The WMO has confirmed that 2025 was among the three warmest years, continuing the trend of anomalously high global temperatures. The last 11 years have seen record warmth, and ocean temperatures are not decreasing.
Key Findings
- The last 11 years have been the warmest on record
- Temporary cooling due to La Niña will not change the long-term trend
- Ocean temperatures continue to rise at the same rate
- The WMO combines eight datasets into a single authoritative source of information
- International data exchange is the foundation of climate monitoring
According to a comprehensive analysis of eight datasets conducted by the WMO, the average global temperature at the surface in 2025 was 1.44 °C (with a margin of uncertainty of ± 0.13 °C) above the average for 1850–1900. In two of these datasets, 2025 ranked second among the warmest years, while in the remaining six, it ranked third.
The last three years, from 2023 to 2025, have been among the top three warmest across all eight datasets. The consolidated average temperature for this period was 1.48 °C (with a margin of uncertainty of ± 0.13 °C) above pre-industrial levels. The entire range from 2015 to 2025 has been the warmest on record.
Despite the fact that 2025 began and ended with the La Niña phenomenon, it still turned out to be one of the warmest years due to the accumulating greenhouse gases in the atmosphere. High air and water temperatures have led to extreme weather events—heatwaves, heavy rains, and powerful tropical cyclones. This underscores the need for early warning systems, noted WMO Secretary-General Celeste Saulo.
"Climate monitoring conducted by the WMO based on comprehensive and scientifically grounded global data collection is more important than ever, as we must ensure that information about the state of the Earth remains reliable and accessible to all," added Celeste Saulo.
The WMO's statement was made in the context of the publication of global temperatures provided by various data providers.
Among them are the Climate Change Service under the Copernicus program of the European Centre for Medium-Range Weather Forecasts (ERA5), the Japan Meteorological Agency (JRA-3Q), NASA (GISTEMP v4), the National Oceanic and Atmospheric Administration (NOAAGlobalTemp v6), the UK Met Office in collaboration with the Climate Research Unit of the University of East Anglia (HadCRUT.5.1.0.0), and Berkeley Earth (USA). In 2025, the WMO also incorporated two additional datasets for the first time: a dynamically consistent temperature dataset (DCENT/UK, USA) and a combined ground temperature dataset from China (CMST).
Six of the mentioned datasets are based on measurements taken at weather stations, ships, and buoys, using statistical methods to fill in data gaps. Two datasets—ERA5 and JRA-3Q—are reanalyses that combine past observations, including satellite data, with models to create consistent time series of various climate variables, including temperature. All key datasets use different methodologies, so they may have different temperature metrics and rankings.
According to DCENT and GISTEMP, 2025 ranked second among the warmest years, while among the other six datasets (Berkeley Earth, CMST, ERA5, HadCRUT5, JRA-3Q, and NOAAGlobalTemp), it ranked third.
The average global temperature in 2025 was estimated at 15.08 °C; however, the uncertainty regarding the actual temperature is about 0.5 °C, which is greater than the uncertainty regarding the temperature anomaly for 2025.
The WMO is a UN agency dealing with weather, climate, and water resources, aiming to provide a single authoritative analysis to support decision-making.
Ocean Temperature
In a separate study published in the journal Advances in Atmospheric Sciences, it was also reported that ocean temperatures in 2025 were among the highest on record, indicating heat accumulation in the climate system.
About 90% of the excess heat from global warming is stored in the ocean, making its heat content an important indicator of climate change. A study led by Lizhin Chen from the Institute of Atmospheric Physics of the Chinese Academy of Sciences showed that from 2024 to 2025, global ocean heat content (OHC) in the upper 2000-meter layer increased by approximately 23 ± 8 zettajoules compared to 2024. This is about 200 times the total electricity generation in 2024.
Regionally, about 33% of the area of the World Ocean experienced the warmest conditions on record (1958–2025), and 57% were among the top five warmest, including the tropics, the southern Atlantic Ocean, the Mediterranean Sea, the northern Indian Ocean, and the Southern Ocean, highlighting global ocean warming in various regions.
According to the study, the global average sea surface temperature (SST) in 2025 was 0.49 °C above the baseline of 1981–2010 and (0.12 ± 0.03) °C lower than in 2024, which is consistent with La Niña conditions, but still ranks third among the warmest years on record.
Notes for Editors
The WMO will provide comprehensive information on key climate indicators such as greenhouse gas concentrations, surface temperature, ocean heat content, sea level, glacier retreat, and sea ice extent in its Statement on the State of the Global Climate for 2025, which will be published in March 2026. It will also include information on phenomena with significant impacts.
The datasets used by the WMO provide nearly complete global coverage of near-surface measurements using statistical methods to fill in data gaps in regions with limited access. Reanalyses allow for a comprehensive global analysis by combining past observations, including satellite data, with models to create consistent time series of various climate variables, including temperature.
To calculate the aggregate temperature values relative to pre-industrial levels, the WMO uses anomalies relative to 1981–2010 for each dataset and then adds a shift of 0.69 °C—the difference between the periods 1981–2010 and 1850–1900 according to estimates from the IPCC. The uncertainty of this shift is 0.12 °C. Anomalies from different datasets are averaged to obtain a single annual value. The spread of datasets is combined with the uncertainty in the shift, resulting in a total uncertainty of 0.13 °C. This method has been used in the State of the Global Climate reports in 2023 and 2024 and adapted for the eight datasets used in 2025. Further details can be found at the link.
