- JoNova - https://joannenova.com.au -

Far Southern Ocean cools. Kiss Goodbye to polar amplication around Antarctica

Antarctic, Map, Latitudes of sea surface that are cooling.

A map of the sea surface zone that has cooled since 1979 — from 56S – 72S . It’s a pretty big area. Click to enlarge.

For years the IPCC have said that warming would be amplified at the poles. They warned us things would heat up twice as fast, which would melt sea ice. The oceans surface in turn would switch from being reflective white to a dark absorbing deep blue. Enormous amounts of energy would then flow into the ocean instead of being reflected back out to space. The more it warmed, the more it would warm — unleashing a devastating feedback loop.

As the Arctic warmed, the merchants of doom were keen to tell us how how right they were and this was evidence of man-made warming. But in the Antarctic exactly the opposite trend was taking place.

Mike Jonas has done what the IPCC should have been doing — investigating the trends in the Sea Surface Temperature in the polar latitudes with satellite records. In the latitude band from 56 to 72 degrees south the oceans have cooled, not warmed. The models don’t even have the sign of the trend correct. At the latitudes where the models expected the most warming, the ocean surface cooled as fast as a tenth of a degree per decade. For the sea surface, that’s surprisingly quick.

It’s more evidence that things are seriously wrong in the global models. The modelers don’t understand the climate, they can’t predict the major processes of ocean currents, cloud changes and albedo. How can they even say they “know” what drives changes in the Arctic if their same models fail so badly on the Antarctic? They claim random luck as “success” and throw a veil of silence over the non-random failures.

As McKitrick and Christy point out:

Swanson (2013) noted that the changes in model output between CMIP3 and CMIP5 improved the fit to Arctic warming but worsened it everywhere else, raising the possibility that the models were getting the Arctic right for the wrong reasons. …

When Jonas tried to get this significant finding published in the peer review, the usual gatekeeping process meant these simple but cutting graphs were rejected — without a right of reply (that’s another story for another day).

But here, I’m happy to publish his work in the exact form he submitted for peer review (so we can discuss the peer review process itself. Though I would have suggested some edits).  Thanks to Mike Jonas for all his work!



Southern Oceans Sea Surface Temperatures contradict a key element of the IPCC Report

Author: M Jonas  (Cover note of submission)

Author’s affiliations: None


The hypothesis – Surface warming is amplified by sea ice- and snow-related feedbacks near the poles – is supported by climate models and was an important factor in the fifth IPCC report. The sea ice part of this hypothesis was tested, using Sea Surface Temperatures of the Southern Oceans. The test showed clearly that the sea ice part of the hypothesis is contradicted by the data, because there was quite strong overall cooling in the latitudes where amplified warming was expected. There must therefore be one or more important large-scale climate processes that are not reasonably represented in the models. From this, it necessarily follows that the climate models are invalid and their Antarctic projections in particular are now untenable. It also necessarily follows that the climate models’ global projections are unreliable.


3poly – 3rd-order polynomial fit, AR5 – Annual Report #5, IPCC – Intergovernmental Panel on Climate Change, NOAA – (US) National Oceanic and Atmospheric Administration, SST – Sea Surface Temperature, WG1 – Working Group #1.

Standard convention is followed re latitudes: “lower latitudes” are nearer the equator, “higher latitudes” are nearer the poles. All temperatures are in deg C.


A significant proportion of the anthropogenic global warming as projected by climate models occurs near the poles, and is amplified by feedbacks. Warming in West Antarctica (eg. Steig, 2009), glacier retreat (eg. Favier, 2014) and ice loss (eg. Rignot, 2008) have been reported, and have been explicitly or implicitly linked by the media and others to the models’ projected global warming – eg. Amos J (2014), Press Association (2015), Fox D (2017).

Sea surface temperatures of the Southern Oceans are examined to determine whether they behave as projected. The relevant latitudes are where sea-ice does or can exist at times (the “sea-ice latitudes”), but the study covers a larger range of latitudes so that the sea-ice and non-sea-ice latitudes can be compared. The sea-ice latitudes cover a large area, so a failure of the temperature in the sea-ice latitudes to behave as projected would indicate that the climate models’ underlying mechanisms or their implementation are invalid.

The climate system is a complex coupled non-linear system (Baede, 2001). ie, the major climate mechanisms interact with and affect each other. Therefore all major climate mechanisms must necessarily be reasonably well represented in climate models in order for the models’ global projections to be at all reliable.

Note: The ocean latitudes examined in this study are from -30 (30S) southwards and are referred to as the “Southern Oceans” to avoid confusion with the official Southern Ocean which lies south of latitude -60 (60S).


The IPCC’s hypothesis is that : Surface warming is amplified by sea ice- and snow-related feedbacks near the poles

2.1 The Mechanism

The fifth IPCC Report (Collins, 2013) states: “Feedbacks associated with changes in sea ice and snow amplify surface warming near the poles (Hall, 2004; Soden et al., 2008; Graversen and Wang, 2009; Kumar et al., 2010).”. Hall (2004) found that surface albedo feedback accounts for about half the high-latitude response to external forcing by CO2.

The sea ice feedback mechanism is explained in clear simple language by Pinson (2017), referring to the Arctic: “As the ice recedes, retreating further and further north, solar radiation that was once reflected back into space due to the high albedo effect of ice is now warming an increasing amount of water, and the radiant heat this creates further exacerbates both regional and global warming.”. The equivalent feedback mechanism operates for ice or snow on land.

There is little doubt about these mechanisms, and there is high confidence that they influence large areas. The fifth IPCC Report (Collins, 2013) states: “In summary, there is robust evidence over multiple generations of models and high confidence in these large-scale warming patterns.”.

2.2 Sea-ice Latitudes

Sea-ice feedback can only occur in places where there is a change in sea-ice.

Sea ice feedback operates almost exclusively in summer, as there is little insolation in winter (Morales Maqueda, 2018). The place where the sea ice feedback operates must be between the minimum and maximum sea ice extents – ie. the places that have less sea ice cover than they used to have, either because the sea ice extents are smaller than they were, or because they have the same coverage for less time each year. Visual inspection of Figure 1 shows that the Antarctic sea ice extent is mostly between about 76S (Ross Sea, RH picture) and 56S (Weddell Sea, LH picture). Any sea ice feedback must therefore originate in or near this latitude range, and it must affect the sea surface temperature (see 6.4 Melting from below).


Antarctica, map, sea ice extent.

Figure 1. Antarctic 1979-2000 sea ice concentration (Lindsey, 2000). The 1979-2000 medians are shown with a yellow line.



Note: The hypothesis being tested is from the IPCC report. The test reported here is therefore not a direct test of the climate models, and in particular it is not a test of any particular climate model. It is a test of certain important statements in the IPCC report. However, because the IPCC report is obviously derived from and supported by climate models, it is reasonable to draw conclusions about the accuracy of the models as a whole, using the demonstrated accuracy (or inaccuracy) of the IPCC’s climate expectations.

The test is conducted by comparing the Sea Surface Temperature (SST) trends by latitude. This use of latitude is particularly suitable for the Southern Oceans, because sea surface temperatures in the Antarctic correlate very well with latitude, with the isotherms approximately following the parallels (Morales Maqueda, 2018).

If the hypothesis being tested is correct, then the SST trends for the sea-ice latitudes (about -76 to -56, see 2.2 Sea-ice Latitudes) should be higher than for other latitudes, and in particular they should be higher than for lower latitudes and, because the test is for amplified warming, they should be higher than the global average.

In the Southern Oceans’ sea-ice latitudes, where any amplified warming from sea-ice feedback must originate, exactly the opposite is shown by the Sea Surface Temperature data (Reynolds, 2002). See Figures 2 and 4. For information about the data and how it was used, see A1 The Data.

Antarctica, map, sea ice extent.

Figure 2. Linear trends in sea surface temperature around the Antarctic, by Latitude, for the four longitude lists described in A1 The Data. High latitudes are on the left, low latitudes on the right.


There is no amplified warming in the sea-ice latitudes (about -76 to -56). In fact, the temperature trend there is mostly quite strongly negative, and is in stark contrast to the lower latitudes where there is significant warming. In Figure 2, an increase in trend can be seen from about latitude -65 southwards, but inspection of the Y axis shows that this increase is from quite a strong cooling rate to around zero. From around latitude -70 to -55, there is cooling, not amplified warming. There is more discussion of this in 6.2 High-latitude temperature trends.

The data used is for the satellite period from Dec 1981 to Oct 2017, covering 35 full years or 36 southern-hemisphere summers. This period should be sufficient. Hall (2004) says “the portion of the scenario run corresponding to the present-day satellite record is long enough to capture [the sea ice- and snow-related] feedback.”. That paper was received in June 2003, and there have been a further 14 years of satellite data since then.

So, the amplified warming from sea ice feedback, as projected by the computer models and as reported by the IPCC with high confidence, is not happening. There could be some limited local sea ice feedback, but there is clearly no large-scale warming. The robust evidence over multiple generations of models must come from errors in the models.


In this section, the evidence will be looked at in more detail, and checked for any feature which might throw a different light on the broad-area analysis above.

4.1 Other Areas

The data used in the above tests was limited to latitudes with complete data for a reasonable distribution of longitudes. Latitudes south of -72 (72S) and longitudes with some land were excluded because they did not have enough complete data. Data is missing where there is ice (sea or land) or snow.

These excluded areas do need to be tested. There are three such areas: The tip of the West Antarctic Peninsula has land which cuts out most of the lines of longitude in that sector, so it is not well represented in Figure 2. The Ross and Weddell seas have very few lines of longitude without any summer sea ice, and are therefore poorly represented in Figure 2. For maps showing these areas, see A2 Supplementary Information.

These areas were tested by replacing all missing data with values of -2.78. This value was chosen because the lowest non-ice temperature in the input data is -1.78. A grid cell that had ice in earlier years but lost it in later years would therefore be treated as if it had warmed by at least 1 degree. An area continously covered with ice would be treated as having constant temperature. NB. The average temperatures and trends in Figure 3 are not real, because of the artificial value used for sea ice, but if there was any amplified warming in these areas then it should show up. It didn’t.


Antarctica, map, sea ice extent.



Antarctica, map, sea ice extent.



Antarctica, map, sea ice extent.

Figure 3. Average summer sea surface temperature in (a) an area around the tip of the West Antarctic Peninsula, (b) the Ross Sea, and (c) the Weddell Sea, from 1982 to 2017, with 3rd-order polynomial fit and linear trend. Grid cells with missing data have been given temperature -2.78.Note: Any 3rd-order polynomial fit is for illumination only, it is not used.


4.2 Possible Bias

It is possible that bias could have been introduced in this study by restricting the sea surface analysis to longitudes with complete data. The rationale is that a location with sea ice in the early years, but that had no ice in later years because it had warmed, would be excluded.

This possible bias was tested for, using the same method as in 4.1 Other Areas. All longitudes were used, across all latitudes with any data, ie. from -78 to -30 (78S to 30S).

Antarctica, map, sea ice extent.

Figure 4. Linear trends in sea surface temperature by latitude across all longitudes, from 1982 to 2017. Grid cells with missing data have been given temperature -2.78.


The pattern is very similar to those in Figure 2, except that the temperature trend falls in latitudes higher than those present in Figure 2. Those latitudes were absent from Figure 2, because of seasonal sea ice. Figure 4 confirms that there is no amplified warming from sea ice feedback over the southern oceans in general.


The hypothesis that surface warming is amplified near the poles by sea ice feedback is contradicted by the data. Warming is occurring in the lower latitudes, but the data clearly shows that little or no warming is occurring in higher latitudes approaching the Antarctic continent, and that over most of those latitudes there is net cooling.

That such a large area of the Southern Oceans is behaving so differently from the models’ projections for a significant period of time implies unequivocally that those climate models’ underlying mechanisms or their implementation are invalid. As explained in 6.1 Model projections and observations and 6.3 Model implications, the climate models’ Antarctic projections are now untenable. From this it necessarily follows, as explained in 1. INTRODUCTION, that the climate models’ global projections are unreliable.


6.1 Model projections and observations

The IPCC Report (Collins, 2013) is very clear: “In equilibrium simulations, amplified warming occurs in both polar regions.”.

It would be reasonable to suppose that 36 years is enough time for the transient response to at least start to move towards equilibrium, and the models do indeed project this: “model trends in surface air temperature are 2.5 to 5 times higher than observed over Antarctica”. The IPCC report provides the following explanation: “The lack of an amplified transient warming response in high Southern polar latitudes has been associated with deep ocean mixing, strong ocean heat uptake and the persistence of the vast Antarctic ice sheet”. But the report then goes on to suggest that the model projections are more accurate than the temperature data: “but here also the observational estimates have a very large uncertainty, so, for example, the CMIP3 ensemble mean is consistent with observations within error estimates (Monaghan et al., 2008)”. The primary justification for this argument is “because station records are short and sparse”.

There are now 36 years of satellite sea surface data uniformly covering the relevant Antarctic latitudes. They show no amplified warming. In fact, they show little or no warming at all. The uncertainty of “short and sparse” land station records is no longer a credible argument for the models’ projections significantly overestimating temperature. Those model projections are now untenable. NB. This argument is not about the uncertainty of the land station records (see A2 Supplementary Information), it is that there is now an additional source of data.

The IPCC’s explanation for the models’ failure, which associated the lack of warming with “deep ocean mixing, strong ocean heat uptake and the persistence of the vast Antarctic ice sheet”, raises another issue: If that explanation is correct, then they have identified some important large-scale climate processes that are not reasonably represented in the models. Without such processes, it is questionable whether the models are fit for purpose.

6.2 High-latitude temperature trends

In Figures 2 and 4, there is a noticeable increase in temperature trend from about latitude -65 (65S) to -72 (72S). But, as noted in 3. THE TEST, this increase is from quite a strong cooling rate to around zero. The average temperatures are taken over the whole 360 degrees of longitude, because the hypothesis being tested relates to the whole area, not to isolated pockets. It is reasonable to expect the overall trend to be nearer to zero at the latitudes with more land (that’s because there is less sea area to provide sea-ice feedback warming for the whole area).

There is a small overall warming in latitudes between about -75 (75S) and -70 (70S). This presumably indicates that there is a pocket (or pockets) of warming between those two latitudes. But the existence of pockets of warming is not enough to confirm an expectation of amplified warming near the poles: that expectation clearly applies to the area as a whole and not just to pockets. In order for the expectation to be met, there would have to be amplified warming over the sea-ice latitudes as a whole. The exact opposite has occurred – the sea-ice latitudes have cooled significantly.

If any of the three separate ocean areas – tip of WA Peninsula, Ross, Weddell – examined in 4.1 Other Areas had shown significant warming, that could conceivably have affected the overall picture of no amplified warming, or they might have given an interesting picture leading to a more nuanced conclusion. They didn’t.

6.3 Model implications

As stated in 1. INTRODUCTION: “Sea surface temperatures of the Southern Oceans are examined to determine whether they behave as projected. The relevant latitudes are where sea-ice does or can exist at times (the “sea-ice latitudes”), but the study covers a larger range of latitudes so that the sea-ice and non-sea-ice latitudes can be compared. The sea-ice latitudes cover a large area, so a failure of the temperature in the sea-ice latitudes to behave as projected would indicate that the climate models’ underlying mechanisms or their implementation are invalid.”.

In more detail:

The climate models operate on a 3D matrix of atmosphere, ocean, cryosphere and land cells, using the estimated interactions between them to calculate their climatic changes over time (IPCC, 2013).

Climatic changes are calculated over all of these small(ish) cells over very short time steps, typically about 30 minutes (UCAR, 2011). Those calculations are repeated until the required overall period has been covered. That overall period may range from years to centuries. Only when all of the calculations for all of the cells over all of the time steps have been completed can the results for the individual cells be combined into regional or global projections.

The models’ regional projections are therefore aggregated from their cells’ projections, and the models’ global projections are in fact aggregated from their regional projections even if the IPCC do not choose to explain it that way.

One of the modellers’ major problems is that the cell-based mechanism that they use is unable to adequately represent the laws of physics. As the IPCC explains: “many physical processes, such as those related to clouds, also occur at smaller scales and cannot be properly modelled. Instead, their known properties must be averaged over the larger scale in a technique known as parameterization.” (IPCC, 2013). That the modellers fully understand the laws of physics is not in dispute, but they may still introduce inaccuracies when they parameterise them.

It therefore follows that if the models are in serious error for a major region – such as all of the Southern Oceans’ sea-ice latitudes – then their implementation of the physical laws and/or their parameterisation must be incorrect. The models use the same rules everywhere, ie. globally. They have to use the same rules everywhere, because the laws of physics are the same across the whole planet. If those rules are incorrect, then the models’ projections for all regions and globally are unreliable.

The IPCC do recognise that they have a problem. They say “GCMs may simulate quite different responses to the same forcing, simply because of the way certain processes and feedbacks are modelled. However, while these differences in response are usually consistent with the climate sensitivity range described in criterion 1, they are unlikely to satisfy criterion 4 concerning the uncertainty range of regional projections. Even the selection of all the available GCM experiments would not guarantee a representative range, due to other uncertainties that GCMs do not fully address, especially the range in estimates of future atmospheric composition.”.

The fact that the IPCC recognises that it has a problem does not mean that the problem can be ignored. It means that they really do have a problem.

As explained above, the problem is much more serious than is acknowledged by the IPCC. The problem is compounded by the complex non-linear interactions between the various climate mechanisms and regions (Baede, 2001).

6.4 Melting from below

The West Antarctic melting has been reported as occurring from below (eg. Alley, 2016). It is possible that this is evidence of local sea ice or snow feedback, but perhaps unlikely because sea ice feedback must initially warm the surface. In any case, it is of no direct interest to this study, because the test here is for sea ice feedback that has amplified surface warming. For that to occur, there must be warmed water at the sea surface.

6.5 Volcanoes

It could be worth investigating whether the melting from below is principally or partly caused by volcanoes – see eg. Van Wyk de Vries (2017) reported in Murphy-Bates (2017), and Pappas (2014). See also A2 Supplementary Information.

Antarctica, map, sea ice extent.


Figure 5. Map of the West Antarctic volcanic region, from Murphy-Bates (2017)


A1 The Data

The aim was to compare sea surface warming rates between lower and higher latitudes around Antarctica. In order to avoid inaccuracy or bias from incomplete data or from changing sample sizes, data was selected from the satellite record as follows
• Summer-only data. This was to ensure that minimal data was missing because of seasonal sea ice. Sea surface temperature is only useful where there is open water, otherwise the temperature of the sea water is concealed by the ice. The same set of calendar months, Dec-Apr, was used for every temperature comparison to ensure consistency.
• Individual longitudes were discarded if they did not have complete data for the range of latitudes being compared. Data was incomplete where there was ice (sea or land) or snow, or possibly simply missing data.
• Data was only compared beween different latitudes if the set of longitudes with complete data for that range of latitudes was reasonably well populated. The four sets of longitudes used were the longitudes free of summer sea-ice to as far south as, respectively, latitude 65S (“List A”), 70S (“B”), 71S (“C”), 72S (“D”). This was a subjective decision, but the data is all available for inspection.
• The full set of 36 Southern Hemisphere summers 1982 to 2017 was used for all temperature comparisons.

Gridded SST data was downloaded from the USA National Oceanic and Atmospheric Administration (NOAA) (Reynolds, 2002). Each grid cell is 1 degree latitude by 1 degree longitude.

The data is presented in the following files:
SSTsDownloaded3090– Sea surface data as downloaded from NOAA (Reynolds, 2002) for latitudes -30 to -90. SST Data was available only from Dec 1981 to Oct 2017.

The files identified in the appendices are all available online.

A2 Supplementary Information

Only selected temperature graphs are shown in this paper. All of the tested combinations of latitude and longitude are shown in file SupplementaryInformation. This file also has maps showing the locations of the areas used in Figure 3.

File SupplementaryInformation also looks at land station temperature data, and supports the IPCC description “sparse”.


Alley KE et al (2016). Impacts of warm water on Antarctic ice shelf stability through basal channel formation, Nature Geoscience, vol.9 April 2016, doi: 10.1038/NGEO2675

Amos J (2014). ‘Nothing can stop retreat’ of West Antarctic glaciers, BBC News.

Baede APM et al (2001). The Climate System: an Overview, in: Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, (Houghton JT et al, 2001), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 881pp

Collins MR et al (2013). IPCC Report AR5 WG1 Chapter 12 Long-term Climate Change: Projections, Commitments and Irreversibility. 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. Cambridge University Press. Accessed 2017.

Favier et al (2014). Retreat of Pine Island Glacier controlled by marine ice-sheet instability, Nature Climate Change 2014/01/12/online 4 117, Nature Publishing Group

Fox D (2017). The Larsen C Ice Shelf Collapse Is Just the Beginning—Antarctica Is Melting, National Geographic magazine.

Hall A (2004). The role of surface albedo feedback in climate. Journal of Climate, 17, 1550-1568

IPCC (2013). What is a GCM? Intergovernmental Panel on Climate Change, Data Distribution Centre. Article accessed May 2018

Lindsey R (2000). Antarctic Sea Ice. NASA Earth Observatory.

Morales Maqueda MA et al (2018). Physics and climatology of sea ice. ResearchGate.

Murphy-Bates S (2017). Scientists uncover Earth’s largest volcanic region two kilometres below Antarctic ice sheet, Daily Mail (UK).

Pappas S (2014). Hidden Volcanoes Melt Antarctic Glaciers from Below, LiveScience.

Pinson J (2017). The Arctic is Warming Twice as Fast as the Rest of the Planet – Here’s Why You Need to Care, One Green Planet.

Press Association (2015). Antarctic ice is melting so fast the whole continent may be at risk by 2100, The Guardian. 

Reynolds RW et al (2002). An improved in situ and satellite SST analysis for climate. Journal of Climate, 15, 1609-1625. Data accessed in Nov 2017.

Rignot E. (2008). Changes in West Antarctic ice stream dynamics observed with ALOS PALSAR data. Geophysical Research Letters, 35, L12505, doi:10.1029/2008GL033365

Steig EJ et al (2009). Warming of the Antarctic ice-sheet surface since the 1957 International Geophysical Year, Nature, 457, 459-462, doi:10.1038/nature07669

UCAR (2011). Climate Modelling. University Corporation for Scientific Research, Center for Science Education. Article accessed May 2018.

Van Wyk de Vries M (2017). A new volcanic province: an inventory of subglacial volcanoes in West Antarctica, School of Geosciences, University of Edinburgh. Accessed Feb 2018.

9.5 out of 10 based on 91 ratings