Abstract
Global mean sea surface temperature (GMSST) is a fundamental diagnostic of ongoing climate change, yet there is incomplete understanding of multi-decadal changes in warming rate and year-to-year variability. Exploiting satellite observations since 1985 and a statistical model incorporating drivers of variability and change, we identify an increasing rate of rise in GMSST. This accelerating ocean surface warming is physically linked to an upward trend in Earth’s energy imbalance (EEI). We quantify that GMSST has increased by 0.54 ± 0.07 K for each GJ m–2 of accumulated energy, equivalent to 0.17 ± 0.02 K decade‒1 (W m‒2)‒1. Using the statistical model to isolate the trend from interannual variability, the underlying rate of change of GMSST rises in proportion with Earth’s energy accumulation from 0.06 K decade–1 during 1985–89 to 0.27 K decade–1 for 2019–23. While variability associated with the El Niño Southern Oscillation triggered the exceptionally high GMSSTs of 2023 and early 2024, 44% (90% confidence interval: 35%–52%) of the +0.22 K difference in GMSST between the peak of the 2023/24 event and that of the 2015/16 event is unexplained unless the acceleration of the GMSST trend is accounted for. Applying indicative future scenarios of EEI based on recent trends, GMSST increases are likely to be faster than would be expected from linear extrapolation of the past four decades. Our results provide observational evidence that the GMSST increase inferred over the past 40 years will likely be exceeded within the next 20 years. Policy makers and wider society should be aware that the rate of global warming over recent decades is a poor guide to the faster change that is likely over the decades to come, underscoring the urgency of deep reductions in fossil-fuel burning.
https://iopscience.iop.org/article/10.1088/1748-9326/adaa8a
Of course the news stories were all teed up ready to publish the moment the embargo was lifted.
Ocean-surface warming four times faster now than late-1980s
Ocean temperature rise accelerating as greenhouse gas levels keep rising
The surface of our oceans is now warming four times faster than it was in the late 1980s
Another day, another climate “crisis” declared by scientists who seem more interested in pushing policy than in practicing rigorous, unbiased science. The latest entry in this parade of alarmism comes from Merchant et al., who claim to have proven that sea surface temperature (SST) is accelerating at an alarming rate due to Earth’s Energy Imbalance (EEI). Their solution? The usual—immediate and severe reductions in fossil fuel use. But before we surrender modern civilization to the dictates of climate activists, let’s take a closer look at this paper and see if its conclusions hold up.
Spoiler: They don’t.
The House of Cards Built on Uncertain Data
The foundation of Merchant et al.’s argument is that Earth’s oceans are warming faster than before, and this acceleration is due to a growing energy imbalance. The problem? The data they use to reach this conclusion is riddled with uncertainty.
They rely heavily on satellite observations to measure EEI, the supposed imbalance between incoming solar radiation and outgoing heat. However, they openly admit that the absolute accuracy of these measurements is not precise enough to detect imbalances as small as ~1 W/m² with confidence. In other words, they are trying to measure a trend that is smaller than the margin of error in the instruments they use. That alone should be enough to dismiss this study.
But it gets worse.
Because the data is unreliable, they supplement it with “reconstructed” EEI estimates before 2000 using a mix of proxy data and modeling. This means that nearly half their dataset is not even direct observation—it’s modeled guesswork. And yet, they use it to claim they can measure multi-decadal acceleration trends with certainty.
Acceleration? Or a Statistical Magic Trick?
If you go looking for acceleration, you’ll probably find it—especially if you design your statistical model to guarantee that outcome. That’s exactly what Merchant et al. have done.
They test three models to explain sea surface temperature trends:
- Linear warming model – Predicts a steady, slow warming trend.
- Quadratic model – Assumes an accelerating trend.
- EEI-driven model – Uses their highly uncertain EEI data to “explain” why warming is accelerating.
Unsurprisingly, they find that the model assuming acceleration fits best, which is exactly what they wanted to prove in the first place. This is not science—it’s circular reasoning disguised as research.
What they don’t do is consider alternative explanations for the changes in sea surface temperature. Natural ocean cycles like the Pacific Decadal Oscillation (PDO) and Atlantic Multidecadal Oscillation (AMO) can drive temperature fluctuations on multi-decadal timescales. These are well-documented, yet Merchant et al. ignore them entirely, preferring to attribute every fraction of a degree increase to human emissions.
Exaggerating Future Warming for Maximum Alarm
The authors then take their dubious acceleration trend and extrapolate it forward to claim that future ocean warming will far exceed previous estimates. They construct three scenarios:
- On-trend – EEI continues rising, leading to catastrophic warming.
- Moderate – Warming still accelerates but not as fast.
- Mitigated – Even with severe emissions cuts, warming will still accelerate.
This is nothing but a glorified version of the climate doomsday scenario playbook—start with a model that assumes acceleration, plug in arbitrary assumptions about future emissions, and produce worst-case predictions that just happen to align perfectly with the climate policy agenda.
Ignoring Uncomfortable Contradictions
The study claims that EEI has been rising since around 2010 and that this is driving SST acceleration. But even they acknowledge that anthropogenic aerosol reductions (i.e., cleaner air) could be responsible for part of the observed trend. If less pollution is allowing more sunlight to reach the ocean, wouldn’t that mean past warming was suppressed by pollution, rather than proving CO₂-driven warming is out of control?
And if that’s true, wouldn’t this mean that climate models have been overestimating the role of CO₂ all along? Merchant et al. conveniently ignore these contradictions because they don’t support their acceleration narrative.
The Real Purpose of This Study: Climate Policy, Not Science
The biggest red flag in this paper is the authors’ final conclusion:
“Policy makers and wider society should be aware that the rate of global warming over recent decades is a poor guide to the faster change that is likely over the decades to come, underscoring the urgency of deep reductions in fossil-fuel burning.”
Notice how they jump from a scientific claim to a policy demand without hesitation? This is the true goal of the paper: to manufacture an urgent climate crisis that justifies drastic intervention.
No mention of alternative explanations for the observed warming.
No discussion of uncertainties in their methods.
No consideration of the costs or consequences of their proposed policies.
Just a predetermined conclusion wrapped in the appearance of scientific rigor.
Final Verdict: Junk Science in Service of Activism
This paper is not an objective scientific analysis—it is an advocacy document masquerading as research. It relies on uncertain data, manipulates statistical models to reinforce a preferred narrative, and exaggerates future warming to scare policymakers into action.
The truth is, given the uncertainties involved, nothing we are seeing is cause for concern. The climate has always fluctuated, and minor changes in sea surface temperature are entirely within natural variability. The only thing accelerating here is the desperation of climate activists to keep their funding and political influence intact.
The next time you see a breathless headline about “accelerating ocean warming”, remember: bad models, biased assumptions, and agenda-driven science are the real drivers behind these claims—not reality.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
No, it’s not. Global mean sea surface temperature is a fundamental diagnostic of oceanic latitude. And that is all that can be said with certainty.
Climate change is like a ghost or a hobgoblin or a unicorn. It’s an abstract concept that doesn’t exist in reality and can’t “cause” anything, yet gets blamed for everything.
Oceans are boiling, therefore their temperatures cannot increase until they are completely vaporized.
Of course oceans are boiling.
Just look at the steam coming off them.
AlGore – ithm is right!
uh .. my life might be too short to really look into such papers, but one sentence in their abstract here triggered a very loud alarm bell.. “””Using the statistical model to isolate the trend from interannual variability, the underlying rate of change of GMSST rises in proportion with Earth’s energy accumulation from 0.06 K decade–1 during 1985–89 to 0.27 K decade–1 for 2019–23. “””
I am not sure how they analyze their “results” but I would expect the EEI 2019-23 to be different than basically any other period given that the world’s largest ever recorded volcanic eruption (Tonga, Jan 2022) changed the stratospheric water content by more than 10% (among other effects).
Could their analysis be biased horribly by omitting this factor?
Uh, well I couldn’t resist..
Merchant’s paper (https://iopscience.iop.org/article/10.1088/1748-9326/adaa8a/pdf) states
“The Hunga Tonga eruption of 2022 is neglected, since the net GMSST effect is likely to be small [10].”
Which is okay for the sea temperature, but problematic when a correlation to EEI is made
J. Jessop has a blog on the question if the Tonga eruption had a massive effect on the EEI:
https://jaimejessop.substack.com/p/the-science-is-clear-warming-since?utm_campaign=post&utm_medium=web
(apparently also engaging A. Dessler who modeled that it didn’t.. in an effort to defend high CO2 climate sensitivities )
Jessop points to a preprint by G. Tselioudis et al. (https://www.researchsquare.com/article/rs-5050391/v1) hinting that cloud cover changes are responsible for changes in EEI.
According to Jessop a very clear sign that Tonga is responsible for a very significant portion of that change
I am guessing I guessed right earlier 🙂
“The Hunga Tonga eruption of 2022 is neglected, since the net GMSST effect is likely to be small [10].”
Isn’t that just precious. Likely to be small. They don’t know, and don’t bother to look. But that’s ok because they already know the answer.
BTW, the signal that they are looking for is also small.
Two lots of only 5 years not even from the beginning of the satellite era. Could it be any more cherry picked?
The second period contains a major El Nino…. The first does not.
It is a silly, meaningless comparison
“Policy makers and wider society should be aware that the rate of global warming over recent decades is a poor guide to the faster change that is likely over the decades to come, underscoring the urgency of deep reductions in fossil-fuel burning.”
But then the authors expect to keep on using FF every minute of every day, naturally. Where do they think these cuts will come from?
I’m not sure I understand how a minuscule rise (0.3% Kelvin) in AIR TEMPERATURE can soak down to affect the ocean abyss? With its rather high specific heat. But I’m just a silly sailor
Probably underwelling!
Adjustments galore. Went looking for accuracy and just what they are able to measure after a post of Bob’s Dec, 2014 and had a real awakening. Seems that overall measuring of ‘sea surface’ has problems. Original bucket & thermometer (no depth control), ship intake (well below surface and varied depth due to load +/-20 feet), buoys (seem to rock in the wave with depth resolution of a meter), then IR satellite (cannot get thru the clouds) to microwave (get thru the clouds, but not the rain & surface mist). Oh and did I mention one of the satellites was doing reasonable until they had to boost the altitude, then had problems with pitch, yaw and just had no idea the height it was flying. The number of adjustments to correct is staggering. Includes (but not limited to); wind speed, rain, cloud amount/percent and cloud water vapor, daytime diurnal warming, high latitudes, aerosols, SSTs <10C, columnar water vapor, higher latitudes show a slight warm bias, seasonal cycle wind direction for SST retrieval, fast moving storms and fronts, wind direction error and instrument degradation.
http://images.remss.com/papers/rsspubs/gentemann_jgr_2014.pdf
Still their abstract reads –
Errors were identified in both the MW and IR SST data sets: (1) at low atmospheric water vapor a posthoc correction added to AMSR-E was incorrectly applied and (2) there is significant cloud contamination of nighttime MODIS retrievals at SST <10C. A correction is suggested for AMSR-E SSTs that will remove the vapor dependency. For MODIS, once the cloud contaminated data were excluded, errors were reduced but not eliminated. Biases were found to be 20.05C and 20.13C and standard deviations to be 0.48C and 0.58C for AMSR-E and MODIS, respectively. Using a three-way error analysis, individual standard deviations were determined to be 0.20C (in situ), 0.28C (AMSR-E), and 0.38C (MODIS).
Story Tip: Study reveals 80 000 people could die yearly by 2100 due to dangerously hot temperatures
From 2000 to 2019 about 4.6 million people died from cold-related causes and about 490,000 people died from heat-related causes.
‘Global, regional, and national burden of mortality associated with non-optimal ambient temperatures from 2000 to 2019: a three-stage modelling study’
https://www.thelancet.com/action/showPdf?pii=S2542-5196%2821%2900081-4
Cold causes the blood vessels to constrict to conserve heat and this raises blood pressure causing more strokes and heart attacks in the colder months.
“We wanted to test this. And we show clearly that we will see a net increase in temperature-related deaths under climate change.”
“In all scenarios analyzed, the number of heat-related deaths is projected to rise over time, though researchers caution that there is still some uncertainty in the data.”
They show clearly with the caveat that there is uncertainty.
FYI: Scenarios = models.
> “Our results provide observational evidence that the GMSST increase inferred over the past 40 years will likely be exceeded within the next 20 years.”
That’s not how projections work. Certainly not without also expanding uncertainty to the point of being a meaningless futurecast.
Regardless, these observations have no causality applied so any tie to fossil fuel use is beyond spurious.
Same quote in my copy-paste buffer for the same reason.
This sort of nonsense ‘research’ will not stop until the funding does.
Fortunately, in the US it has been ordered to stop at 1700 today, per Trump’s first press conference at 1300 EST. No Fed grants to states, NGO’s, Universities for DEI, wokeness, transgenders, green new scam, illegal aliens,… The wailing is already great.
As a particularly funny example, a trans ‘woman’ in MA already sued Trump 47 because he is being moved from a women’s prison to a man’s prison—arguing cruel and unusual punishment under 8A! Don’t think that is gonna get very far, even in MA.
Baits people to ask inappropriate questions like “if there’s really no difference between the two, then why should I care?”
It might get far in MA- this place is nuts. Just watched online a committee set up by the state’s enviro and energy agency. It’s mostly women of course. Their talk was focused on “the energy transition”. It was 2 hours long and you’d never know watching it that Trump is in the White House crushing everything they believe in.
“Trans-women are women”
If true, why include “trans.”
With the new WE greenhouse efficiency metric, we know any warming seen over the last 25 years was 100% due to increases in absorbed solar energy. Hence, this paper is obvious academic fraud. Wouldn’t be nice if the new administration put together a team of scientists to investigate these clowns and the journal which published what is clearly nonsense?
The idea that an increase in the atmospheric concentration of CO2 or other greenhouse gases could heat the oceans was introduced into the climate models in the late 1970s. For example, the 1979 paper by Manabe and Stouffer claimed an ocean warming of about 4 °C for a quadrupling of the CO2 concentration. The surface energy transfer, particularly the wind driven evaporation was ignored and the ‘slab’ ocean simply added heat capacity to the surface thermal reservoir. This was a flat ocean without wind or waves. In the real world, the oceans cannot be heated by an increase in the atmospheric concentration of greenhouse gases.
The ocean surface is almost transparent to the short wave solar radiation. Approximately 90% of the solar energy is absorbed in first 10 meter layer of the ocean. The increase in surface temperature during the day is small, typically 1 °C or less. The downward long wave IR (LWIR) flux from the lower troposphere to the surface ‘blocks’ most of the upward LWIR flux emitted by the surface. The net LWIR cooling flux is limited to the LWIR flux from the surface that is emitted into the LWIR atmospheric window. In order to dissipate the excess absorbed solar heat, the bulk ocean temperature increases until the heat is removed by wind driven evaporation (latent heat flux). Within the ±30° latitude bands, the long term sensitivity of the latent heat flux to the wind speed is near 15 W m-2/m s-1. For each increase in wind speed of 1 meter per second, the increase in latent heat flux is 15 Watts per square meter.
The penetration depth of the LWIR radiation into the oceans is less than 100 micron (0.004 inches). Here it is fully coupled to the wind driven latent heat flux. The net LWIR flux and the latent flux should not be separated and analyzed independently of each other. At present, the increase in the atmospheric CO2 concentration is approximately 2.5 parts per million (ppm) per year. The corresponding increase in downward LWIR flux to the surface is near 40 milliWatts per square meter per year (40 mW m-2 yr-1). This is dissipated by an increase in wind speed near 3 millimeters per second. In the equatorial Pacific Ocean, long term average wind speeds are typically 5 ±2 m s-1.
There is no requirement for an exact energy balance at the ocean surface between the absorbed solar energy and the surface cooling. Natural variations in wind speed produce quasi-periodic variations in ocean surface temperature such as the El Nino Southern Oscillation (ENSO). The inverse relation between the ENSO and the Southern Oscillation Index (SOI) is well known. The SOI is based on the surface pressure difference between Tahiti and Darwin, Australia. It is an indicator of the wind speed in ENSO region of the Pacific Ocean. The 2016 ENSO peak was produced by a decrease in wind speed near 2 m s-1 over a period of about 6 months. It is not just the surface temperature that changes. The ocean warms down to depths of at least 75m. The heating is produced by a decrease in latent heat flux at the surface of approximately 30 W m-2. There is additional heating related to a decrease in the speed of the S. Pacific equatorial current.
This is discussed in more detail in the book ‘Finding Simplicity in a Complex World’ (2023). The 2016 ENSO peak is explained at the bottom of the home page on the book website.
Ocean heating by greenhouse gases is a fundamental error in all of the ‘equilibrium’ climate models from the late 1970s onwards. This also means that there can be no ‘Climate Sensitivity’ to CO2.
“Ocean heating by greenhouse gases is a fundamental error in all of the ‘equilibrium’ climate models from the late 1970s onwards. This also means that there can be no ‘Climate Sensitivity’ to CO2.”
The oceans aren’t “heated” byGHGs.
They reduce the ability to cool.
Same effect they have on the Earth in general.
The ocean skin is warmed however, and the reduction in deltaT between lower water and the skin means a lower heat flux to be in contact with the air.
Hence a slower cooling….
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017JC013351
“The additional energy from the absorption of increasing IR radiation adjusts the curvature of the TSL such that the upward conduction of heat from the bulk of the ocean into the TSL is reduced. The additional energy absorbed within the TSL supports more of the surface heat loss. Thus, more heat beneath the TSL is retained leading to the observed increase in upper ocean heat content.”
Pass the collection plate.
Is there no way to hold the authors of crappy reports accountable?
Not when the journal editors are part of “the cause”.
When they talk about “Global mean sea surface temperature”- how do they define surface? Even if it goes up a bit- isn’t the average depth of the ocean something like 12,000 feet! Seems like a mosquito bite to an elephant.
Average ocean depth is just shy of 4km.
https://ecowatch.noaa.gov/thematic/sea-surface-temperature
Survival times in the North Sea have not improved.
Ok. So the study is garbage, but they got paid, right? Probably by you and me. That is the goal of all such guys.
The authors are of the UK so maybe you paid if that’s where you pay taxes. I’m in the USA so likely didn’t pay for this one. Of course, I don’t know where their support came from.
Don’t be so sure US money didn’t find its way.
Biden funded a lot of UN climate groups, so it is possible US money paid in part or in full for this.
In climate4you dot com Ole Humlum has charts for global cloud cover. Click the climate+clouds button and see for yourself. Cloud cover is down by a few percent in recent decades. Hmmm.
If they had taken the time to consider their observations in more detail, they would have observed that the SST of oceans of the NH have increased by 0.7C over the satellite era. The SH oceans have only increased by 0.2C.
Then if they cared to look at how the peak solar intensity is shifting northward across the globe, they would realise that the precession cycle is responsible for the observed changes.
In fact the Southern Ocean has actually cooled over the satellite era.
Since ARGO began in 2004, there has been no acceleration in the ocean warming. The CERES data was calibrated to the ocean warming from 2005 to 2015. I pointed out a couple of years ago that CERES needs to be recalibrated because it is parting ways with OHC. This whole scary story is based on a calibration error or misunderstanding where heat is accumulating or being lost.
This chart shows that the oceans of the northern hemisphere have only retained 7.9% of the net radiation absorbed into and above the water surface.
?ssl=1
So the heat retention in NH oceans is dominated by heat advection to land. Small changes in advection, caused by changing solar intensity, will have more influence in the NH than in the SH.
Another point of comparison: (sea-surface temperature vs. lower-troposphere [over oceans] )
We quantify that GMSST has increased by … + 0.17 ± 0.02 K decade‒1 — this source
— vs. —
Global area-averaged temperature trend (Jan. 1979 through Dec. 2024) remains at + 0.13 C/decade over oceans) [lower troposphere]
— https://www.drroyspencer.com/
Also, at depths below the ‘sea surface’, the trend is much smaller (Annual Report compiled by Ole Humlum).
So it seems this new report is an outlier.
Story Tip
Floating solar caused increased greenhouse gas emissions on small ponds, study finds | Tallbloke’s Talkshop
Has Professor Chris Merchant, lead author at the University of Reading and National Centre for Earth Observation, taken into account the main longer-term (~60+ years) oceanic overturning cycles such as the AMO in his deliberations?
It would be rather embarrassing if the AMO, which last flipped to -ve in the early 1960’s, flips -ve again in the next decade.
This paper isn’t doing anything except stating the obvious. The earth is warming and the only source of energy that can be doing that is the sun. Hence there must be an energy imbalance at the top of the atmosphere and not surprisingly the increase in the temperature of the ocean follows the imbalance in the energy balance. Anything else would violate the laws of thermodynamics.
The obvious often needs to be presented to AGW cultists Izaak.
As in –
“It’s The Sun, Stupid”
But you are missing the point. So while the sun is clearly the source of the incoming energy the reason why there is an energy imbalance is not the subject of the paper.
In fact they are very careful to avoid any discussion of what is causing the energy imbalance. They go as far to say that “This leaves unanswered the important question of what has caused the EEI trend. While the investigation of this crucial question is beyond the scope of this paper…”
So all we are left with is the completely unsurprising result that if there is an energy imbalance at the top of the atmosphere then the planet responds by warming up. Nobody should be surprised by such a completely banal and obvious conclusion. As I said anything else would violate the conservation of energy.
You seem to have missed this point:
“However, they openly admit that the absolute accuracy of these measurements is not precise enough to detect imbalances as small as ~1 W/m² with confidence. In other words, they are trying to measure a trend that is smaller than the margin of error in the instruments they use. That alone should be enough to dismiss this study.”
The definition of an intensive property includes that it is constant throughout the body. An average of intensive properties of samples is meaningless otherwise, unless sampling every little bit in uniform infinitesimal sizes. A practical size would be where the variation throughout the sample is much less than the precision you want for the result (and measured with that sort of precision). Creating a profile from the much less data that you have and pretending that you sampled enough is valid, but likely to introduce a systematic error.
If a trend in changes in local temperatures were just due to the change in a global average, that would be like a constant intensive property. But it’s not. The heat gets distributed differently over the decades, so the same problems as averaging the temperature measurements.
You need to take the results with a pinch of salt. An error analysis as if it’s a true intensive property is not good enough. You can’t trust experts whose checking consists solely of whether the results show warming not being consistent with a rise in global CO2 after 1960 being the cause.
Precision, or accuracy? The two are different, and it matters. I believe you are talking about accuracy, not precision.
I’m not going to bother looking this up, however . . .
In one of Richard Feynman’s lectures he admonished researchers to try to identify all the alternative reasons for an event other than the one favored. If the alternatives are not identified and refuted your result is poor. “The first principle is that you must not fool yourself—and you are the easiest person to fool.”
Seems this paper suffers from these problems.
Thanks CTM.
The expression analysis of alternatives is one I am familiar with over the past 5 decades.
You are correct. Fenman is correct.
CERES’s OLR data when seasonal trends are removed shows only noise at the 1 sigma level about the mean and no trend. This was explained at WUWT sometime ago. ARGO data finds a OHC rise corresponding to about 0.05 C in the first decade of measurement. That is for the 0-700m depth. Unfortunately, ARGO reaches only between 60N and 60S, missing most of the Southern ocean and all the Arctic ocean, the most important 25% of the ocean. This puff piece supporting global warming is not interesting, except it writes that the ‘rise’ is unexplained.
It is likely non-existent or the teams acquiring the data would have made the discovery.
“CERES’s OLR data when seasonal trends are removed shows only noise at the 1 sigma level about the mean and no trend. “
That is not what this paper found …..
https://iopscience.iop.org/article/10.1088/2752-5295/ac6f74
“Monthly time series of CERES EBAF Ed4.1 net radiation at the TOA (positive down) relative to an estimated mean of 0.7 W m⁻² for 2000–2015 (see Loeb et al (2018)) (blue). The mean annual cycle is removed, and the heavy black line is the 12 month running mean (includes data through 2021). The total solar irradiance contribution is in red (updated, see Coddington et al (2016)) with the mean (1361 W m⁻²) removed, and converted to a radiative forcing by dividing by 4 and multiplying by 0.7 to account for the albedo.”
Lots of assumptions and averages in that.