CHRONOLOGY OF THE ICE AGES
WILLIAM F. TANNER*

From: JASA 17 (December 1965): 112-116,

The Pleistocene ice age in which we now live has been characterized by fluctuating continental ice sheets. Two of these are still in existence: those on Antarctica and Greenland. Two have vanished temporarily: those on North America and Europe. The existence of a fifth sheet (Asia) is still uncertain. The sizes of these glaciers can be visualized in terms of their effects on the oceans: if the two remaining ice caps were to melt, sea level would rise by about 400 feet; if the two former ones were restored, sea level would drop about 450 feet. At least two, and perhaps as many as five, major glaciations are known from widely-spaced parts of geological time; the most recent one (the Pleistocene) has had four important ice advances and retreats.

A simple geological model can be used to account for the beginning of the current ice age, and a combination geological-meteorological model can be used to explain the fluctuations. These models permit the following predictions: (1) Although only about half of the maximum ice has melted, no additional warming is in sight (that is, we have reached what is essentially the most melting and the highest sea level possible in the present cycle); and (2) The over-all glaciation, although momentarily is an "interglacial phase," has barely gotten started.

Radioactivity dates show that conditions have been shifting, in the general direction of glaciation, throughout much of Cenozoic time. Carbon-14 dates provide a fairly detailed history of the most recent sycle '(the Wisconsin) within Pleistocene time. Deglaciation (and sea level rise) began roughly 20,000 years ago, and this change halted about 5,000 years ago. Deglaciation was, however, sporadic, and modem sea level has fluctuated both above and below its present position.

North America and Europe are still "bouncing back" - at measurable rates - as a result of the removal of the load of ice. The ice on North America alone was enough to provide each presently living person on the earth with an individual supply of approximately 25 billion pounds. This quantity of ice provides a good measure of the storage of water necessary, on the continents, during times of lowered sea level.

Although there are extremes in both directions, a conservative summary of the glacial history of the earth must contain these two statements:

1. North America has undergone, probably, three major glaciations, at intervals of approximately one billion years (on the radioactivity calendar).

2. Other continents have had quite different glacial histories.

Among the experts, one can find those who hold that only the Pleistocene glaciation (last million years) was real, and that all the earlier ones arise because of our inability to interpret the rocks correctly. On the other hand, there are those who profess to find glacial debris from all eras of history, suggesting that glaciation, on a quickly-recurring basis, is a normal event for our earth. The present writer takes, along with perhaps a large majority of geologists, a middle position. This position-with a few glacial age~ at widely-scattered times--forces us to reject any idea that the Pleistocene ice age was unique, the result of long-term cooling of the earth into a perpetually frozen state such as that of the planet Pluto.

The first widely-accepted glaciation for North America took place about two billion years ago, during the accumulation of Keewatin and Timiskaming rocks, in the Great Lakes area-approximately the locale of the Pleistocene ice age in which we now live. The second widely-accepted glaciation in North America occurred about one billion years ago, during the deposition of Huronian rocks, also in the Great Lakes area. The third widely-accepted glaciation in North America is the Pleistocene ice age.⁹

A number of geologists have reported glacial deposits of late Paleozoic age in North America. The best examples of these may be found in rocks in the Ouachita Mountains, not far from the Oklahoma-Arkansas line. The present writer has examined these supposed tillites, and does not see how they can be assigned a glacial origin. Furthermore, much other available evidence indicates that the climate, at that time, in what is now the Southwestern U.S., was tropical humid.

Glacial debris is, however, known from late Paleozoic rocks of many other continents, specifically South America, Africa, Asia (India), Australia, and Antarctica.9 We conclude, therefore, that North America has undergone more-or-less cyclic glaciation (about once every billion years), but that other continents may have had different histories. This should not surprise anyone: Africa had no continental ice sheet during the Pleistocene, when North America, Greenland, Europe, and Antarctica were largely covered with ice.

*William F. Tanner is in the Geology Department of Florida State University, Tallahassee, Florida. Paper prepared for the 19th Annual Convention of the American Scientific Affiliation, August 1964, at John Brown University, Siloam Springs, Arkansas.

North America and Asia (excluding India), appear to have had similar glacial histories: both are now close to the North Pole, and both lie on the margin of the Pacific Ocean. South America, Africa, Australia, India, and Antarctica have had similar histories; they are all clustered about the South Pole. Europe's position isn't precisely clear, but its glacial history, over the long term, may have been like that of North America.

Throughout Cenozoic time, the North American climate has been getting colder. The total drop has been abqut 20' F, over a period of perhaps 60 or 70 million years.2 This is probably due to a migration of the poles; or, if you prefer to think of it this way, a migration of North America toward the North Pole.8 This rather steady temperature drop does not support the popular concept that some important event took place, roughly one million years ago, which brought on the ice age. Rather, we have a history of a slow, long-term change from the sub-tropical conditions of Mesozoic time, through the temperate conditions of most Cenozoic time, into the glacial conditions of the Pleistocene.

Much geological thinking about the great ice age has been colored by the assumption that the advent of glaciation was sudden and unheralded. Hence we have, in the literature, a continuing search for a "key event," such as a major volcanic eruption, designed to put many cubic miles of volcanic dust into the atmosphere, thereby chilling the entire globe. The field evidence, however, supports the notion of a long, slow change, the culmination of which has been the development of four (or perhaps five) major ice areas.

The first of these (Antarctic) probably began to grow in middle Tertiary time, perhaps 30 or 35 million years ago. When it reached full size is not known, but once it did, it probably did not melt again, as did the North American ice sheet. The earliest appearance of a large ice cap in Canada has not been dated. Mountain glaciation, in the western part of the United States, increased more-or-less systematically throughout the Cenozoic, complicated by changes of elevation, in local areas, due to tectonic uplift here and there, or by reduction of mountains by erosion. There is at least a suggestion that a late Tertiary ice sheet existed in Canada, but it could not have been anything like as extensive as later continental glaciers in the same area.

The Pliocene-Pleistocene boundary is commonly drawn at that time when the first of four large sequential ice caps in North America began to expand to continental proportions. As has been stated above, this was essentially the culmination of a long history of gradual chilling. The precise date for this important event is unknown, and, in view of the nature of the growth of a glacier, probably not recoverable.

Traditionally, the Pleistocene has been allotted approximately 106 years and a few geologists have assigned to it two million years. More recently, geologists analyzing deep sea cores have observed that late Pleistocene rates of sedimentation were so fast that the entire epoch might not have lasted more than about 300,000 years.4 And, even more recently, development of a new radioactivity dating technique7 has permitted a reworking of Pleistocene history, with the conclusion that the original date is more nearly correct. It should be kept in mind, here, that we actually do not have suitable absolute dates for the Pleistocene: what we have is a set of inferences, which can, of course, be modified somewhat to give variable results.

Osmond's success in dating a Florida reef limestone has provided the basis for a reassessment of ice age history. He obtained a date of 130,000 years for a shallow-water coral reef remnant now about 8 meters above sea level, in the Florida Keys. The question is, What part of Pleistocene time does this date represent? At first glance, it might seem that almost any moment within the Pleistocene could be represented; on reconsideration, however, it becomes highly probable that this date identifies a single, brief part of the total ice age history. To make this clear, we must examine the Pleistocene in considerable detail.

In North America, the great ice age is typically separated into seven divisions: four of these are 11glacials" and three are "interglacials." Geologists have given specific names to all seven; however, we can identify them just as accurately by using numbers. That is, we refer to the first glacial, the first interglacial, the second glacial, etc., in terms of the passage of time. Any geologist should be able to understand these terms, even tho4h he doesn't ordinarily use them. Europe has had a similar Pleistocene history.

For approximately 5,000 years, climactic conditions have been almost stable. Although temperature and precipitation have fluctuated somewhat, the quantity of ice locked up in the world's glaciers has remained pretty much the same, and therefore sea level has held almost Still.4 There does not seem to be any evidence that the planet will warm up appreciably in the near future. In fact, most geologists have held either (a) That the Pleistocene is completely over, or (b) That we live in a fourth interglacial, with a fifth glacial still ahead of us. The former position does not seem tenable, in view of the fact that we still have two major ice caps (Antarctica; Greenland), Therefore we adopt the second position, committing ourselves to a prediction that the present interglacial (fourh interglacial) will be followed by a fifth glacial.

The observation that the ice age isn't really over (we still have two ice sheets, which show little or no sign of melting) leads to the suggestion that these two major ice caps are permanent features of the Pleistocene landscape. That is, Pleistocene history has three different aspects:

1. In non-glaciated areas, such as north Africa and the Amazon basin.

2. In permanently-glaciated areas, such as Antarctica and Greenland.

3. In areas where glaciers come and go, such as Canada and northern Europe.

From this analysis, we can move readily into an examination of the behavior of Pleistocene sea levels. Obviously, sea level will stand low when glaciers are well developed, and will stand high when glaciers are missing. Pre-Pleistocene M.S.L. was roughly 100 meters higher than at present. If the Antarctic and Greenland ice caps are "permanent" sea level has not returned to its "original" position during any of the interglacials. Actually, there does not seem to be any field evidence of such a high M.S.L. since the Pleistocene began.

Minimum sea levels, during the ice age, have been close to 130 meters below the present position. That is, the total fluctuation has been about 230 meters, but, since the Pleistocene started, the fluctuation of M.S.L. has been limited to about the lower 130 meters of this range. Present sea level is close to normal interglacial sea level. Hence we can draw the conclusion that the Florida coral (dated as being 130,000 years old), mentioned earlier, represents a former interglacial. The remaining question is, Which of the three available interglacials is represented? The most likely answer is, The latest (i.e., the third).

It is highly probable that corals have flourished in South Florida during each interglacial, when coastal waters were relatively warm, as they are now. However, high-level coral ridges which were built during the first interglacial would be severely eroded during the second glacial, when sea level was lowered and hill-side slopes were therefore steep. And during the second interglacial, remnants of these old coral ridges would be subjected to wave attack as well as to encrustration and burial by later growths. Modern (i.e., fourth interglacial) corals cover perhaps 100 to 1,000 times as much area as do all other interglacial corals. If we extend this ratio backwards into time, and use the conservative end of the range, we can postulate the following:

Modem corals 98.9899%
Third interglacial 1.0000%
Second interglacial 0.0100%
First interglacial 0.0001% 

which is also an estimate of the probabilities. We therefore adopt the tentative conclusion that the 130,000 year figure represents the third interglacial. On this basis, the entire Pleistocene has continued for a period of time which falls somewhere between a minimum of about 600,000 and a maximum of 900,000 years. More detailed theoretical work, using a method which cannot be presented in the brief time available here, indicates that the longer end of the range may be more likely than the shorter. These results are in rough agreement with a chronology worked out from North Atlantic sea bottom cores.³

Regardless of our success in dating the Pleistocene, we have enough information from previous periods to know that North America has been undergoing systematic cooling for perhaps the last 100 million years. The ice cap in the vicinity of the South Pole was probably initiated some tens of millions of years ago. The Pleistocene, then, is the last half-million, or last million, years of a much longer history of dropping temperatures. On the basis of long-term trends alone, and suppressing later fluctuations which do not appear to have altered these trends, we can venture the opinion that even colder weather lies ahead of us.

The fourth, and most recent, glaciation, was the Wisconsin. The fourth, and current, interglacial, is commonly designated as the Recent. Using C14 dates which are now available, it is possible to construct a fairly detailed history of late Wisconsin and Recent times. Glaciation was widespread (in the Wisconsin) up to about 20,000 years ago. Shortly thereafter, melting began. -This continued, with a single important interruption, for about 15,000 years. The interruption, characterized by an ice advance in the Great Lakes area, occurred about 10,000 years ago. Melting was essentially over roughly 5,000 years ago, and sea level was within a meter or two of its present position. The great American and European ice caps had vanished completely.⁴

The interval from about 20,000 to about 5,000 years ago was the time of transition from the Wisconsin to the Recent. If we choose to draw the WisconsinRecent boundary at the time when the last part of the continental glacier melted, then this transition must be placed entirely in the Wisconsin glacial. Many older geologists would, however, put the transition entirely in the Recent.

The transition interval was a time of rising sea level, on a world-wide basis. During this time, M.S.L. went up by about 130 meters, or at an average rate of close to one centimeter per year. There were, however, short, sharp, fluctuations, some of which may have achieved rates as high as 10, or 100, cm per year. One of these fluctuations, affecting all of the coastal low-lands of the world, coupled with an unusually wet, stormy season and perhaps violent coastal erosion, is the best which the geologist can produce in an effort to match the Noachian deluge.

In the United States, Wisconsin time was characterized by relatively low temperatures and high precipitation. As a rule of thumb, we can estimate that Wisconsin temperatures were about 10o F colder than those we know today⁵ and rainfall was about twoto-four times as great. Southwestern states such as Texas were well vegetated (more rainfall), Florida was much bigger (lower sea level), and northeastern states were uninhabitable (covered with ice). The big climatic change, into the Recent, has been a tendency toward higher temperatures and reduced precipitation. We cannot, however, peg temperature and rainfall directly to the amount of ice. That is, we cannot say: Glacial, and hence wetter; or, conversely, Interglacial, and hence drier. This is because the feedback between glacier and climate is so complex that it would be quite possible to have either an increase or a decrease in precipitation correlated with any particular temperature change which might develop. Trying to understand the ice age snow budget strictly on the basis of a single parameter such as rainfall would be much more difficult than trying to understand your bank balance strictly on the basis of a single parameter such as deposits. Withdrawals might also affect the balance; you could, conceivably, either increase or decrease withdrawals, and yet get either an increase or a decrease in the balance. Despite this uncertainty, we can say, in general, that Recent time has been warmer and drier than was the Wisconsin, and that any future glaciation (in the present sequence) will be accompanied by lower temperatures and greater rainfall.

The chronology upon which this section is based has been obtained from C14 dates. Up until fairly recently, no radioactivity dates at all had been available for Pleistocene time. The Wisconsin chronology which was widely published a couple of decades ago, was based largely on varve counts. The varve chronology, as published, is quite different from the C14 calendar.

Varves are thin couplets of sedimentary materials, deposited in meltwater or pro-glacial lakes. Each couplet consists of a coarse layer (silt) and a fine layer (clay). The coarse layer grades gradually upward into the finer layer; this couplet is separated from adjacent pairs by sharpIy-defined lines. In general, each set of two layers represents one year of deposition. The coarse layer is a summer deposit, when meltwater carries silt and perhaps even sand. The fine layer is a winter deposit, when clay particles settle out of suspension. Individual varves have thicknesses which, commonly, can be recognized in distinct, but near-by, lake deposits. That is, the varves in one lake can be matched, in many instances, with those in another lake roughly one kilometer away. If a sequence of varves can be identified in the sediments of one lake, and then matched in nearby lake deposits, one of the two sites probably extends into, more ancient times, and hence the combined sequence may strefich over more than the span found in either one. Correlation with a third site may very well extend the sequence even more, toward either the present or the past.⁵

By making varve counts at geographic intervals of about one km, DeGeer was able to work out a fairly accurate chronology of late-glacial events in Sweden. Deglaciation, in that area, covered about 8,000 years, in the first study; later study added 8,700 years of post-glacial history, but did not bring the sequence down to historical times. The total period covered was 16,700 years, somewhat less than the 20,000 available (according to, the C14 scale). Furthermore, the sequence was not continuous; there were several gaps, as DeGeer and his students moved from place to, place, where they were unable to correlate varves, and had to substitute estimates based on their general understanding of Pleistocene history.⁵ (p. 394)

Antevs has been the leading proponent of varvecounting in North America. His early studies led him to the conclusion that the time elapsed since the ice sheet disappeared "can not even be estimated" with any degree of reliability. Later he estimated 28,000 years for the retreat of ice from Long Island to an important point in Ontario, and 12,000 additional years since then, making 40,000 years since the Wisconsin ice sheet began to melt. This is the chronology that many American geologists had accepted before the advent of C14 dating. Unfortunately, only 1R,000 couplets were actually studied in the field; the other 21,000 16couplets" were interpolated in three big gaps, 80, 180, and 290 km wide. There was also at least one smaller gap.5 (p. 396) This means that the 19,000 counts were made in 5 different (and uncorrelated) areas; hence the total elapsed time might be as small as 4,000 years.

In addition, there are two other complicating factors: there was no absolute date to tie the varve sequence to, and there is ample evidence that (during the stormy years), more than a single couplet may accumulate in one 12-month period. In other words, Antevs' total apparently needed reducing below the 19,000 actual count, and the reduced figure needed to be "anchored" to some known event. Even without any reduction, however, it should be noted that 19,000 is not as much as the 20,000 provided from C14 dates.

On the other hand, there is strong evidence to support the observation that ice appears to have been present over the site of Boston about 13,000 years ago.⁶

In general, Wisconsin time is considered to have included two major glacial episodes, with a brief warming trend between. C14 dates provide details for the melting of the second of the Wisconsin glaciers. According to . the C14 calendar, this melting occurred largely between 20, and 10 thousand years ago, for North America, with some other melting, perhaps elsewhere, accounting for a slow rise of M.S.L. until about 5,000 years ago. The varve record can be made to conflict with this, but does not necessarily do so.

North America appears to have undergone approximately three major glaciations, at intervals of about one billion years. This is not the glacial history of some other continents, such as Africa, Australia; and South America, which have had important ice caps at other times. In general, this pattern of on-again, off-again glaciation, ever since there was enough water at the surface to form extensive ice sheets, does not support the theory of a gradually cooling earth. Instead, wide-spread ice cover seems to be due to the operation of some more-or-less regular (but perhaps complicated) mechanism on the earth itself. The ability of this mechanism, which must be very delicately adjusted, to produce successive ice ages over an interval of several billion years, indicates that the genm eral temperature of the air has not fluctuated greatly during that time.

The most recent ice age is the Pleistocene. It is the culmination of a cooling trend, particularly in North America, which began (perhaps) in the Cretaceous period of the Mesozoic era, roughly 100 million years ago. This cooling produced mountain glaciers in the Rocky Mountains at various times since then, and may have been responsible for a succession of small continental ice sheets, in Canada, of which we no longer have any record. The Pleistocene started not more than 2 million years ago, nor less than about 300,000 years ago. A "best guess" is somewhere between 600,000 and 900,000 years ago. Earlier glaciations were pre-Pleistocene. As North America cooled, and glaciers began to form, sea level dropped. By middle Tertiary time, some 30 million years ago, an Antarctic ice cap had been formed. It is unlikely that sea level has returned fully to "normal" at any time since.

If we count the present as the fourth interglacial, the Pleistocene can be subdivided into four glacials and four interglacials. This statement presupposes that a fifth glacial lies ahead of us. All of the systematic (Le., not ad hoc) hypotheses for the cause of the ice age require additional glaciations yet to come.

The present (Le., fourth) interglacial is representative of the previous three, as far as temperature, rainfall, and sea level are concerned. It is unlikely that sea level has stood much higher than it does now, at any time during the Pleistocene. Although it may rise a few meters more, practically all of the evidence suggests that the next major change will be a drop, some time in the next few thousands or tens of thousands of years.

The fourth, or Wisconsin, glacial, was terminated by a melting which lasted, roughly, from 20 to about 5 thousand years ago. Noah's flood was, apparently, one event in that history of deglaciation. For the most recent 5,000 years, world wide temperatures have fluctuated only modestly, and sea level has remained within a meter or two of one position. All of well-known human history has taken place since melting reached a minimum and sea level has become fairly well stabilized. Melting today is on the order of some tens of cubic miles per year, which is negligible when spread over the surface of the ocean. Up until the last few years, at least, the climatic trend has been one of almost imperceptible warming; this can be expected to continue, with minor reversals, for some time yet.

1. Arnold, J. R., Nuclear Geology (edited by Henry Faul), John Wiley and Sons, New York, 349-354, 1954.

2. Clark, Thomas H., and C. W. Stearn, The Geological Evolution of North America, Ronald Press, New York, 1960, p. 273.

3. Ericson, David B., Maurice Ewing, Goesta Wollin, and Bruce C. Heezen, "Atlantic deep-sea sediment cores," Bulletin, the Geological Society of America, 72, 193-286, 1961.

4. Fairbridge, Rhodes, Physics and Chemistry of the Earth IV (edited by L. H. Ahrens, Frank Press, K. Rankama, and S. K. Runcorn), Pergamon Press, New York, 99-185, 1961.

5. Flint, Richard F., Glacial Geology and the Pleistocene Epoch, John Wiley and Sons, New York, 1947.

6. Kaye, Clifford A., and Elso Barghoorn, "Late Quaternary sea-level change and crustal rise at Boston, Massachusetts, with notes on the autocompaction of peat," Bulletin, Geological Society of America, 75, 63-80, 1964.

9. Schwarzbach, Martin, Climates of the Past, D. Van Nostrand Co., Princeton, N.J., 1963.

10. Tanner, W. F., "Geology and the Great Flood," Journal, American Scientific Affiliation, 13, 117-119, 1961.