Clive Best

Dear Clyde. Thanks for your proposal. I made this into a guest Post to allow for a wider discussion, hope you agree – Clive Best.

I am a long-term admirer and reader of your website. I have degrees in physics — a bachelor’s from Carnegie Institute of Technology and a master’s from University ofFlorida. I worked 40 years as an aerospace computer programmer and systems engineer with General Electric and Lockheed Martin and now am retired.
A statement of yours about ice ages inspired me: “We need a model to determine which picture fits best!”
My humble effort at such a model follows.
Clyde Carroll

Contents:
Chapter 1 The recent one million years
• Why do ice contractions stop, and expansions start?
• Why do expansions stop, and contractions start?
• Why are expansions slow and contractions fast?
• What determines the amount of ice in a cycle?
• What determines the duration (period), of cycles?
Chapter 2 Between one million and five million years ago
• Why do cycles average 41 thousand years?
• Why is there no evidence for the sawtooth (slow forming fast melting) shape?
• Why is ice amplitude early-on smaller?
• What caused the character of cycles to change at the Threshold?
Chapter 3 The Future
• How do human influences measure up to natural influences?
• A few more things.

Chapter 1. The recent one million years

A perhaps new idea is that in the recent one million years or so, ice ages are comprised of two ways that alternately dominate dispersal of sun supplied equatorial heat. In a swing toward less polar ice, the earth’s apparent present phase, globe girding ocean currents circulate sun warmed surface waters from tropical latitudes to near earth’s poles. Ocean waters cool as these currents melt polar ice. During a swing toward ice expansion tropical waters largely cool by evaporation, “sweating.” Humid warm air rises and carries to far north and south latitudes where heat radiates into space. Atmospheric moisture cools and drops substantially as snow. Each of these ways of heat dispersal leads naturally to the other, creating a back-and-forth repetition of polar ice cycles. Call it the Alternating Conveyance of Heat, or ACH, model.

Why do contractions stop?

In author’s conjecture the depletion of ice susceptible to ocean current melting is a too-little-ice trigger that ends a polar ice contraction.
Geologic features provide historical evidence of ice contractions. Solidified layers of ancient sea floor mud show melting and glaciation caused seas in the four recent ice ages to rise and fall by about 122 m (400 ft)*4. Herein sea and ocean are synonymous. Noted scientist Steven E. Koonin writes that these ancient sea level estimates have a typical uncertainty of 10 m (33 ft)*5. As oceans recede glaciers grow over much of Canada, to as thick as 3.2 kilometers (2 miles)*6. Ocean and other waters that end up on land as glaciers, the ACH model assumes, get there as usual namely by evaporations in warmer locations flowing through the atmosphere until they precipitate largely as snow. Floating ice also grows and shrinks during an ice age, but unlike the land based floating ice following Archimedes’ principle does not significantly affect ocean levels. Stacked-up ice layers from annual snows on the land of Antarctica and Greenland are a some 800-thousand-year record*7. Like rings of a tree trunk ice layer numbers and composition reveal much about the date and environment when formed. Ice ages are readily apparent, with layers being closer together at the warmer climate peaks. Counting annual ice layers determines the post Threshold irregular durations cited above. An analysis shows that the “most frequent” period of the completed recent eleven ice ages cited is about 82*8 thousand years.

Today’s sea level continues to rise*9. Today’s sea is approaching the high extreme of the preceding four ice ages and that extreme Koonin writes , is about 6 m (20 ft) above the current sea level*10. See Figure 1.

Figure 1*11 (see figure at reference 11)
Figure 1 is a 4-million-year graph of the “018 ice proxy.” An increase of glaciers, as measured by this earth glacier proxy, implies a lower sea level. Climate scientists Richard A Muller and Gordon J MacDonald write the 018 ice proxy is “…probably the most dependable measure of total ice volume.” *12 More on the 018-ice proxy is in Chapter 2.

A problem in considering “6 m (20 ft) above the current sea level” is the approximately 10 m (33 ft) of “typical uncertainty” that scientists assign to those estimates of ancient levels. In other words, within typical uncertainty, sea levels could be at a peak. Or they could rise as much as another 16 m (53 ft) and still not exceed the high extreme – including typical uncertainty – of the prior four ice ages.
Notably, oceans could rise even more, an estimated 70 m*13 (230 ft) from today if earth glaciers were all to melt. For reference, approximately 99%*14 of earth’s present glaciers are in the ice sheets of Greenland and Antarctica.
What is going on?

We know Greenland and Antarctic ice stopped melting near present levels by natural causes in the preceding four cycles. How do we know? If melting had not stopped much of the annual layers of polar ice that we now can observe would not exist. In other words a further significant rise – greater than about 6 m (20 ft) – of necessity would need the “water spigot” of ice-melts from the ancient glaciers that presently sit on Greenland and Antarctica. Why did glaciers in these recent cycles stop melting when oceans rose to near today’s levels? See Figure 1 again. Author’s conjecture is that melting stopped because ocean thermal currents that melted other glaciers could not reach them. More ocean thermal current particulars may elucidate.

Ocean thermal currents are key to melting polar ice. Think of the Gulf Stream. The GS is part of a giant system of ocean heat “conveyer belts.” The GS carries warm surface waters from the Caribbean Sea north, warming Europe, and veering toward Greenland where it melts ice it contacts. The GS further receives cold rivers of land-based glacier melts – melts partly caused by the weather warming proximate GS. The GS itself thereby cools and becomes denser. In the process GS waters sink, returning deep toward the tropics. Eventually, in a complex pattern of linked global thermal currents, deep cooled ocean waters resurface in tropical latitudes where they warm. Re-warmed surface waters again stream pole-ward, to melt more ice. This giant heat conveyer system is powered by the sun, that creates warm currents that polar ice melting causes to cool and sink. These sinking melts keep giving the thermal conveyer belts a push.
Ocean thermal currents approach near to a stop when, as at the present point in a glacial cycle, there is not enough available melting ice to keep powering their motion at the old rate. The massive Antarctica and Greenland glaciers can persist long after surrounding snow and ice have melted. They are like hills of snow on a cleared parking lot. Parking lot snow hills can last a relatively long time after being piled up and after all other nearby snow has melted. Being massive, on relatively high ground and mostly within cold-skimpy-sun arctic circles, Antarctica and Greenland glaciers have survived through several ice age cycles the relatively full melting that occurs with other glaciers. Confirmation of the end of an ice contraction is the end of rising oceans. As noted, today’s sea level is still rising, perhaps irregularly, and approaching the high extreme of the four preceding ice ages – which, you will recall, is about 6 m (20 ft) higher plus or minus an uncertainty of 10 m (33 ft.)

Why do expansions start?

The start of the next expansion is a too-little-ice trigger. It is an equilibrium point marking where glacier formation begins to overtake glacier melting. The ACH model posits this trigger will mark the start of the next ice age and the stopping of the present one.
With too little ice to melt ocean thermal currents approach a near stop.
Without adequate ocean current cooling, tropical ocean surface water temperatures increase. They increase due to a continuing and essentially constant rate of sunshine.
In ice cycles within the recent one million years when ocean thermal currents became substantially diminished what would take their place to achieve a balance in equatorial ocean near-surface temperatures? Author’s conjecture is that atmospheres thermal currents do that. Tropical waters heat up, and evaporating water increasingly lofts moisture upward. A higher equilibrium temperature of surface water is reached when evaporative cooling matches the sunlight’s heating. How much do tropical ocean surface temperatures rise? They were 2 to 3 degrees C (3.6 to 5.4 deg F)*16 warmer than today, at the estimated ocean near-surface temperature some 127 thousand years ago, near then cycle-extremes of high sea level and temperature. Today as in the past at similar cycle points the fluid atmosphere is also a global heat conveyer, though on a smaller scale than the ocean. For example, as in the ocean, atmospheric heat is transported via energized (warmed) molecules and atoms. With earth oceans weighing about 280*17 times the atmosphere*18, the oceans have a lot more similarly sized molecules and atoms to be enlisted in a heat conveyer. Acknowledging, though, the atmospheric conveyer makes up for some, but not all, of its handicap of being smaller than the ocean conveyer by its property of moving faster. Gaseous air generally conveys its warmed molecules substantially faster than do the ocean’s liquid waters. For example, the atmospheric jet stream is a belt of wind that blows from west to east at speeds up to about 400 kilometers per hour (250 miles per hour)*19. A Gulf Stream extreme speed typically is about 9 kilometers per hour (5.6 miles per hour)*20.

What is helpful in outlining a transition from mainly an ocean to mainly an atmospheric heat conveyer is some specifics on present atmospheric thermal currents. Notably the earth’s weather-atmosphere (also known as the troposphere) is a relatively thin layer, about 10 kilometers deep at the equator and 6 K at the poles (or 6.2 and 3.7 miles.) The troposphere contains about 80 percent of the atmosphere’s mass and is the effective limit to which significant convection, the atmospheric conveyer, will rise. The earth’s weather is essentially in this layer which is about .08% of earth’s diameter. It is a layer that proportionally is thinner by about five times the skin on an apple. The weather atmosphere consists of thermal circulation “cells.” Air currents in these cells carry heat away from equatorial regions. Meteorology shows there are presently three atmospheric circulation cells in each hemisphere. The cells in one hemisphere are largely symmetrical to those in the other.

In the northern hemisphere a main circulation cell encircles the earth and borders near the zero-degree-latitude equator. That cell’s outlines are those of a thin but gigantic girdle with belt near a third of the way to the north pole and lower hem near the equator. It is known as a Hadley cell, one of two. Rising, warm and humid tropical air flows north, nearer the top of this thin cell to approximately 30 degrees N latitude. The air is cooled by expansion and by radiating heat into space. The Hadley cell moisture largely precipitates by the time air reaches near thirty-degrees north. The dryer, cooled, denser air sinks. Due to a vacuum, caused by the Hadley’s tropical rising air, the falling air heads equatorward near earth’s surface and starts a new trip around the Hadley circuit. The sinking denser air near thirty-degrees north latitude, as well as the sun-driven rising moist warm air near the equator, power movement of the Hadley atmospheric heat conveyer.

There are two more atmospheric circulation cells in the northern hemisphere. These cells operate between approximately thirty and sixty-degrees and between sixty and ninety-degrees north latitudes. Like a river causing swirling eddy currents in adjacent pools, the higher latitude atmospheric cells are thermal eddy currents in the wake of the relatively strong air flows in a Hadley main cell. Symmetrically, the southern hemisphere’s three circulation cells operate between like south latitudes. A schematic sketch is at the web address reference.*21

In heating a house steam heat is a less energy efficient system than hot water. Similarly a vaporous atmospheric heat conveyer is less efficient than ocean currents. As the hefty “varsity” heat-conveyer oceans are taken out of the game – by lack of ice to readily melt – the less weighty atmospheric “junior varsity” begins to take over. Rising tropical ocean surface temperatures put the JV into the global-scale heat conveyer game.
At today’s point in the ice age cycle glaciers, perhaps irregularly, are still slowly shrinking. Today’s atmospheric circulations do not transport the moisture necessary to produce new snow in polar regions at a rate needed for significant glacier growth. So, an ice age scale of snows precipitating near the poles can be attributed to what mechanism? There are many ways that atmospheric thermal currents could reorganize themselves to transport poleward large additions of evaporating tropical moisture. But author’s conjecture is that the atmospheric reorganization may become two enlarged Hadley-like cells. The two Hadleys become greatly enlarged due to the increases in vaporous water being lofted upward by rising tropical sea temperatures. Like a swollen river overwhelms adjacent pools and eddy currents, growing Hadleys largely subsume the atmospheric cells nearer the poles. By the time polar ice is significantly expanding possibly one large Hadley-like atmospheric cell spans the northern hemisphere and another the southern. After they have expanded to near the arctic circles these enlarged Hardleys begin delivering to polar regions an ice age scale of precipitation and snow. But early, before reaching near arctic latitudes, increased Hadley moisture would likely manifest as greater amounts of rain and more flooding than usual at mid latitudes — above 30 degrees. As the boundary of a growing Hadley approaches a pole, increasingly precipitation in and near polar regions would be as snow. At its coldest, for example near the present ongoing cycle’s high extreme of glacial ice about 20 thousand years ago, earth’s average surface temperature was lower by approximately 5 deg C (9 deg F)*22. What causes this temperature drop?
First, as a basis to convince oneself of the validity of temperature estimates tens and hundreds of thousands of years ago, some details on the means of figuring these temperatures may help.

As a temperature proxy geoscientists measure the tiny proportions of heavy water in samples from arctic glacier annual ice layers. Water is partly made up of rare heavy H20 molecules that are heavy because they include a rare overweight H, hydrogen, atom – called an isotope. Because of their extra weight heavier than usual H2O molecules are slightly less likely than regular H2O to be in glacier ice. Why? As tropical waters sweat it is the lighter molecules that are more likely to evaporate. As this atmospheric moisture proceeds poleward it further decreases concentrations of heavy water due to the few heavier molecules remaining being the ones more likely to precipitate along the way. The decrease of heavy molecules in evaporated and precipitated H2O is temperature dependent. This temperature dependence allows scientists to analyze the rare heavy water concentrations in arctic annual ice layers and estimate the average atmospheric temperature when a layer was formed. Analysis of other temperature proxies, lend confidence to average temperature estimates, such as the approximate 5 deg C (9 degrees F) atmospheric temperature difference between the high and low-extremes of temperature in the four ice ages preceding the present ongoing one. Graphs of changes in estimated near-earth-surface temperature averages, within and spanning several ice age cycles, are readily available*23.

The cause of the average atmospheric temperature drop during ice expansions in author’s conjecture is that clouds do that.
Sunlight reflected by near hemisphere-spanning Hadley cell clouds accomplishes the bulk of temperature drops. Rivers of moisture-laden air, moving pole-ward at altitude, become blankets of warm clouds that increasingly cover earth’s latitudes approaching arctic circles. Growing cloud blankets would perform two functions. 1) They form a giant near-white shade, reflecting large amounts of the sun’s light that would otherwise warm the surface below. 2) They are a gigantic heater, radiating heat into space – much like a home’s old fashion steam heating system radiates heat into mostly empty rooms. Because of 1) the earth surface becomes colder. Because of 2) the cloud blanket’s moisture cools, condenses and falls largely as snow at latitudes approaching the poles. In these latitudes snow accumulates. For example, during a time about 127 thousand years ago in this most recent cycle earth’s atmospheric temperature was near a nadir and glacier levels were near a peak. Flowing back toward the equator are cold dry polar winds, near earth’s surface on the Hadley-like heat conveyer.

Why do expansions stop?

In author’s conjecture what stops an expansion is the too-much-ice trigger. The advancing glaciers intrude significantly into the edges of receded oceans, restarting the ocean heat conveyers.
Recall that as polar ice accumulates ocean levels drop by approximately 122 m (400 ft). Much “new land” is created by receding shorelines. Growth of glaciers within a cycle thus uncovers additional land on which more glaciers may sit. How much more land is made available? Examination of present ocean depth contour maps shows that with 122 m (400 ft) of drop in ocean level, shorelines would recede, and land would be exposed including at north polar icecap latitudes. At these latitudes much of North American, European, and Asian continental shelves, and the Arctic Ocean, North Sea, and Bearing Sea would become new land during recent cycles of ice expansion. By inspection of present-day ocean depth contour maps, that increase in the amount of northern hemispheric land available for snow and glacier accumulation is on the order of the present area of Greenland*24. In the recent cycles the lands uncovered by receding ocean levels in ice expansion phases allowed northern hemisphere snowpacks to continue to grow and cover them. In a race, the advancing glaciers and receding oceans finally significantly intersect. By contrast, near Antarctica shallow waters also become exposed land but to a much lesser extent, simply because oceans neighboring Antarctica have fewer shallow near-shore areas. Opportunity for Arctic glacier growth is on a much larger scale than for the Antarctic.
Especially in the northern hemisphere, advancing glaciers begin to encroach on mid latitude oceans. The approximate North American latitude limit for advancing ice in the most recent ongoing cycle is marked by Long Island on the east and Vancouver Island on the west. These islands are the result of massive land erosion. They are the moraines, the debris bulldozed up by advancing arctic glaciers as they pushed equatorward to their farthest points. Glaciers finally run out of arctic and near-arctic land on which to expand, and they increasingly protrude beyond ocean shores, for example, as ice shelves partially sitting on ocean bottom near a cycle’s ocean-low-water line. This is the too-much-ice trigger that ends an ice expansion and begins its contraction.

Why do contractions start?

The too-much-ice trigger also starts the next contraction. The ocean heat conveyer, when it is triggered back to life after advancing glaciers encroach significantly into the edges of receded oceans, begins again significantly melting polar ice.
In the process, ocean levels start again to rise. As oceans rise, ocean currents encounter more and more ice to melt. Relatively warm ocean waters began more significantly to melt proximate polar ice. Once it begins regrowing, the heavy weight “varsity” ocean current heat conveyer outperforms the “junior varsity” atmospheric conveyer — and cools tropical waters faster than the sun warms them. Tropical ocean temperatures drop below levels needed to maintain production of rising H2O vapors that power the atmospheric conveyer. The JV atmospheric heat conveyer begins to leave the field, being replaced by the varsity heavy weight ocean conveyer. The hemispheric sized Hadley-like cells begin to shrink into their weaker and smaller configurations. Presently ice susceptible to melting by ocean thermal currents is close to being depleted. It is near to becoming too-little-ice and triggering the beginning of the next ice age expansion. In the past one million years of cycles what has happened about eleven times is nearly ready to repeat. In view of glacier ice graphs such as Figure 1 and others extending for many more past ice ages, scientists Muller and MacDonald write: “Data such as these are what led us to state … that the next ice age is about to hit us, any-millennium now.*25 ”

Why are expansions slow and contractions fast?
Considering the ACH model, the earth itself answers the slow-fast question. Expansions are slow and contractions fast because ocean heat conveyers are more efficient than the atmospheric conveyers. Relative efficiency of the two conveyers is confirmed in earth’s polar ice history. In Figure 1, for example, in the most recent almost-cycle it has taken about 100 thousand years to build ice that since then is taking only some 20 thousand years to largely melt. In all but one of the Figure 1 cycles this slow ice building and fast ice retreat is readily visible. In authors conjecture the one exception where build and retreat times appear around the same is a measurement error. In more recently published climate data this cycle is also a “sawtooth” shape — where sawtooth is the descriptive term for slow ice building and fast retreat*26.
If the atmospheric heat conveyer were the more efficient the ocean conveyer could not take over for a period within a cycle. The atmospheric conveyer would keep evaporating moisture until it could not, turning earth into a near-ice-ball. Thankfully each ice age advance in Figure 1, and before, has reversed.

What determines the amount of ice in a cycle?

Author’s conjecture is that the amount of ice in a cycle primarily depends on the amount of land available on which snow can build glaciers. And the more suitable land the more ice. What is the basis for this?
For a mechanism to account for ice accumulation consider only the ACH model’s outlined “four elephants in the climate room:” 1) the atmospheric heat conveyer that creates polar ice, 2) the ocean heat conveyer that melts polar ice, 3) sun’s radiant energy that powers the conveyers, and 4) the land – including land herein called ice-warehouse-land in and near arctic circles.
Of the ACH model’s four major elements — atmosphere, oceans, sun and land — three are relatively constant over a five-million-year span. You will recall during the recent five million years the sun’s radiated energy has increased about 0.3 percent. Author judges that a change of 0.3 percent is not significant to sun’s function of supplying energy to the ACH model. Also assumed to be relatively constant are the model’s two heat conveyers. From Koonin, “The amount of water on the earth is essentially fixed.”*27 Generally, mass does not readily appear or disappear. Each conveyer’s mass – largely the ocean’s sum of water molecules and the air’s sum of 78 percent oxygen and 21 percent nitrogen – likely does not undergo significant change during five-million-years. True for example, near the present ongoing cycle’s high-ice extreme about 3 percent*28 of ocean water otherwise available to its conveyer function is missing and stored in glaciers. Author judges that a change of 3 percent is not significant to the ocean’s conveyer function. Other possible means for additions or subtractions to the conductor fluids are by chemical or nuclear reactions. But such reactions, on a scale to significantly influence the functioning of the conveyers, would likely leave behind geologic evidence for example in a sequence of changes to the distribution of earth’s related elements and compounds. That such large-scale reactions are happening at present or in the past fifty million years is not apparent from research by author. For example CO2 concentration in the atmosphere is about 420 parts-per-million.*29 But CO2 now makes up a still small part, approximately 0.04 percent, of the atmosphere. That leaves the fourth major element in the ACH model, available ice-warehouse-land, as likely the primary influence for how much glacial ice accumulates in a cycle. Occam’s Razor principle: “The simplest answer is often the best.”

What determines the period of cycles?
In the recent one million years author’s conjecture is that ice cycles are those of a “relaxation oscillator.” Why bring this up? In a relaxation oscillator a cycle’s period is mainly determined by the “time constant”, or time between like equilibrium points — between two successive high, or two successive low, ice extremes. As cited above, after Javier Vinos, the “most frequent” period between ice peaks in the eleven ice cycles since the Threshold is about 82 thousand years. Thus the ACH model’s period is about 82 thousand years.
Elaboration on relaxation oscillations may be helpful. After Wikipedia:
“Relaxation oscillations are characterized by two alternating processes on different time scales: a long relaxation period during which the system approaches an equilibrium point, alternating with a short impulsive period in which the equilibrium shifts. The period of a relaxation oscillator is mainly determined by the time constant*30.”
Tailoring the relaxation oscillations characterization to the ACH model: “Since the Threshold, glacial ice oscillations are characterized by two alternating processes on different time scales: a long relaxation period during which ice production approaches an equilibrium point (too-much-ice) alternating with a relatively short impulsive period in which the equilibrium shifts (to too-little-ice.) The period of the relaxation oscillator is mainly determined by the time constant – namely, time between successive like-equilibriums.”

As above, that period of post Threshold cycles is about 82 thousand years.

Chapter 1 References
1. Reserved
2. Reserved
3. Reserved
4. Koonin, op cit. page 151
5. Koonin, op cit. page 150
6. https://en.wikipedia.org/wiki/Laurentide_ice_sheet#:~:text=It%20was%20up%20to%203%C2%A0km%20(2%C2%A0mi)%20thick%20in%20Nunavik%2C%20Quebec%2C%20Canada%2C
7. See graph’s blue line of data.
8. Javier Vinos, op cit. Page 11, Table 2.1
9. Koonin, op cit. page 151, Figure 8.2
10. Koonin, op cit. page 150, Figure 8.1
11.
12. Richard A Muller and Gorden J. MacDonald, “Ice Ages and Astronomical Causes” page 108, Springer-Praxis Publishing, Chichester, UK
13. https://www.usgs.gov/water-science-school/science/glaciers-and-icecaps Go to this address then hit page down (pg dn) 4 times to “Some glacier and icecap facts”
14. https://www.antarcticglaciers.org/glaciers-and-climate/what-is-the-global-volume-of-land-ice-and-how-is-it-changing/
Go to this address then hit page down (pg dn) 16 times to “Summary”
15. https://en.wikipedia.org/wiki/Solar_luminosity
16. Koonin, Steven E op. cit. Figure 1.8, page 43
17. https://en.wikipedia.org/wiki/Ocean Find “The total mass of Earth’s Hydrosphere…”
18. https://en.wikipedia.org/wiki/Atmosphere_of_Earth Find “The atmosphere has a mass”
19. https://en.wikipedia.org/wiki/Jet_stream Find “250 mph”
20. https://oceanservice.noaa.gov/facts/gulfstreamspeed.html find “maximum speed”

21. ?1376037245
22. Koonin, Steven E. op. cit. page 43
23.
24.
25. Richard A Muller et al, op. cit. page 5
26. Richard A Muller et al, op. cit. page 6
27. Koonin, Steven E. op. cit. page 130
28. 3%=122m (sea level drop if all glaciers melted) divided by present 3,962m avg ocean depth.
29. .webp
30. https://en.wikipedia.org/wiki/Relaxation_oscillator – note: Referenced quote not found at this address as of 3/30/2026/
To the extent that any of the above may be subject to copyright, author relies on the Fair Use Exception of 17 U.S.C. paragraph 107.