2025, but it has climbed to roughly 0.32 watts per square metre per decade since 2005. The ocean is absorbing heat faster now than it was twenty years ago.
The sea surface temperature story is slightly different. Global mean sea surface temperature in 2025 was about 0.49 degrees Celsius above the 1981 to 2010 baseline, which makes it the third-warmest year on record at the surface, cooler than 2024 because of the La Nina development. But the subsurface heat content tells the opposite story, because the heat that La Nina drives downward is stored, not erased. This divergence between surface and subsurface is the central methodological point of the paper: sea surface temperature alone is a poor thermometer for the planet, because the ocean moves heat between the surface and the deep interior on annual and decadal timescales.
The implications are mechanical rather than rhetorical. The added heat expands seawater, contributing roughly 2.5 millimetres per year to global sea level rise through thermal expansion alone, on top of meltwater from glaciers and ice sheets. The same heat fuels the intensification of tropical cyclones and the marine heatwaves that are reshaping ecosystems. The paper is careful to note that the 2025 increase is also influenced by recent reductions in reflective sulfate aerosols, which means part of the acceleration is a one-time effect of cleaner air layered on top of the greenhouse gas trend.
The annual ocean heat report has become a ritual, and the ninth record in a row risks numbing the reader. But the acceleration in the decadal rate, and the way the heat keeps piling up even in a La Nina year, are the numbers that matter. The ocean is not a buffer that fills up and stops. It is a bucket that keeps getting poured into.