Sea surface temperature (2024)

The sea surface is the boundary between the ocean and atmosphere. The sea surface temperature (SST) can be used to understand the flows of energy between the two and hence the role of the oceans in shaping the weather and climate, and vice versa. Records of SST, using observations made from ships, go back to the mid-19th century and are some of the longest instrumental records available for understanding the climate. Since the early 1980s, there have also been measurements from satellites and buoys.

SST patterns influence key elements of the climate system, such as atmospheric circulation, rainfall patterns and tropical cyclones. The top few metres of the ocean can hold as much energy as the entire atmosphere, highlighting the importance of SSTs for climate monitoring. They also form a key part of the datasets used to estimate global temperature change. SST is also important for forecasting at a range of timescales, from hours to years ahead. Marine heatwaves, ocean fronts and regions of upwelling can all be monitored using measurements of SST.

Globe

Figure 1. Annual mean sea surface temperature anomalies (°C) for 60°S–60°N, relative to the average for the 1991–2020 reference period. Data source: ERA5, ESA SST CCI Analysis v3, ERSSTv5 , HadSST.4.0.1.0, HadSST4.0.1.0 . Credit: C3S/ECMWF.

The average sea surface temperatures (SSTs) over the extrapolar ocean^[2] (60°S–60°N) has increased markedly since records began in 1850. During the second half of the 19th century, the SST was relatively stable. During the first two decades of the 20th century, there was a period of cooling, followed by a period of warming during the 1930s and 1940s. There was then little overall change from the 1950s to the mid-1970s. In the late 1970s, the SST abruptly started to rise and this warming continues.

Between the late 19th century (1880–1900) and the last five years (2019–2023), the average SST over the extrapolar ocean has increased by about 0.9°C. The increase from 1980 to 2023 has been around 0.6°C. The latest five-year average is around 0.3°C above the average for the 1991–2020 reference period. See note [3]. The chief driver of year-to-year variability is the El Niño Southern Oscillation (ENSO) – periods of warmer (El Niño) or cooler (La Niña) than average SSTs in the central and eastern tropical Pacific. El Niños temporarily increase the global average SST and La Niñas temporarily decrease it. ENSO is discussed in more detail below.

Differences between datasets are larger in the earlier parts of the records. This reflects both the sparse data coverage and uncertainties related to measurement techniques^[4]. In the 19th century, for example, SST measurements were made by drawing a water sample up to the ship’s deck in a wooden or canvas bucket. Canvas buckets became more common over time but were found to cool unacceptably under certain conditions; they were later replaced by specially-designed insulated buckets. From the 1930s onward, measurements started to be made in ship engine rooms, using the water taken in at depth for cooling and other purposes. More recently, dedicated sensors have been fitted to the hulls of ships. Since the late 1970s, measurements have also been made by moored and drifting buoys. The approach used to account for the effect of these instrumental changes may differ between SST datasets.

All datasets show that SSTs have increased in most parts of the ocean in the last three decades. However, this increase has not been uniform across all parts of the ocean. Calculating the linear trend for the last three decades gives an indication of the recent rate of warming or cooling across the global ocean; however, SST change over time is rarely linear.

Amongst the fastest warming areas are parts of the Arctic Ocean, including the Barents and Kara Seas, the Baltic Sea, the Black Sea and parts of the extra-tropical Pacific, such as the North Pacific. However, even over periods of a few decades, local trends can be affected by, and also mask, internal variability. An area of cooling in the North Atlantic, to the south of Greenland and Iceland, is consistent with centennial trends estimated from in situ data sources, and one of the few areas globally to have cooled on those timescales^[6]. The marked cooling trends in the eastern tropical Pacific are related to prolonged La Niña conditions from late 2020 to early 2023.

Europe

Figure 3. Annual average SST anomalies (°C), relative to the average for the 1991–2020 reference period, for the European ocean domain (WMO Region VI). Data source: ERA5, ESA SST CCI Analysis v3, ERSSTv5, COBE2-SST, HadSST.4.0.1.0. Credit: C3S/ECMWF/UK Met Office.

The SST averaged over sea waters in the WMO Region VI – Europe domain has increased since records began in 1850. The rate of warming has not been uniform. There were periods of faster warming – between the 1920s and 1940s, and between the 1990s and the early 2000s – and periods of slower warming or cooling – between the 1940s and 1980s, and from the early 2000s to the late 2010s. The record-high SST in 2023 represents a marked jump above the levels of the past two decades. A similar pattern of changes has been observed for surface air temperatures over land. On decadal timescales, this broadly follows the same pattern of relatively large variability in temperatures seen in the North Atlantic Ocean; often referred to as Atlantic Multidecadal Oscillation. This variability is thought to be a combination of internal variability and possibly a response to changes in aerosol loading over the ocean^[7].

The European average SST for 2019–2023 is about 1.1°C above the 1880–1900 average and about 0.4°C above the average for the 1991–2020 reference period. The total increase since 1980 has been around 1.1°C. See note [8].

ENSO

Figure 4. Monthly sea surface temperature anomalies (°C) for the Niño 3.4 region (5°S–5°N 170–120°W), relative to the average for the 1991–2020 reference period. Horizontal lines are marked at ±0.5°C to indicate typical thresholds used to identify El Niño and La Niña events. Data sources: ERA5, ESA SST CCI Analysis v3, ERSSTv5, COBE2-SST, HadSST.4.0.1.0. Credit: C3S/ECMWF.

In the tropical Pacific, prevailing winds and currents cause the upwelling of cold water along the western edge of the Americas and along the eastern tropical Pacific. In some years, the winds strengthen, increasing the upwelling of cold water, which in turn leads to a further strengthening of the winds. The presence of cooler-than-average SSTs in the eastern tropical Pacific is characteristic of La Niña events. In other years, a weakening of the winds can lead to a reduction in upwelling and a warming of surface waters. These above-average SSTs in the eastern equatorial Pacific characterise El Niño events. El Niño and La Niña events affect weather patterns around the world. The interlinked variability of ocean and atmosphere in the tropical Pacific is often referred to as the El Niño Southern Oscillation (ENSO). The SST conditions associated with ENSO are routinely monitored in four areas in the central and eastern tropical Pacific, referred to as the Niño 1+2, Niño 3, Niño 3.4 and Niño 4 regions^[9]. The Niño 3.4 index is the one most commonly used for climate studies as it captures an area of important variability for El Niño and La Niña.

Consecutive years with La Niña events are more common than consecutive years with El Niño events, and not every year has a La Niña or El Niño event. Historically, there have been periods with more frequent and stronger El Niño events and others when La Niña events predominate. Following a strong El Niño event of 2015–2016, the period from late 2016 to early 2023 was dominated by La Niña conditions. In 2023, there was a transition to El Niño conditions. The 2023–2024 El Niño event was not as strong as the events of 1982–1983, 1997–1998, and 2015–2016.

Further reading

Notes

[1] The grey shading indicates the 95% uncertainty range. The range represents uncertainties due to measurement error and residual systematic errors in the data as assessed for the HadCRUT.4.0.1.0 dataset but does not include uncertainty due to incomplete spatial coverage. Comparable uncertainty estimates are available precomputed for ERSSTv5 for the global average (not shown) but not for the regional averages. Uncertainty in ERSST is represented by an ensemble, but this is not currently updated.

[2] The extrapolar ocean is defined here as the ocean domain between 60°S and 60°N and corresponds approximately to the region of ocean free of seasonal sea ice. It is where direct SST observations from satellites (since the early 1980s) and in situ measurements are mostly available.

[3] The table below provides estimates of sea surface temperature changes for 60°S–60°N from the individual datasets used in this indicator. The column ‘total linear change’ corresponds to the change estimated by calculating the trend (based on least-square linear regression) for the 1980–2023 period and multiplying the result by the number of years (43 years).

	2019–2023 minus 1880–1900 (°C)	2019–2023 minus 1991–2020 (°C)	Total linear change from 1980 to 2023 (°C)
ERA5	N/A	0.26	0.55
ESA SST CCI v3	N/A	0.25	0.55
ERSSTv5	0.82	0.26	0.54
COB2-SST	0.87	0.25	0.62
HadSST4	0.97	0.28	0.62
Mean	0.89	0.26	0.58

The estimated change from the IPCC AR6 (Chapter 9, Fox-Kemper et al., 2021) report is 0.88°C (0.68–1.01°C) from 1850–1900 to 2011–2020, and 0.60°C (0.44–0.74°C) from 1980 to 2020. More on change in sea surface temperature can also be found in IPCC AR5 (Chapter 2, Hartmann et al., 2013).

[4] Poor geographical cover of early records plays a role in the uncertainty, however, when collocating data so that the spatial coverage is the same, there are still substantial differences. Reducing modern SST fields down to 19th century coverage makes little difference. See, for example, Figure 7 in Kennedy (2013).

[5] The linear trends are calculated using ordinary least-square regression. Note that this does not imply that the temperature changes are themselves linear.

[6] More information about the cooling region in the North Atlantic, often called the ‘North Atlantic warming hole’, can be found in Keil et al. (2020).

[7] For more on the potential drivers, see, for example, Booth et al. (2021).

[8] The table below provides estimates of sea surface temperature changes for WMO Region VI – Europe domain from the individual datasets used in this indicator. The column ‘total linear change’ corresponds to the change estimated by calculating the trend (based on least-square linear regression) for the 1980–2023 period and multiplying the result by the number of years (43 years).

	2019–2023 minus 1880–1900 (°C)	2019–2023 minus 1991–2020 (°C)	Total linear change from 1980 to 2023 (°C)
ERA5	N/A	0.37	1.09
ESA SST CCI v3	N/A	0.34	1.03
ERSSTv5	1.00	0.37	1.20
COB2-SST	1.01	0.34	1.08
HadSST4	1.19	0.37	1.33
Mean	1.10	0.35	1.15

[9] The four Niño regions are Niño 1+2 (0–10°S, 90–80°W), Niño 3 (5°S–5°N, 150°–90°W), Niño 3.4 (5°S–5°N, 170–12°0W), and Niño 4 (5°S–5°N, 160°E–150°W).

References

Booth, B., et al., 2021: Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature 484, 228–232. doi.org/10.1038/nature10946

Fox-Kemper, B., et al., 2021: Ocean, Cryosphere and Sea Level Change. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [Masson-Delmotte, V., et al., (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Hartmann, D. L., et al., 2013: Observations: Atmosphere and Surface. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T. F., et al., (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. doi.org/10.1017/CBO9781107415324

Keil, P. et al., 2020: Multiple drivers of the North Atlantic warming hole, Nature Climate Change, 10, 667–671. doi.org/10.1038/s41558-020-0819-8

Kennedy, J. J., 2014: A review of uncertainty in in situ measurements and data sets of sea surface temperature: Reviews of Geophysics, 52, 1–32. doi.org/10.1002/2013RG000434