CMIP5 Ice cover

Will the Arctic and Antarctic sea ice melt completely by the end of this century? I show below what one CMIP5 model – the Centre National Recereche Metorological CNRM-CM5 is predicting. The images are generated from the sea ice field output by the model between 1861 and 2100.


Plotted here are the Antarctica sea ice levels for Sep between 1861 and 2099. This is the maximum month for sea ice. What is not shown is the Antarctica ice sheets on land. Only if these melt will it make a difference to sea levels. Freezing and melting oceans with the seasons do not change sea level.

Arctic-winterThis is how the  Arctic winter ice looks during the same 240 year period.

The summer ice in the Arctic is predicted to dissappear by 2040-2060. A minimal amount still remains in the Antarctic summer in 2060 but that too dissapears by 2100.




Whether summer ice in the Arctic will completely disappear is something that we will surely know within the next ten years.

Not all models agree on the rate of ice loss either. I show below a comparison of CSIRO and GISS models for a 12 monthly running average of total ice cover. For some reason they don’t even agree about the cover in 1860. Perhaps there is a problem of definition.

Comparison of ocean ice cover from 1861 to 2100. The red curve is a running 12 month average from GISS-E2-H, while the black curve is CSIRO-MK3

Comparison of ocean ice cover from 1861 to 2100. The red curve is a running 12 month average from GISS-E2-H, while the black curve is CSIRO-MK3

This entry was posted in AGW, Climate Change, climate science, GCM, Institiutions, IPCC, Science and tagged , . Bookmark the permalink.

6 Responses to CMIP5 Ice cover

  1. Euan Mearns says:

    Clive, what is the source of the data 1861 to 1979? If its true must surely show that the process of ice retreat is natural and linked to the ending of the LIA. I also wonder how unusual it is to have an ice free Arctic? Do you have any information on that? I once came across a paper that implied the Arctic was ice free (summer?) during the LIA. This could help explain the Bond drift ice data – ice bergs don’t get trapped in Arctic sea ice. And it also has profound implications for salinity distributions in the northern high latitudes.

    • Clive Best says:

      The source is the French climate model which runs from 1861 to 2100 based on RCP8.5. I have now got the output from 31 climate models and am writing some software to interpret it. This was just an exercise in visualising the model ice cover. As far as I know the ice data should follow the historical data up until 2000. I am not making any comment about whether to believe the models or not. I am just trying to display the reults I think if there is still summer ice in 20 years time then the models will have been proved wrong.

      It is also true that the ice cover was far greater in 1861 than today and much of that change prior to 1970 must have been natural.

  2. edmh says:

    Looking at a broader picture in Century or Millennium steps, according to the GISP2 Ice Core data the real decline towards the next glacial age started some 3000 years ago, round about 1000BC.

    The GISP2 temperature record shows a distinct ‘Tipping Point’ at ~1000BC and temperatures then start their decline at a significantly increased rate.

    Having been roughly flat for the first 7000 years including the Holocene Climate Optimum, the Holocene rate of temperature decline escalates from roughly 0.05°C / millennium 8000BC – 1000BC, to about 0.5 °C/ millennium, 1000BC – 2000AD.

    The GISP2 ice core records from Greenland show.

    1 the last millennium of our benign Holocene 1000AD – 2000AD was the coldest of the whole current Holocene interglacial.

    2 each of the notable high points in Holocene temperatures, (Holocene: Climate Optimum – Minoan – Roman – Medieval – Modern), has been progressively colder than the previous high point.

    3 for its first 7-8000 years the early Holocene, encompassing its high point “climate optimum” had a pretty flat temperatures on average a drop of only ~0.05 °C per millennium.

    4 but the recent Holocene for the last 3000 years since 1000BC has seen a temperature diminution at at least 10 times that earlier rate.

    5 our happy Holocene interglacial is about 10-11000 years old and judging by earlier Interglacials the epoch is probably drawing to its close, in this century the next century or this millennium.

    6 so any minor warming after at the end of the 20th century to the Modern high point eventually will be seen as just a little noise in the system in the longer term progress of continuing cooling over the past 3000+ years.

    7 other published Greenland Ice Core records (NGRIP1, GRIP) corroborate this finding. They also exhibit a similar pattern of a prolonged relatively stable early Holocene period followed by a subsequent much more rapid decline in the more recent past.

    Global warming protagonists should accept that the Holocene is in long term decline and that any action taken by man-kind is unlikely to make any difference whatsoever.

    And were the actions by Man-kind effective at averting warming would be simply reinforcing the catastrophic and eventually disastrous cooling that is bound to return in due course, this century, next century or this millennium


  3. DrO says:

    There are so many profound problems with the climate models, it is hard to know where to begin, and I have listed many of these in the past and also provided a few explanatory notes, such as those at ( With that in mind, I offer just a couple of the many possible concerns with these IPCC et al type “games”.

    1) Omitted from your post is that the Antarctic has been and continues to set record new increases in sea ice extent in each and every year in recent records. Roughly speaking, the losses in the Arctic are about offset or surpassed by the gains in the Antarctic.

    Amongst other things, I note the, perhaps spurious, correlation between the ice extent in the North and South with the motion of the magnetic poles. Just one of the many large forces too poorly understood by man, and certainly not sufficiently quantifiable to even get a handle on how such poorly understood forces/dynamics impact the planet’s climate … and of course, omitted from the “models”.

    2) The IPCC et al have started to move away from “proper physical” (albeit still entirely inadequate) models more towards statistical/time series models, which is much worse still. Crudely put, cherry pick some data, fit a curve to obtain a trend, and guess a sufficiently large variance to be able imply any doom and gloom you wish … lies , damn lies, and statistics.

    Ignoring non-stationarities, non-linearities, and a vast number of crucial forces, and messing with the data at will, allows such models to be set-up to predict anything you like.

    ASIDE: not to mention things like the “administrative adjustments” in the land based temp records, which are self-contradicted by themselves and the satellite records etc etc , but that’s for another day.

    3) If you are going to use such models, and attempt to incorporate the variances over thousands of years to make 40 – 80 year prediction, then your prediction quality may be vanishing, as the “noises” would destroy the resolution required, and of course destroy many periodicities, and aperiodicties that exist and are required in a proper model.

    Finally, as always, the “context” and objectives of the models/predictions must form a coherent basis. It is meaningless to predict +/- 10C over 80 years, if you “believe” that +/-2 is already death for the planet. Your models must be able to predict to a very much finer resolution, and it is known with certainty that is not possible.

    … and of course the biggest issue, as always on this subject, while mathematics proves the utter impossibility of any prediction any where near the resolution required for the IPCC et al “objective” (which are decidedly to do with transferring cash from the rich to the poor), even if a (infinitely difficult) “temperature” prediction were possible … it is then a difficulty of “uncountable infinity” to predict if that change is good, bad, or indifferent for mankind.

    • Clive Best says:


      I agree with many of the points you make above. It is also true that Antarctic sea ice has grown in recent years and reached a mximum this september. No-one expected that or the hiatus in warming. There are still many uncertainties in climate models particularly regarding aerosols, clouds and the water cycle.

      So I am not defending climate models. I am just trying to visualise their results in a simple, easy to understand format. The public have lost interest because the climate doesn’t change on any timescale of direct concern to them and they have more pressing concerns. Sea level may rise by 20cm in one lifetime. Yet that is nothing compared to the diurnal tidal range which can reach up to 10m!

      So our immediate problem is not global warming as such, but the activists who are trying to overturn modern life in western countries to solve an as yet opaque problem. The developing world will continue burning fossil fuels so long as they are cheapest. Eventually we will discover whether AGW is a severe problem or not. It would have been better to invest the trillion dollars already spent on non-hydro renewable energy into nuclear fusion. If we had done that then we would have a long term solution for the world, avoiding nuclear proliferation problems.

  4. Ron Graf says:

    It would have been better to invest the trillion dollars already spent on non-hydro renewable energy into nuclear fusion. If we had done that then we would have a long term solution for the world, avoiding nuclear proliferation problems.

    The best thing the Russians could do for mankind now is announce a crash program to obtain controllable fusion. I can’t imagine the west would let them win the race.

Leave a Reply