Minds in Ablation Part Two: Stability of the Arctic Ice Pack: Theory

Copyright © 1998, 1999 by Sean Mewhinney (df736@freenet.carleton.ca).

Stability of the Arctic Ice Pack: Theory
In Which Ginenthal Discovers the Hypsithermal

Ginenthal's argument is based on one of these internal feedback loops in a part of the climate system that seems to be quite sensitive: the ice covering the Arctic Ocean. Ice reflects most of the solar radiation that falls upon it. If the ice cover were permanently removed, the surface of the Arctic Ocean would be greatly warmed by the heat of the sun, and much of this heat would be transferred to the air above it. The two great continental ice sheets today -- the Greenland and the Antarctic -- are as much as three kilometers thick, and would require a very long time to melt away under a warmer climatic regime. But the Arctic ice pack averages only three meters thick, and is much more vulnerable. In fact, half of it melts every summer, in the marginal areas of the Arctic basin, but the central Arctic is perennially ice-covered. Climatologists have long speculated on the role of the Arctic ice pack in initiating glacial cycles. In the past, Donn and Ewing theorized that an ice-free Arctic Ocean is necessary to initiate the growth of ice sheets in northern Europe and North America, by providing an abundant source of moisture for high-latitude snowfall. 1 Others believe that without an ice-covered Arctic, the formation of such ice sheets would not be possible. The stability of the ice pack has been the subject of many studies.

C. E. P. Brooks believed that the ice pack was balanced on a knife edge -- that a rise of as little as 2 degrees Fahrenheit (1.1° C) in the temperature of the air might be sufficient to bring about its complete dissolution in summer, and that once melted away, it would not reform in winter until atmospheric temperatures fell again. In other words, the Arctic Ocean was bistable between two states, ice-covered and ice-free, and a slight shift in temperature would be enough to nudge it from one state to the other. This belief was not supported by any calculation. It was merely a conjecture. Further, he believed that the Arctic was in fact ice-free during the hypsithermal -- from approximately 8,000 to 3,000 years ago. This is Ginenthal's point of departure. Brooks presented these views in his Climate Through the Ages, first published in 1926. A revised edition appeared in 1949. 2

Ginenthal's next stop is two computer experiments with general circulation models published in 1973. Both of these studies addressed the question, what effect would disappearance of the Arctic ice pack have on atmospheric temperatures. R. L. Newson's experiment predicts mean annual temperature increases "near the model surface" of 10° C in southern Greenland, 20° C in northern Greenland, and even more than that -- 30° C -- over Hudson Bay. 3 Warshaw and Rapp's results show mean annual warming, at the 1,000-millibar level (near sea level), averaged for each parallel of latitude, ranging from 5° C at about 62° N to over 15° C north of latitude 77° N. (Greenland ranges from about 60° to 84° N.) 4

Ginenthal argues that such warming would melt away the Greenland ice sheet in a comparatively short time, and he goes on to cite various botanical and faunal evidence at great length, some of it from Greenland, intended to emphasize how warm the hypsithermal was. But before moving on to the next links in Ginenthal's argument, I want to have a brief look at more recent theoretical work on the stability of the Arctic ice pack. Note that there are already three separate questions, which could be stated as follows:

  1. How much must air temperature over the Arctic Ocean rise to melt the pack ice in summer?
  2. How much additional temperature increase, if any, would be required to prevent it from freezing up again in the winter? Or more simply, would it refreeze?
  3. If the ice did melt, and not reform, how much further would the temperature of the air rise in the Arctic?

Brooks is hardly the last word on questions one and two. Researchers have returned to these questions again and again, using more and more sophisticated tools of analysis, with a growing body of observation to improve the accuracy of values used for the parameters in their calculations, and they have come up with a variety of answers.

Some used simple energy budget calculations, with mean annual values, or adjusted for the seasons. Others used computer models of varying degrees of sophistication. There are three coupled systems in nature to model: the ocean, ice, and atmosphere, with inputs and outputs between all three. Some tradeoffs are necessary. Even with the fastest supercomputers available at the time, many simplifying assumptions must be made to save computing time, and even the most sophisticated of these models are rather primitive representations of reality. These studies do not all address the same questions.

In the sixties, Donn and Shaw calculated the mean annual energy budget for the Arctic in its present ice-covered state and in a hypothetical ice-free state. They concluded: "The results are tentative but seem to suggest, on the basis of present knowledge, that an ice-free polar surface could be maintained if the present ice cover were removed..." 5

Using a simple two-season energy budget, Budyko concluded that an increase of 4° C in summertime air temperatures, if maintained for four years, would be enough to melt the thickest Arctic sea ice. A smaller increase, on the order of 2° C, if maintained long enough, would eventually destroy the ice. Once ice-free, the sea would not refreeze in winter unless temperatures fell again. But "it is plausible that over a sufficiently long period thermal anomalies can combine in such a way as to lead to either the formation or destruction of old ice." 6 In 1966, Budyko was confident that "probable errors in determining the principal parameters of calculation do not greatly influence our results." 7 But when he presented the same calculation in 1974, he was more cautious:

The results... corroborate the possibility of an ice-free regime in the Central Arctic, because the water temperature in winter turns out to be higher than the freezing point of salt water, approximately -1.8° C.... the results... suggest a great instability of the ice-free regime, as the water temperature in winter exceeds the freezing point by only 1°.... one should bear in mind that the probable error of even the most accurately calculated temperature cannot be less than several degrees. 8

From his own calculations, in 1968 Fletcher concluded: "...the heat budget for an ice-free Arctic sea would be slightly positive; that is to say, the pack ice, once removed, might not return." 9

A study of convection in the Arctic Ocean carried out by Foster does not directly address the question of whether the ice cover is stable, but is relevant: Foster found that heat absorbed from the sun by open water remains in a shallow layer much closer to the surface than previously thought. Most of this heat is lost in melting ice in the summer, and is not available to prevent freeze-up when winter returns. 10

Parkinson and Kellogg used a computer model with four vertical layers on a coarse geographic grid, representing ocean, ice, snow, and air. In their first experiment, they found that a temperature increase of 5° C caused the ice to disappear completely in August of the third year, but it still refroze in October. Then they stepped up the temperature increase to 6° C in the summer months and as much as 9° C in the winter. Ice still returned in November and persisted until August. 11

Stigebrandt's Arctic dynamical sea-ice model incorporates terms for river runoff, ice drift out of the Arctic basin, and vertical mixing. It predicts that if mean ice thickness falls below about 1.5 meters, the ice pack will become unstable, and may disappear. 12 "For... contemporary climatic conditions the model predicts that a perennial ice cover on the Arctic Ocean would not reform once disappeared." 13

Hunt set out to test the hypothesis that land at high latitudes is necessary to initiate ice ages. His 1984 study couples a general circulation model of the atmosphere to a sea-ice model on a polar ocean with no land above 70° N. He allowed an ice pack to form under the present mean annual conditions of vapor, clouds, etc. averaged for each latitude. The albedo of the surface was then changed from that of ice to that of open ocean, allowing the temperature to rise. "The largest polar warming, 4.5 K ... was insufficient to raise the temperature above the freezing point of seawater." 14 Hunt concluded that "an initial ice-free Arctic Ocean would be expected to refreeze."15

Semtner used a three-layer sea-ice model coupled to a 13-layer ocean model, superimposed on a geographic grid representing the distribution of land and sea, as well as ocean depths. When temperatures are increased by a uniform 2° C throughout the year, "The result is a dramatic disappearance of the sea ice in late summer..." However, "no catastrophic disappearance of ice throughout the year occurs.... the wintertime extent of sea ice is only slightly reduced by increased atmospheric temperature..." 16 Because the model predicts ice thicknesses under present temperatures as little as half what is observed, Semtner felt that the results of this perturbation experiment should be interpreted as "more representative of, say, a 4° C change..." 17

This brings us up to 1987, where our little survey ends. There is a lot of scatter in the answers to the first question, how much warming is needed to melt the ice, but no one who has attempted the calculation agrees with Brooks that as little as 2° F (1.1° C) is enough. The stability of the ice cover does not seem to be as precariously balanced as he thought. As to the second question, stated simply as, would the sea freeze up again the next winter, once open, there are more nos than yeses. But if we look at the later, more sophisticated computer modelling studies, the yeses outnumber the nos. When the problem is analyzed, it looks as if the Arctic ice pack is tougher to get rid of than Brooks thought. It may be able to withstand quite a substantial temperature increase.

How Warm Was the Hypsithermal?

Newson and Warshaw and Rapp have already answered our third question: if the pack ice could be made to disappear, how warm would the atmosphere over the Arctic become? Their answer was, quite a lot. I have no quarrel with these studies, although as Budyko points out, 18 most of this increase comes in the winter. Ablation takes place only in the summer.

But this, too, is a theoretical question. The question, "How warm was the hypsithermal?" is a question of fact. Ginenthal seems to think that Newson and Warshaw and Rapp support Brooks' conjecture that the Arctic was ice-free during the hypsithermal. But they support exactly the opposite conclusion, for the simple reason that there is no evidence of such a large temperature increase in postglacial time. Neither Brooks nor any of Ginenthal's other sources claim such a large increase.

Ginenthal speaks of "the period known as the hipsithermal [sic] ... when the temperature was greater by 4 to 5 F."19 His note 34 cites page 297 of Brooks, where we find this sentence: "This was the beginning of the `Climatic Optimum,' with temperatures up to 5° F. higher than the present... [2.8° C]" Ginenthal correctly notes that any change in global temperatures is amplified in the polar regions. He quotes wholesale passages out of Orr's Between Earth and Space on the effects at temperate and Arctic latitudes of a temperature increase in the first decades of this century, averaging 1 to 2° F (.55 to 1.1° C) in the northern hemisphere (or, according to Borisov, 0.6° C for the earth as a whole). He then asks: "What would happen in Greenland during the hipsithermal [sic] ... under a temperature regime 4 or 5 F hotter than the present?" 20 And again a few pages later: "What would happen to Greenland with a 4 to 5 F rise in Earth temperature for, perhaps, 5,000 years?" 21

But Brooks was speaking of the maximum temperature difference in northern Europe, not a global average. The sentence quoted above comes from Part III of his book, "The Climate of the Historical Past," in chapter XVII, "Europe."

Another example: As an illustration of hypsithermal warming, Ginenthal takes this passage from Borisov: "The vegetative zones advanced toward the pole. On the Eurasian continent this latitudinal shift amounted to 4-5 degrees in the west and to 1-2 degrees in the east." 22 On another page, he quotes Orr to the effect that "A one-degree [Fahrenheit] shift in mean annual temperature is equivalent to roughly a hundred miles of latitude." 23 Let's put these two statements together and see what we get. A hundred miles is about 1.45 degrees of latitude, so an average 4.5-degree latitudinal shift in the west is equivalent to 3.1° F, or 1.7° C. And an average 1.5-degree shift in the east is equivalent to 1° F, or 0.6° C, accepting Orr's equation as valid.

This is in line with more recent estimates. Botanical and faunal data of various kinds from hundreds of sites have been compiled and used to prepare maps of temperatures and precipitation during the hypsithermal. Maps prepared by several different investigators are collected in Frenzel et al., Atlas of Paleoclimates. 24 On page 73, Klimanov's map shows annual mean temperatures higher than those of today for the period 6,000 to 5,500 years ago of 2° C in southern Greenland and 3° C in central and northern Greenland. (See Fig. 1.) In Climate Change: The IPCC Scientific Assessment, 25 a map on page 205 prepared by Borzenkova and Zubakov shows summer temperatures 3° C higher than present in Greenland. (See Fig. 2.) According to the text, "The greatest relative warmth in summer (up to 4° C) was in high latitudes north of 70° N."

Thompson Webb is one of the key figures in a project to map global temperatures and precipitation for the period 6,000 years before present, from pollen, lake-level, and plankton data. He is on record as saying, "current data suggest that the global mean temperature at 6000 B.P. was probably within 1° C of today's temperature." 26

If the Arctic ice pack melted during the hypsithermal, then according to the studies of Budyko, Newson, and Warshaw and Rapp, the earth should have experienced much greater warming. The evidence from land and sea is that this did not happen.

Arctic Ocean Sediment Cores

Since the 1960's several hundred sediment cores have been retrieved from the central Arctic. In the world's deep oceans, sediment rich in biological remains normally accumulates at rates varying from 1.5 to 3 centimeters per thousand years. 27 In contrast to this, the sedimentation rate in the Arctic cores has been only about 1.5 to 2 millimeters per thousand years over the last 700,000 years or so, with the average at the low end of this range for the last 70,000 or 80,000 years. It is very poor in biological material, consisting largely of ice-rafted gravel. Everybody who has studied these cores has agreed that the Arctic has remained ice-covered for at least the last 70 or 80 thousand years. If the sea had been open for several thousand years as Ginenthal claims, and therefore capable of supporting large-scale photosynthesis, it would be quite obvious from a tenfold increase in sedimentation and a dramatic change incomposition in the cores. Hunkins and Kutschale note:

...Sedimentation in the Arctic Ocean has continued unchanged over the last 70,000 years. In most Atlantic cores, a sharp boundary is found marking the end of the last glaciation about 11,000 years ago. This boundary is not evident in the Arctic Ocean sediments. Presumably any marked change in the ice cover would have changed the sedimentation regime. This implies that the ice cover has existed relatively unchanged through the last glaciation and through post-glacial time to the present. 28

The question, "Was the Arctic ice-free during the hypsithermal?" is answered by the sediment cores: It was not.


1. William L. Donn and Maurice Ewing, "The Theory of an Ice-Free Arctic Ocean," Proceedings of the Seventh Congress of the International Union for Quaternary Research, Vol 5, equals Meteorological Monographs Vol. 8, no. 30 (Feb., 1968), pp. 100-105.

2. C. E. P. Brooks, Climate Through the Ages: A Study of the Climatic Factors and their Variations (N.Y., Dover, 1970 repr. of 1949 ed.).

3. R. L. Newson, "Response of a General Circulation Model of the Atmosphere to Removal of the Arctic Ice-Cap," Nature Vol. 241 (Jan. 5, 1973), pp. 39-40.

4. M. Warshaw and R. R. Rapp, "An Experiment on the Sensitivity of a Global Circulation Model," Journal of Applied Meteorology Vol. 12 (Feb., 1973), pp. 43-49.

5. William L. Donn and David M. Shaw, "The Heat Budgets of an Ice-Free and an Ice-Covered Arctic Ocean," Journal of Geophysical Research Vol. 71, no. 4 (Feb. 15, 1966), p. 1092. See also Donn and Shaw, "The Maintenance of an Ice-Free Arctic Ocean," Progress in Oceanography Vol. 4 (1967), pp. 105-113.

6. M. I. Budyko, "Polar Ice and Climate," Proceedings of the Symposium on the Arctic Heat Budget and Atmospheric Circulation (1966), p.19.

7. Budyko, op. cit., p. 16.

8. M. I. Budyko, Climate and Life (Academic Press, 1974), p. 285.

9. J. O. Fletcher, "The Influence of the Arctic Pack Ice on Climate," Proceedings of the Seventh Congress of the International Union for Quaternary Research, Vol. 5, equals Meteorlogical Monographs Vol. 8, no. 30 (Feb., 1968), p. 99.

10. Theodore D. Foster, "Heat Exchange in the Upper Arctic Ocean," AIDJEX (Arctic Ice Dynamics Joint Experiment) Bulletin no. 28 (March, 1975), pp. 151-166.

11. Claire L. Parkinson and William W. Kellogg, "Arctic Sea Ice Decay Simulated for a CO2-Induced Temperature Rise," Climatic Change Vol. 2 (1979), pp. 149-162.

12. Anders Stigebrandt, "A Model for the Thickness and Salinity of the Upper Layer in the Arctic Ocean and the Relationship between the Ice Thickness and Some External Parameters," Journal of Physical Oceanography Vol. 11 (Oct., 1981), pp. 1407 ff.

13. Stigebrandt, "On the Hydrographic and Ice Conditions in the Northern North Atlantic During Different Phases of a Glaciation Cycle," Palaeogeography, Palaeoclimatology, Palaeoecology Vol. 50 (1985), pp. 312-313.

14. B. G. Hunt, "Polar Glaciation and the Genesis of Ice Ages," Nature Vol. 308 (March 1, 1984), p. 49.

15. Hunt, op. cit., p. 50.

16. Albert J. Semtner, Jr., "A Numerical Study of Sea Ice and Ocean Circulation in the Arctic," Journal of Physical Oceanography Vol. 17 (August, 1987) pp. 1095-1097.

17. Semtner, op. cit., p. 1095.

18. "Thus, we can conclude that the polar ice decreases the temperature in the central Arctic in winter by 30° to 35° C, and in summer by 5° C." Budyko, "Polar Ice and Climate," p. 16.

19. Ginenthal, near the beginning of Part VI.

20. Ginenthal, Part VI.

21. Ginenthal, still in Part VI.

22. Ginenthal's note 35, citing Borisov, p. 36.

23. Ginenthal's note 51, citing Orr, pp. 160-161 of the 1961 ed.

24. B. Frenzel et al., eds., Atlas of Paleoclimates and Paleoenvironments of the Northern Hemisphere: Late Pleistocene-Holocene (Budapest & Stuttgart, 1992).

25. J. T. Houghton et al., eds., Climate Change: The IPCC Scientific Assessment (Cambridge Univ. Press, 1990).

26. T. Webb III and T.M.L. Wigley, "What Past Climates Can Indicate about a Warmer World," in M. C. MacCracken & F. M. Luther, eds., Projecting the Climatic Effects of Increasing Carbon Dioxide, U. S. Dept. of Energy Report ER-0237 (1985), p. 251.

27. G. Wollin, D. B. Ericson, and M. Ewing, "Late Pleistocene Climates Recorded in Atlantic and Pacific Deep-Sea Sediments," in Karl K. Turekian, ed., The Late Cenozoic Glacial Ages (Yale Univ. Press, 1971), p. 199.

28. Kenneth Hunkins and Henry Kutschale, "Quaternary Sedimentation in the Arctic Ocean," Progress in Oceanography Vol. 4 (1967), p. 94. See also Kenneth Hunkins, Allan W. H. Bé, Neil D. Opdyke, and Guy Mathieu, "The Late Cenozoic History of the Arctic Ocean," in Turekian, op. cit., pp. 215-237.

Back to Minds in Ablation index page.
Back to my catastrophism page.
Back to my interests.
Back to my home page.
Search my pages.

Last modified by pib on October 28, 2003.