Ice Ages Confirmed

Alan Feuerbacher

Up until about 1960, many pieces of the ice age puzzle were unexplained. Ice Ages: Solving the Mystery195 describes the situation:

Once they had accepted and enlarged upon Agassiz' glacial theory, geologists faced the challenge of explaining the ice ages. What agent stimulated the ice sheets to grow and expand? Why, having spread out to cover nearly one-third of the earth's land area, did those ice sheets then retreat? Most intriguing of all: would they return? These were the central questions of the ice-age mystery.

Many theories were advanced. Some that seemed at first to provide plausible answers were later rejected when fresh evidence proved them wrong. Others that were untestable had to be laid aside -- judged according the the Scottish verdict, "Not proven."

Several attempts to solve the ice-age mystery ran into difficulty because they focused too narrowly on the fluctuations of the ice sheets themselves, failing to see them as only one part of a global climate system -- a system that includes all of the mobile elements of the planet: ice sheets, ocean, and atmosphere. The three elements of this air-sea-ice system are interconnected in such a way that they behave like a huge machine. A change in one part brings about corresponding changes in the other parts of the system....

Any valid theory of the ice ages must take into account that the growth or decay of a large ice sheet would have a large effect on the other elements of the climate system. For example, if an ice sheet is to expand, water must be drawn from the oceans, carried through the atmosphere to the site of the ice sheet, and precipitated there as snow. Variations in the volume of global ice are therefore linked inextricably with variations in sea level. Furthermore, any change in the area of an ice sheet must bring about a change in the radiation balance of the globe. When an ice sheet expands, heat is lost through reflection, global temperatures drop, and more ice is formed. Conversely, when an ice sheet shrinks, temperatures rise, and further shrinkage occurs. This "radiation-feedback effect" plays an important role in several theories of the ice ages because it explains how a small initial change in the size of the ice sheet is amplified.

The main objective of most theories is to discover the cause of this initial change. Since Agassiz' 1837 "Discourse," literally dozens of explanations for the ice age have been suggested.

One theory proved resoundingly successful in explaining the ice ages. About 1911, a Yugoslavian professor of mathematics, Milutin Milankovitch, began work on a theory "capable of describing the climates of earth, Mars, and Venus -- today and in the past." To describe what Milankovitch did, I first must describe the earth's changing relationship with the sun. A good explanation of the changes is paraphrased from Time-Life's Ice Ages, which also contains a helpful diagram:196

The 100,000 year stretch: The orbit of the earth gradually stretches from nearly circular to an elliptical shape and back again in a cycle of approximately 100,000 years. This is called the orbit's eccentricity. During the cycle, the distance between earth and sun varies by as much as 11.35 million miles.

The 41,000 year tilt: The earth's axis is never perpendicular to the plane of its orbit; over the course of about 41,000 years the angle varies between 21.5 and 24.5 degrees. Because of the tilt, the solar radiation striking any point on earth fluctuates during the yearly orbit, producing seasons. When the tilt is greatest, summers are hottest, winters are coldest.

The 22,000 year wobble: Even while the shape of its orbit and the tilt of its axis are changing, the earth wobbles slowly in space, its axis describing a circle once every 22,000 years. Because of this movement, known as axial precession or the precession of the equinoxes, the distance between the earth and the sun in a given season slowly changes. Today, for instance, the shape of the orbit places the planet closest to the sun in the Northern Hemisphere's winter and farthest away in summer. The combination tends to make winters mild and summers cool -- and favors ice-sheet growth. However, 11,000 years ago, the arrangement was just the opposite, setting the stage for the Northern Hemisphere ice sheets to decay.

These variations were calculated mathematically by several workers during the latter half of the 19th century. Milankovitch used the existing calculations of variations in eccentricity, precession and tilt to calculate how much solar radiation strikes the surface of the earth during each season and at each latitude. He published his first results in 1920, which contained a graph showing how summer radiation at latitudes 55Deg, 60Deg, and 65Deg North varied over the past 650,000 years. His next results were published in 1930, and included radiation curves for each of eight latitudes ranging from 5Deg North to 75Deg North. The curves calculated for high latitudes are dominated by the 41,000 year tilt cycle, while those for low latitudes are dominated by the 22,000 year precession cycle. By 1941 he had finished his calculations. The value of the Milankovitch theory197

was that it made testable predictions about the geological record of climate. It predicted how many ice-age deposits geologists would find, and it pinpointed when these deposits had been formed during the past 650,000 years.

These predictions were contained in three nearly identical radiation curves that showed past changes in summertime radiation at latitudes 55Deg, 60Deg, and 65Deg North. In theory, each radiation minimum caused an ice age.

A number of findings up through the late 1950s threw much doubt on the Milankovitch theory. But eventually the theory was confirmed. Let's now look at the evidence. In the early 1950s,198

One of the first American geologists to advocate the systematic use of the radiocarbon method in the study of Pleistocene drifts was Richard F. Flint at Yale University. After collecting a large number of datable materials from the Wisconsin drift of the eastern and central United States, Flint sent them off.... for radiocarbon analysis. Flint's results showed that the drift actually recorded at least two glaciations -- perhaps more. Previously, it had been supposed that a single glaciation was responsible for the Wisconsin drift, but the radiocarbon results made it clear that this hypothesis could no longer be maintained. The older tills in the drift were, for the most part, beyond the range of radiocarbon dating; but the youngest till was well within the datable range, and Flint and [others] were able to show that the great ice sheet had reached its maximum extent 18,000 years ago. Then, about 10,000 years ago it rapidly disappeared.

For a time it seemed that the results of the radiocarbon revolution were consistent with the Milankovitch theory. Although it was true that the 18,000-year date for the last glacial maximum was 7000 years younger than the 25,000-year date calculated by Milankovitch for the last radiation minimum, such a discrepancy could easily be explained as the time needed for a sluggish ice sheet to respond to a change in the earth's radiation budget. In fact, Milankovitch himself had predicted that just such a lag should occur, and estimated its duration as about 5000 years.

However, the discovery of a 25,000-year-old peat layer in Farmdale, Illinois, finally shattered belief in the Milankovitch theory. Such a deposit could only have been formed during an interval of relatively warm climate. Exactly how warm was uncertain, but the date for that warm interval coincided exactly with the date of a radiation minimum. When deposits of the same age and type were found at other locations in the Midwest, in eastern Canada, and in Europe, the geological evidence against the astronomical theory seemed to be overwhelming.

The program of radiocarbon dating allowed more and more geologists to fix their field observations on a firm time scale. This led to the development of a new method for constructing a climatic curve that could be directly compared with the radiation curve. Geologists accomplished this by finding dates for a large number of till and loess samples along some convenient north-south line. This till-loess boundary could then be graphically represented as a function of time. The resulting jagged line showed the position of the southern margin of the ice sheet as it advanced and retreated at that particular longitude over the course of thousands of years.... [One diagram along a line from Indiana to Quebec] indicated that major glacial advances occurred 60,000, 40,000, and 18,000 years ago.

Although evidence concerning ice ages continued to be gathered on land, it was fragmentary and piecemeal. But in the ocean, continuous sediment records going back hundreds of thousands of years were found. In 1872 the British research ship H.M.S. Challenger observed that tiny planktonic animals called foraminifera (informally forams)199

are found in all the world's oceans -- some species only in warm waters, others only in cold waters. For those who would later seek to unravel the mystery of the ice ages, this was a highly significant finding; it meant that an examination of fossil foram sequences on the seabed could indicate whether the ocean was warm or cold at the time that the creatures died and sank to the bottom....

Studying cores taken in this way from the Atlantic floor in the mid-1920's, the German paleontologist Wolfgang Schott made a critical finding. He recognized in the sediments three distinct layers with different populations of forams. The uppermost layer, laid down in the recent geological past, continued a high concentration of warm-climate species, including Globorotalia menardii. The second, older layer was richer in cold-climate types, and menardii was completely absent. But in the oldest of the three layers, menardii was back again, together with a high proportion of other warm-climate species. Schott deduced that the layer devoid of menardii was deposited during the last ice age, when the Atlantic was cooler, while the other two strata were laid down during the preceding and present interglacials.

Lest the reader think this sequence of warm-cool-warm forams is consistent with the Society's view that a large influx of cold water inundated a tropical earth, note first that the sequence is wrong. There should have been warm species, then cold species, and nothing else. In addition, if the earth were in a hothouse condition, how could there be cold water species at all? If much water fell from a vapor canopy it would become extremely hot as it fell. Also note that the cores described above were about three feet long. Cores taken using newer techniques are up to 45 feet long, and contain many warm-cool cycles of forams, as well as other evidence showing multiple glacial cycles. I'll discuss the evidence for multiple cycles presently. Meanwhile, Time-Life's Ice Ages describes an outstanding correlation between events on the land and events in the sea:200

In the early 1950s.... David B. Ericson.... [and] Geochemists at [Columbia University's] Lamont [Geological Observatory] determined that the boundary between the topmost, menardii-rich layer and the second layer had been formed rapidly, some 11,000 years ago. Ericson noted that the timing of this rapid change from cold to warm climate coincided with the dates that had been deduced from radiocarbon dating of glacial debris on land. In a paper describing their findings, Ericson and his colleagues noted that "further correlation of events both in the ocean and on land during this interval may lead to an understanding of some of the factors causing glaciation."

Meanwhile, other scientists were conducting a parallel line of research that involved chemical analysis of fossil forams. The method they used had been suggested in 1947 by Harold Urey, a Nobel laureate at the University of Chicago. It consisted of measuring the ratio of two oxygen isotopes -- atoms that are virtually identical but different in atomic weight -- that are absorbed from sea water by the shells and skeletons of marine organisms. Urey and his associates had found that organisms from cold water contained a higher proportion of the heavier isotope, designated oxygen 18, or O-18, than did organisms living in warmer water. The remains of the warm-water creatures had a higher ratio of the lighter isotope, known as O-16.

In the 1950s, the Italian-American geologist Cesare Emiliani applied Urey's theories to eight deep-sea cores. After radiocarbon-dating the upper sections of the cores and then estimating the rates of sedimentation, Emiliani decided that there had been no fewer than seven complete glacial-interglacial stages during the past 300,000 years and that they had occurred in a time sequence that agreed fairly well with the variations predicted by Milankovitch. In its broad outlines, Emiliani's work also agreed with Ericson's findings, although there were some major differences; certain periods that Ericson's foram analysis had identified as warm were shown by Emiliani's methods to have been cold.

So spirited was the debate over the contradictory findings that in 1965 the National Science Foundation held a special conference to try to settle the dispute. John Imbrie, then a professor of geology at Columbia University, attended the meeting, and later told the story of the controversy and its aftermath in his book, Ice Ages: Solving the Mystery. At the conference, Imbrie pointed out that Ericson and Emiliani had all but ignored the possibility that factors other than temperature may cause variations in foram concentrations. Imbrie decided then and there to develop an analysis technique that took into account such things as water salinity and the amount of food available, as well as water temperatures in winter and summer.

In his book, Imbrie says that he and his assistant201

.... developed a multiple-factor method for climatic analysis that took into account abundance variations in all 25 species of planktonic forams. In many respects, their approach was a computerized extension of the technique used by Wolfgang Schott in 1935.

At a meeting held in Paris in 1969, Imbrie announced the results that he had obtained when he studied a Caribbean core with this multiple-factor technique. Whereas Emiliani's research indicated that surface water temperatures in the Caribbean had dropped by almost 11Deg F. during the last ice age, Imbrie's multiple factor method showed a drop of only 3.5Deg F. When the core was analyzed for oxygen-isotope ratios, the zones that Ericson had identified as cold were shown to be warm by both the isotope and multiple-factor methods. Imbrie said:202

Apparently, some environmental factor other than surface water temperature (but often correlated with it) caused Globorotalia menardii to appear and disappear cyclically in deep waters of the Atlantic Ocean.

At the Paris meeting Imbrie talked after the lecture with a British geophysicist named Nicholas Shackleton. They203

realized that their independent work on the problem had led them to the same answer: Changing ratios of oxygen isotopes in marine fossils are caused primarily by fluctuations in the size of ice sheets, not by variations in sea temperatures. Their tentative conclusion was based on the fact that because O-18 is heavier than O-16, water molecules containing O-18 do not evaporate as readily; therefore, water rising from the oceans in the form of vapor and subsequently falling as precipitation contains a smaller proportion of O-18 than do the oceans themselves. If water deficient in O-18 were to be locked up on land in the ice sheets, the proportion of the heavy isotopes in sea water would rise, and this increase would be reflected in the ratios of the oxygen isotopes present in forams and other marine organisms.

This type of enrichment process using differences in the weight of two isotopes has had practical application. In the purification of uranium for atomic bombs one process uses a gas centrifuge to separate molecules of uranium hexafluoride gas containing the lighter, bomb grade U235 isotope from the heavier U238. I hardly need comment on how effective this process is.

Evidence confirming the Milankovitch radiation curves continued to appear. In 1965204

geochemist Wallace S. Broecker reported some interesting findings that he and some colleagues had made when they dated fossil coral reefs in the Florida Keys and the Bahamas. Since coral can grow only at certain depths, it provides an accurate record of former sea levels. Broecker's studies indicated that the sea had stood much higher 120,000 and 80,000 years ago, presumably during periods of warm climate when vast amounts of water had been released from the melting ice sheets. Noting that present sea levels are also considerably higher than they have been at times of great glaciation, Broecker observed that these three known periods of high sea levels closely correspond to the warm periods calculated by Milankovitch in his radiation curve for lat. 65Deg N.

Soon, other researchers were reporting similar findings elsewhere. Brown University geologist Robley K. Matthews, for example, investigated the terraced coastlines of Barbados and determined that the steplike terraces had been formed by the growth of coral reefs at former sea levels. According to his calculations, the age of one terrace reef was 80,000 years, that of another was 125,000 years -- a near-perfect match with the findings of Wallace Broecker. But Matthews found something quite different, too: a middle terrace that indicated a time of high sea level about 105,000 years ago.

Sadly for believers in the astronomical theory, the Milankovitch curve did not show a radiation maximum in that time period -- not, at least, at lat. 65Deg N., where the effects of axial-tilt variations space the radiation peaks at intervals of some 41,000 years. But when Broecker -- curious about the seeming anomaly of the 105,000-year-old Barbados shoreline -- examined additional Milankovitch curves, he found that those for lower latitudes showed peaks corresponding to all of the dates assigned to the three Barbados terraces. At these latitudes, it seemed, the precession cycle's 22,000 year period -- the time it takes for the earth's axis to wobble in a complete circle in space -- was influential enough to modulate the effects of axial tilt. Reef terraces in Hawaii and New Guinea yielded similar data, indicating that past periods of high sea level could indeed be explained by application of the astronomical theory of ice ages.

A striking photo of the New Guinea terraces appears on page 145 of Ice Ages: Solving the Mystery. These terraces are compelling evidence, as well, for the theory of plate tectonics. The island of New Guinea is being uplifted by the Australian tectonic plate plunging under the island arc of which New Guinea is a part. This uplift is seen in the terraces, and shows how well geological evidence in the fields of ice age research and plate tectonics correlate.

Sunken reefs were found as well. These sometimes contain caves that could only have been formed when the reef structure was above sea level. In "Jamaica a series of drowned and overgrown ridges can be seen at 25, 40 and 60 meters below the present sea level. These drowned reefs were formed during periods of intensive glaciation 8,000, 11,000 and 14,000 years ago, when the sea level was considerably lower than it is today."205 Time-Life's Ice Ages continues with the evidence:

On a different scientific front, researchers were seeking to refine geological chronologies by matching switches in the magnetic polarity of undersea sediments. Tiny iron particles in most rocks become permanently magnetized in alignment with the earth's magnetic poles at the time the rocks are formed, and scientists had recently found that the same process occurs in ocean sediments. The phenomenon of magnetic reversal -- which is probably caused by disturbances in the earth's molten core -- had first been noted in 1906 by Bernard Brunhes, a French geophysicist who discovered that the iron-rich particles in an ancient lava flow had been magnetized so that the north and south magnetic poles were interchanged. During the 1960s, scientists using the recently developed potassium-argon dating technique -- which measures the rate at which a radioactive potassium isotope found in rocks changes to an isotope of argon -- determined that the earth had reversed its magnetic polarity a number of times during the past four million years. The last switch took place about 700,000 years ago. All deposits laid down since this occurrence, which marked the start of the period called the Brunhes Epoch, have "normal" magnetic polarity; strata laid down during the 300,000 years prior to the event show reverse polarity; and before that the orientation was normal.

Ice Ages: Solving the Mystery says on page 148, concerning the geologists who established the magnetic reversal time scale:

Cox and his colleagues proved that the field-reversal theory was correct by showing that each reversal had been a globally synchronous event. They argued that it would be unreasonable to suppose that lava flows all over the world had undergone self-reversal simultaneously. To demonstrate synchroneity, they dated lava flows occurring just above and just below a large number of reversals. This dating effort, carried out by a group of investigators at the University of California.... was based on the potassium-argon method -- a technique that worked particularly well with lava flows. The results not only established the synchroneity of magnetic reversals, but also focused attention on the reversal dates themselves. These dates proved to be the long awaited fixed points around which a firm Pleistocene chronology could be constructed.

Time-Life's Ice Ages continues:

Since deposits from land and sea and from all parts of the world contain the same magnetic records, the identification of the times of the reversals is a means of correlating the geological chronologies estimated by different methods in different regions. Older deposits revealed that a magnetic switch took place about two million years ago. This is a significant date, for it corresponds roughly to the time that geologists assign to the beginning of the Pleistocene epoch and the start of the glacial-interglacial cycle that has continued to the present.

Now, armed with a way to date sea-floor sediment cores that contained evidence of past climate changes, scientists would be able to determine whether previous cold and warm periods coincided with the cycles predicted by Milutin Milankovitch. But such proof would also have to explain the fact, deduced from continuing studies of sediments on land and on the sea floors, that the 22,000- and 41,000-year cycles -- which Milankovitch had believed to be most critical to radical changes in climate -- seemed to be superimposed over longer cycles of 100,000 years, a figure reminiscent of [19th century geologist James] Croll's theory that variations in the earth's orbital eccentricity are paramount in bringing about climate changes. It appeared that the great Pleistocene ice ages had developed slowly in cycles of about 100,000 years; after a number of oscillations, each ice age had come to an abrupt end. Only if Milankovitch's shorter cycles could be related in some way to these 100,000-year periods could his astronomical theory be accepted as an explanation of the cause of the ice ages.

By about 1968, research on loess deposits shed more light on the pulsebeat of climate. It turned out that loess deposits are not the result of just one glaciation. They sometimes contain a record of many glacial cycles. Ice Ages: Solving the Mystery206 describes the investigations of a geologist name George Kukla in 1968 at a brickyard quarry on Red Hill near the city of Brno in Czechoslovakia:

Kukla's interest in loess was an outgrowth of his fascination with Czechoslovakian caves. In many of these caves, thin layers of loess blown in during Pleistocene ice ages have been found to contain the bones of Neanderthal and other Stone Age people. By tracing these layers of loess outside of the caves, and correlating them with the thicker sheets that cover the sides of nearby hills, archaeologists could place the human artifacts in an historical sequence.

The Pleistocene ice sheets that flowed outward from centers in Scandinavia and the Alps never reached the Red Hill region, yet the climate there changed drastically. As early as 1961, George Kukla and his colleague Vojen Lozek had explained why the nonglaciated areas of Czechoslovakia and Austria were ideally located to record the fluctuations of Pleistocene climate. When the ice sheets were large, Central Europe was a polar desert -- dry, treeless, and swept by bitter winds that deposited layers of loess. But when the glaciers were small, Czechoslovakia had a climate even warmer and wetter than today's: broad-leafed trees grew in forests, fertile soils were formed, and Stone Age hunters lived under temperate conditions. Thus, as the Scandinavian and Alpine ice sheets alternately expanded and contracted, the boundary between prairie and forest marched back and forth across the nonglaciated corridor of Central Europe.

Long before they were aware of the magnetic time scale, Czechoslovakian geologists had demonstrated that at least ten repetitions of the soil-loess cycle were recorded in the region of Brno alone. But it had not been possible to determine how long each cycle was. In 1968, Kukla and his colleagues at the Czechoslovakian Academy of Science returned to their brickyards, examined each layer of soil and loess, and found five magnetic reversals. With the time scale now fixed, the average length of each cycle could easily be calculated: the main pulse of late Pleistocene climate was a steady beat of one cycle per 100,000 years.

Investigations carried out over the preceding decade had established that the sedimentary cycle was not really a simple repetition of soil (layer 1) and loess (layer 2) in a symmetrical pattern (1-2-1-2). Instead, it was a four-fold cycle made up of three kinds of soil (1, 2, and 3) and loess (4), forming a sawtoothed, asymmetrical sequence (1-2-3-4-1-2-3-4). The first soil in the sequence formed in a warm, moist climate. The second layer was a black soil, identical to that forming now in the moister parts of the Asiatic steppe, and containing fossils indicative of a climate somewhat cooler and dryer than that of the preceding forest phase. Above the black soil was a layer of brown soil, typical of the more temperate parts of Arctic regions today. This soil, the third layer in the sequence, contained fossils indicative of a climate colder and dryer than the steppe, but not as cold and dry as that which accompanied the deposition of the overlying loess sheet that formed the fourth and final phase of the cycle.

These observations led Kukla to an important conclusion: the cooling phase of the climatic cycle lasted much longer than the warming phase. Moreover, transitions from dusty, polar desert phases to deciduous forest phases were so abrupt that they appeared in the quarry walls as distinct lines. These lines, named "Marklines" by Kukla, were useful in distinguishing one sedimentary cycle from another, and in correlating the cycles between widely separated regions.

In the Soviet Central Asian republic of Tadzhikistan an extensive loess area was excavated. The loess deposits are as much as 200 meters deep in some places. According to a Scientific American article,207

They contain evidence of an apparently continuous sequence of warm-to-cold climatic oscillations that span the past two million years.... One immediately apparent feature of the Tadzhikistan loess exposures is the alternation of thick layers of unaltered loess and distinct "horizons" of soil. The soil horizons were formed when the surface of the loess was altered in periods of relatively moist and warm climate. In many places the soil structure is complex.... The plant pollen and snail species found in the loess.... indicate that it accumulated when the climate was considerably cooler and drier than it is today. Thus the alternating layers of loess and soil are evidence of major climatic oscillations in the region....

Various paleomagnetic events are detectable in the loess of Tadzhikistan, the most important one being the transition some 690,000 years ago between the Matuyama [previous reversed magnetic period].... and the Brunhes period of today. In the Tadzhikistan loess sections currently being studied the total number of buried soil complexes varies, the maximum being 37. In the six sections where the Matuyama-Brunhes boundary has been detected, however, nine of the soil complexes consistently lie above it. The number of soil complexes above the.... boundary.... corresponds fairly closely to the number above the boundary in the loess of central Europe. The number of soil complexes is also in agreement with the record of climatic oscillations preserved in deep-sea sediments.

Ice Ages: Solving the Mystery describes what was happening on another front. As Kukla was doing his research,208

Jan van Donk at Lamont was completing isotopic measurements of forams in Caribbean core V12-122. Along with Broecker, van Donk was attempting to improve the geological time scale. Because the core did not extend to the base of the Brunhes Epoch, the magnetic time scale could not be applied directly. However, the core did contain the U-V boundary -- which Ericson had dated as about 400,000 years old by interpolation in cores long enough to contain the last magnetic reversal. This estimate, falling as it did in the middle of the rather broad range of dates obtained by uranium and thorium methods, became the cornerstone of Broecker and van Donk's chronology, and led them to conclude that the major cycle in the isotopic record was 100,000 years. Moreover, they noted that this primary climatic cycle had an asymmetrical, sawtoothed shape: "Periods of glacial expansion averaging about 100,000 years in length were abruptly terminated by rapid deglaciations." They labeled these episodes of rapid warming, "terminations."

Not until September 1969, when Broecker and Kukla met at the international scientific congress in Paris, did they realize that their separate lines of research had led them to many of the same conclusions: the major ice ages of the Pleistocene were spaced about 100,000 years apart, developed slowly, and terminated abruptly. The marklines in the Czechoslovakian brickyards corresponded to the terminations in the Caribbean cores.

More evidence was found in the Atlantic Ocean. Ice Ages: Solving the Mystery describes what was found:209

While Broecker and Kukla discussed the shape of Pleistocene cycles -- and Imbrie and Shackleton exchanged views on Pleistocene temperatures -- William Ruddiman and Andrew McIntyre were hard at work at the Lamont Observatory, developing a new method for studying the history of the ocean. By selecting cores along a north-south line, and recording the changing distributions of temperature-sensitive species along that line, they were able to trace the shifting course of the Gulf Stream. During interglacial intervals, the current had flowed northeast across the Atlantic from Cape Hatteras towards Great Britain. But during the ice ages, it had taken an easterly course towards Spain. As the ice sheets expanded and contracted, and forests and prairies swept back and forth across Europe and Asia, the Gulf Stream swung back and forth like a gate hinged on Cape Hatteras. By counting the number of "swings" of the current, and keying these into the magnetic time scale, Ruddiman and McIntyre found that there were eight climatic cycles within the Brunhes Epoch. Like the Arctic ice sheets and the Eurasian forests, the ocean's currents marched to a 100,000-year beat.

But up until this time most scientists were not yet convinced of the Milankovitch theory. Ice Ages: Solving the Mystery describes the situation:210

By 1969, the magnetic time scale had proved its value as a basis for studying the history of the ice ages, and had made it possible to identify the dominant pulsebeat of climate as the 100,000-year cycle. But the advent of that time scale had so far done little to generate much support for the astronomical theory. On the contrary, it was something of an embarrassment that the 100,000 year cycle had not been predicted by that theory. Not until after the facts were in had Kukla and [others] suggested how the Milankovitch theory could be modified to account for that cycle.

Most scientists, therefore, would only be convinced that the astronomical theory was correct if it could be shown that the small oscillations superimposed on the 100,000-year cycle were those that Milankovitch had predicted. If these shorter climatic cycles turned out to correspond to the 41,000-year cycle of axial tilt, and to the 22,000-year cycle of precession, then the astronomical theory of the ice ages would be confirmed. To demonstrate such a correspondence, however, parallelisms between astronomical and climatic curves must be demonstrated in records sufficiently detailed to exhibit the 22,000-and 41,000-year cycles. Once more, the problem of testing the astronomical theory of the ice ages hinged on increasing the accuracy of the geological time scale.

Time-Life's Ice Ages continues its narrative:

In the spring of 1971, as part of the International Decade of Ocean Exploration, a group of scientists and researchers organized a series of studies known as CLIMAP -- the Climate Long Range Investigation, Mapping and Prediction project. One of their first missions was to analyze sea cores and deduce the climate changes that have taken place during the 700,000-year Brunhes Epoch.

To achieve the goal, investigators needed a core rich in forams that could be analyzed for oxygen isotopes. In December, CLIMAP scientists located such a specimen -- it had been raised from the western Pacific early in the year -- and after confirming that it dated back beyond the magnetic reversal that marked the start of the Brunhes Epoch, they shipped samples of the core to Nicholas Shackleton at Cambridge University.

Shackleton, an expert at analyzing the isotopic contents of marine fossils, studied the core samples and plotted two isotopic curves, one showing the ratio of light and heavy oxygen isotopes in the remains of surface-dwelling forams, and the other plotting isotopic variations in forams that lived on the sea floor. If, as Cesare Emiliani had theorized some years earlier, the proportion of oxygen isotopes in marine fossils is governed by sea temperatures, the second curve should have shown much smaller deviations than the first: No matter what the climate, the temperature of the water at the bottom of the ocean remains close to freezing. In fact, as Shackleton showed the CLIMAP team in mid-1972, the two isotopic curves were nearly identical.

It was just as Shackleton and John Imbrie had surmised in Paris three years before. Imbrie said that both of Shackleton's curves211

reflected changes in the proportion of light isotopes in the ocean -- not changes in water temperature. And, because sea water was mixed rapidly by currents, any chemical change in one part of the ocean would be reflected everywhere within a thousand years. All along, Emiliani's curve had been a chemical message from the ancient ice sheets. When the glaciers expanded, light atoms of oxygen were extracted from the sea and stored in the ice sheets -- altering the isotopic ratio of oxygen in sea water. When the glaciers melted, the stored isotopes flooded back into the ocean, returning it to its original composition. The effect of local variation in temperature was too small to be detected.

In short, the two curves were not directly indicating changes in climate; instead, they documented a consequence of climate change -- the waxing and waning of glaciers, the comings and goings of the ice ages. Time-Life's Ice Ages continues:

Shackleton's painstaking analysis yielded additional revelations. The core showed a definite sequence of 19 stages of warming and cooling over the past 700,000 years, making it possible for scientists to estimate the duration of each stage. More significantly, there were clear indications that major climate changes had occurred at intervals of some 100,000 years, the same time period suggested by the notion that climate was affected primarily by changes in the eccentricity of the earth's orbit around the sun.

The eccentricity of the earth's orbit is important to climate because212

the intensity of radiation during any particular season is largely controlled by the precession cycle -- the amplitude of which is exactly proportional to eccentricity. When the orbit is unusually elongate, the winters are colder than average and summers are warmer. Therefore, if the temperature during one particular season is critical to the expansion or retraction of ice sheets, it follows that the 100,000-year cycle must be reflected in the climatic record.

Back to Time-Life's Ice Ages:

This 100,000-year cycle was so dominant on Shackleton's curve that he could not determine whether the less-pronounced fluctuations reflected the 41,000-year axial-tilt cycle and the 22,000-year-old precession cycle. Milankovitch's astronomical theory of the ice ages remained unproved. Before long, however, another CLIMAP researcher, James D. Hays of Columbia University, clarified the situation by looking at two sediment cores from the southern Indian Ocean -- one that had been raised in 1967 and one that had been brought up in 1971. While the cores' records did not extend all the way back to the start of the Brunhes Epoch, they went far enough -- 450,000 years -- to provide a sufficient time span for valid analysis. Moreover, the sediments had built up more rapidly than those in Shackleton's Pacific core, providing a thicker accumulation for each cycle -- and thus a more detailed account of the climate changes that had occurred.

When Hays and Shackleton examined the evidence from the Indian Ocean cores, they found clear imprints of the 100,000-year cycle. They also saw unmistakable signs of the shorter cycles of 41,000 and 22,000 years. "We are certain now," they announced, "that changes in the earth's orbital geometry caused the ice ages. The evidence is so strong that other explanations must now be discarded or modified."

Earlier conclusions drawn from radiocarbon dating -- which at first seemed to invalidate the Milankovitch theory -- had already been modified considerably. As geological knowledge expanded, researchers realized that a slight waning of the ice sheets occurred some 25,000 years ago, indicating that climates would have been warm enough to produce such apparent anomalies as a 25,000-year-old peat layer in Illinois.

Ice Ages: Solving the Mystery gives a more detailed account. Researchers were having difficulty in showing conclusively that the 41,000-and 22,000-year cycles were actually represented in some of the sea cores.213

Why was it proving so difficult to find out what the higher frequencies in the climatic curves were? Reviewing this problem in the fall of 1972, Hays thought he knew the reason: the cores that had so far been analyzed spectrally213a had accumulated too slowly.

He argued that when accumulation rates were as low as one or two millimeters per century -- as they were in most Pacific and Caribbean cores -- the burrowing activities of animals living on the sea floor would blur the record of the higher frequency cycles. To make a valid test of the Milankovitch theory, therefore, it would be necessary to analyze an undisturbed core whose accumulation rate had exceeded two millimeters per century.

Hays and his CLIMAP colleagues were already studying all of the available cores as part of their effort to map the ice-age ocean. After some reflection, Hays decided that they would now search for a particular type of core: one that had a suitably high sedimentation rate -- that was located in the high latitudes of the southern hemisphere -- and that contained shells of both forams and radiolaria. Such a core, Hays reasoned, would provide more information than one located in the northern hemisphere. Variations in the isotopic composition of foram shells would provide a record of ice-sheet fluctuations in the northern hemisphere -- for nearly all of the glacial expansion and contraction that influenced the isotopic composition of the ocean took place there. At the same time, changes in radiolarian populations could be analyzed by the multiple-factor technique and made to reveal what the history of water temperature had been over the coring site. By comparing the two signals -- isotopic and radiolarian -- Hays hoped to be able to answer a question that had first been raised by [19th century geologist] James Croll: do climatic changes in the southern hemisphere coincide with those in the northern hemisphere?

In January 1973, Hays located a core in the Lamont collection that seemed to meet his requirements. Core RC11-120 had been raised from the southern Indian Ocean six years earlier by Geoffrey Dickson aboard the Robert Conrad. After counting the radiolaria and sending samples to Shackleton for isotopic analysis, Hays was gratified to find that the deposition rate was high enough for his purposes (three millimeters per century). When the data were plotted, the answer to Croll's question was immediately apparent: climatic changes in the northern hemisphere were essentially synchronous with those of the southern hemisphere. Although this result alone was important enough to justify his efforts, Hays was disappointed to find that the core only extended back about 300,000 years, to the base of Stage 9 in Emiliani's isotopic scheme. To provide a suitable record for spectral analysis, a core extending back at least 400,000 years would be needed.

When it became clear that the needle Hays was looking for was not to be found in the Lamont haystack, he decided to search elsewhere. In July, he went to Florida State University in Tallahassee, where an extensive collection of Antarctic cores was maintained. There, he continued the search for cores taken near the site of RC11-120. Soon he came upon several cores taken by Norman Watkins aboard the Eltanin in 1971. With the assistance of two graduate students, Hays began to open the Watkins cores. Later he would recall: "The cores were kept in cold storage, and we were all shivering in our parkas. But when core E49-18 was opened, we stopped shivering. I knew right away we had something interesting because the color-banding matched perfectly with the oscillations in Shackleton's oxygen curve for V23-238." Counting down, Hays found that the core extended to Stage 13 -- giving it an age of 450,000 years. He had found his needle at last.

Hays's off-the-cuff stratigraphic analysis turned out to be correct. Core E49-18 did indeed extend back to Stage 13. Unfortunately, the top three isotopic stages had been lost when the core was taken; but with the isotope stratigraphy now available, these could be patched in from the nearby core RC11-120. Together, these two cores contained a detailed and undisturbed record of climate extending back 450,000 years -- and their accumulation rate was high enough to have preserved cycles as short as 10,000 years.

When the radiolarian and isotopic data had been graphed, Hays and Shackleton were elated. For the isotope curves in the Indian Ocean matched the general pattern that Emiliani had established for Stages 1 through 13 in a number of other cores. But now, as Hays had anticipated, frequencies higher than the 100,000-year cycle were clearly visible (Figure 40). Realizing that an opportunity to make a definitive test of the Milankovitch theory was at hand, he asked Imbrie to carry out a spectral analysis.

The first objective was to find out exactly what the frequencies of variation in tilt and precession had been over the past 450,000 years (Figure 41). These frequencies, rather than the frequency of the eccentricity cycle, would be crucial to the coming test -- because they alone were unambiguously predicted by the Milankovitch theory. Imbrie knew that Anandu D. Vernekar, at the University of Maryland, had recently recalculated the astronomical curves, and Imbrie obtained copies of the calculations from him. After processing Vernekar's information statistically, Imbrie found that, as expected, the tilt curve showed a single cycle of 41,000 years. But the spectrum for the precession curve contained not one, but two distinct cycles -- a major precessional cycle of 23,000 years, and a minor cycle of 19,000 years. Concerned that his calculations had been wrong, Imbrie laid his results before Belgian astronomer Andre Berger. After examining the trigonometrical formulas from which the precession calculations were derived, Berger announced that the double cycle that Imbrie had found was not a statistical error: variations in earth-sun distance do in fact occur as 23,000-year and 19,000-year cycles.

Berger's endorsement set the wheels in motion. According to the expanded version of the astronomical theory developed by Mesolella and Kukla, climatic oscillations should occur as four distinct cycles: a 100,000-year cycle corresponding to variations in eccentricity; a 41,000-year cycle corresponding to variations in axial tilt; and 23,000- and 19,000-year cycles corresponding to variations in precession. In the summer of 1974, Imbrie performed the long-awaited test. Spectral analysis indicated that, as expected, the dominant climatic pulse was a 100,000-year cycle, which appeared on both the isotopic and the radiolarian spectra as a large peak. But three other peaks -- smaller but nevertheless distinct -- also appeared on the spectra (Figure 42). On the isotopic spectrum these cycles were 43,000 years, 24,000 years, and 19,000 years long. On the temperature-radiolarian spectrum, they were 42,000 years, 23,000 years, and 20,000 years long.

These results were everything for which Imbrie and his colleagues had hoped. Each of the cycles found in the Indian Ocean cores matched the predicted cycles within five percent. That such a coincidence might occur by chance alone seemed highly unlikely. Before long, Nicklas G. Pisias provided additional evidence in support of the astronomical theory. Using a more powerful spectral method, he found a statistically significant 23,000-year cycle in core V28-238. CLIMAP investigators -- realizing that their isotope records from the Pacific and Indian oceans matched the corresponding parts of isotope records already known from other oceans -- felt justified in concluding that the succession of late Pleistocene ice ages had indeed been triggered by changes in the earth's eccentricity, precession, and tilt.

If the astronomical theory were correct, it should be possible to do more than demonstrate by spectral analysis that the astronomical frequencies appeared in climatic curves. It should also be possible to discover how rapidly the ice sheets had responded to each type of astronomical variation. For example, if the ice sheets responded instantly to a change in axial tilt, then the fluctuations of the 41,000-year climatic cycle should have occurred simultaneously with variations in tilt. But if, as seemed more likely, the ice sheets were slow in responding to a change in radiation caused by changes in tilt, the 41,000-year cycle of climate should follow regularly behind the orbital curve.

Discovering that a statistical technique called filter analysis was available to examine the 41,000-year and 23,000-year frequency components of a climatic curve separately, Imbrie applied this method to the records from the two Indian Ocean cores. The result showed clearly that the 41,000-year climatic cycle did lag behind variation in axial tilt by about 8000 years. And, for at least the major portion of the record under study, the 23,000-year climatic cycle lagged systematically behind variations in precession. Moreover, these lags were regular enough to confirm the inference that variations in tilt and precession set the pace for climatic change.

Convinced now that major climatic changes were caused by astronomical variations, and that the 41,000-year and 23,000-year climatic cycles followed systematically behind variations in tilt and precession, Hays, Imbrie, and Shackleton announced their findings in an article in Science, which appeared on December 10, 1976: "Variations in the Earth's Orbit: Pacemaker of the Ice Ages."

A century after Croll published his theory and 50 years after Milankovitch mailed his radiation curves to Koppen and Wegener, two cores from the Indian Ocean confirmed the astronomical theory of the ice ages. At last, geologists had clear evidence that the motions of the earth in its orbit around the sun triggered the succession of late Pleistocene ice ages. Exactly how this triggering mechanism operated, and why the 100,000-year cycle of orbital eccentricity appeared to be so strongly impressed on the climatic record of the last half-million years were still unknown. But, for the moment, it was enough to know that Milutin Milankovitch, traveler through distant worlds and times, had led the way to solving a major part of the ice-age mystery.

In March 1941, looking back on a lifetime devoted to finding the cause of the ice ages, Milankovitch had reflected that:

These causes -- the changes in insolation brought about by the mutual perturbations of the planets -- lie far beyond the vision of the descriptive natural sciences. It is therefore the task of the exact natural sciences to outline this scheme, by means of its laws ruling the universe and by its developed mathematical tools. It is left, however, to the descriptive natural sciences to establish an agreement between this scheme and geological experience.

The spectral, or Fourier analysis used to obtain frequency components from the sea core isotopic measurements, the temperature-radiolarian measurements, and the Milankovitch theory's mathematics, is a mathematical technique taught to all undergraduate physics and electrical engineering students. The technique has been reduced to a rather simple algorithm called the Fast Fourier Transform that can be easily programmed on a small computer. The technique is completely straightforward. When properly applied, it correctly extracts frequency component information from a signal in an objective manner. The technique has no more room for subjective judgement than the simple algorithms for multiplication or long division we all learned in beginning arithmetic. The point is that all a geologist has to do is put his data into a computer programmed to do Fourier analysis, and out comes the frequency components.

The way this applies to the preceding discussion is that once the base measurements were made, Fourier analysis extracted the frequency components. There was no subjective judgement in figuring out the frequency components. The frequencies that the analysis said were in the data were really in the data.

It is amazing that Milankovitch's mathematical theory, first published about 1920, could be so accurately confirmed 50 years later by two independent geological phenomena. This theory was further confirmed in a most unusual way. The Innocent Assassins described how lake varves of two types form -- see Part 14 of this essay. The book then describes what was found in similar ancient lakes:214

It might reasonably be expected that such lakes existed in time past too -- having vanished long ago but leaving a fossil lake bed perhaps protected by other, covering sediments. And so it is; the number in fact is legion, and we can take note of only a few examples. Let us start with a look at the lakes that existed during the Ice Age, or more precisely, during the interglacial periods -- times when the climate was as warm as it is now. (We, of course, live in another such interglacial period, the warmest part of which is already long past -- and which, as Angerman River and Lake Valkiajarvi teach us, has lasted for about 10,000 years.)

Fossil lake bottoms of the same type -- but overlain by "cold" sediments laid down during a glacial age -- are known in various parts of the world. In Germany the annual varves of one of these lake bottoms were counted, while at the same time, by analysis of the fossil plant pollen in the sediment, it was shown that the entire history of the interglacial was recorded there. The record starts with a tundra vegetation, shows the immigration of birch and pine, attains a culmination with oak and other broad-leaved hardwoods, and then returns by stages to the tundra. All this, as shown by the Quaternary geologist H. Muller, was enacted within 11,000 years -- a length of time comparable with that of our current interglacial.

Compare this sequence with the sequences seen in the previously described stages of loess formation in Czechoslovakia and Tadzhikistan.

The chronology, of course, is again a floating one. All we can say on this basis is that it dates back some considerable time, because it was followed by a cold climatic oscillation, which in turn gave way to the present interglacial period. Radiometric dating, however, indicates that the interglacial occurred about 120,000 years ago.

Based on still earlier interglacial lake beds in Germany and England, similar studies indicated durations of from 16,000 to well over 20,000 years. Yet all the interglacials appear as relatively short episodes in the long, cold-dominated history of the Ice Age.

With such rapid climatic changes, lakes tend to have had a rather short life, just a few tens of thousands of years -- short, that is, when we start looking at lakes that existed before the Ice Age. We are then well back in the Tertiary period, when climatic change, as a rule, seems to have been slower, or at any rate, not at all extreme. And we meet with chronologies on another scale entirely.... [In many ancient lakes] annual varves prove the lakes to have existed for hundreds of thousands of years.

Still, there are longer sequences. During the Eocene epoch, a lake basin formed in North America, covering large parts of what are now the states of Colorado, Wyoming, and Utah. The area had previously been traversed by rivers flowing east, but in connection with the rise of the Rocky Mountains a barrier was formed to the east. Thus arose Fossil Lake, and waxed greater and greater, finally to cover an area of about 13,000 square kilometers, with a depth of up to 100 meters. It was thus about half the size of Lake Erie, and twice that of Great Salt Lake.

Judging from the sediments that lie beneath and above those of Fossil lake, it existed for about one-third of the Eocene epoch. The lake-bottom sediments show a fine lamination with alternating light, lime-rich bands and dark ones containing a great deal of organic matter. There is also rich fossil flora and fauna (especially fish and insects). The flora indicate a climate of a subtropical type with two annual rainy seasons, and if the varves are interpreted on that basis (two dark bands to a year), it can be seen that the entire pile of sediments was formed during a time period of 6.5 million years.

According to radiometric dating, the Eocene epoch started about 55 million years ago, and ended about 35 million years ago. Its total duration would then be about 20 million years. One-third of that is 6.7 million years, or very close to the number of years that can be counted in the lake-bottom silts of Fossil Lake. The agreement is almost too good to be true, but there you are. Radiometric dating is supported by the chronology based on annual varves.

The longest consecutive sequence of annual varves that I have happened upon, in a far from systematic search, comes from eastern North America and constitutes about 40 million years. It dates from the later part of the Triassic period and the early part of the Jurassic period, and so has an age of about 220-180 million years before the present. My authority is Paul E. Olsen.

The geography of the earth at the beginning of the Triassic period was very unlike that of the present day. All continents were then collected into a single supercontinent, called Pangaea ("all land"). Then began the birth of the Atlantic: a great rift valley running in a northeast-southwest direction started to form. It has its counterpart in the present-day Rift Valley of East Africa, which also marks the place of a future ocean. The Triassic valley, a string of at least thirteen elongated basins, extended from Nova Scotia to North Carolina. In the basins, very fine-grained lake sediments were deposited (forming the so-called Newark Supergroup) with annual laminae, totaling some 40 million. Parts of the rift valley are now covered by the sea, but the southern part is dry, as are patches further north.

The sediments preserve a record of climatic changes, especially an alternation between dry periods, when the water was low, and times of high precipitation and high waterlines. During the former, annual varves tend to be very thin, drying cracks are formed, and there are numerous footprints of reptiles, in some cases so perfectly preserved that you can count the scales on the soles. At high water, the drying cracks are absent, and the sediments contain a great amount of organic matter, especially fish remains.

As it happens now, these climatic changes are cyclical: they tend to return at regular intervals. Their periodicity is complicated, however, because it is a combination of several cycles differing in length. (Each cycle represents a sequence from low through high to low water.) The most important periods are about 25,000, 44,000, 100,000, and 400,000 years in length, respectively. These are figures that cause the geochronologist to smile in recognition. They have to do with changes in the rotation of the earth around the sun,

(is this sounding familiar?)

known from astronomical calculations, and have turned out to drive the climatic changes during the Ice Age as well -- that is, during the last two million years or so of earth history. Now the analysis of the Newark Supergroup sediments proves that the same factors affected climates as long as 200 million years ago.

And so the geological time scale, originally dated by radiometry, is corroborated by two additional and completely independent methods of study: from analysis of annual varves, and from astronomical observations.

I consider the preceding descriptions -- of the confirmation of the Milankovitch theory and its correlation to ancient lake sediment layers -- to be unequivocal evidence that the ice ages really occurred. The burden of proof that they did not occur, and that the Flood produced all the observed evidence, is on whoever disagrees with this conclusion.

In the following section I present more evidence that ice ages occurred. If the reader is able to show that this is actually evidence for the Flood, I would like to hear about it.


195 ibid, pp. 61-62.

196 Windsor Chorlton, op cit, pp. 106-107.

197 John Imbrie, et al, op cit, p. 113.

198 ibid, pp. 120-121.

199 Windsor Chorlton, op cit, p. 129.

200 Windsor Chorlton, op cit, p. 131.

201 John Imbrie, et al, op cit, p. 139.

202 ibid, p. 139.

203 Windsor Chorlton, op cit, pp. 132-141.

204 ibid, p. 132.

205 Thomas F. Goreau, Nora I. Goreau and Thomas J. Goreau, "Corals and Coral Reefs," Scientific American, p. 133, New York, August, 1979.

206 John Imbrie, et al, op cit, pp. 153-154.

207 Richard S. Davis, Vadim A Ranov and Andrey E. Dodonov, "Early Man in Soviet Central Asia," Scientific American, p. 130, New York, December, 1980.

208 John Imbrie, et al, op cit, pp. 156-158.

209 ibid, p. 158.

210 ibid, p. 159.

211 John Imbrie, et al, op cit, p. 164.

212 ibid, pp. 158-159.

213 ibid, pp. 167-173.

213a Spectral analysis is a mathematical technique also known as Fourier analysis, which decomposes a signal into its frequency components. This technique is bread and butter to physicists and electrical engineers. It is used extensively in signal processing applications such as radar.

214 Bjorn Kurten, The Innocent Assassins, pp. 88-94, Columbia University Press, New York, 1991.

Index ยท French