August 7, 2009 | David F. Coppedge

Paper View:  A Geology Paradigm Suffers a Paradox

A pair of geologists found a paradox in a paradigm.  That paradigm is the belief that ancient ocean levels rose and fell in cycles as ice sheets retreated and advanced, and the cause of the cycles was periodic changes in earth’s orbit.  They modeled this process and couldn’t get it to work.  They couldn’t get the sea level to rise as high as the evidence shows it did.  And it would have required enormously high fluctuations in atmospheric carbon dioxide to get rid of the ice sheets each time.

Horton and Poulsen [U of Michigan, Ann Arbor], writing in Geology,1 called their paper the “Paradox of late Paleozoic glacioeustasy.”  Time out for definitions: eustasy refers to global sea levels as measured from a static reference (like the center of the earth); glacioeustasy is, by extension, the global sea level during ice ages.  In theory, the earth’s cyclic orbital variations over long time scales2 force ice ages to advance and retreat.  In the Paleozoic, when the land terrain was generally lower, large ice sheets could form and melt over vast areas, producing cyclothems, or sedimentary deposits alternating between strata buried under water (subaqueous) and under air (subaerial).  Sounds plausible, but how accurate is this picture scientifically?  How well is it understood?  After all, Horton and Poulsen began by saying that this paradigm has been used to explain our planet’s past: “Models of Euramerican cyclothem deposition invoke orbitally driven glacioeustasy to explain widespread cyclic marine and nonmarine late Paleozoic sedimentary sequences.”  Presumably, a geologist doesn’t just want to catalog the rocks, but explain how they got that way.

Things get messy real quick when fitting the field data into the paradigm.  It appears global sea levels would have had to alternate by 100 meters or more to produce the beds.  Their model, however, couldn’t get the water to rise more than 25m.  And only by adjusting the model with enormous fluctuations in atmospheric carbon dioxide could they get the ice sheets to melt away.  “These results present a potential paradox: while our model is able to simulate widespread Gondwanan glaciation, it is unable to reproduce significant orbitally driven glacioeustatic fluctuations without very large magnitude carbon cycle perturbations.”

Their Introduction describes the paradigm: from 326 to 267 million years ago (mya), conditions “primed the late Paleozoic paleoenvironment for glaciation.”3  The continents were joined into a supercontinent called Gondwana, and carbon dioxide levels were low.  Ice sheets formed and grew when orbital changes cooled the atmosphere; they retreated when warm times returned.  “The ice sheets of Gondwana left not only direct geological evidence of continental glaciation, but also indirect sedimentary signatures of their waxing and waning,” they said.  The presence of North American deposits that appear cyclic “has been used to infer that late Paleozoic depositional environments were largely controlled by glacioeustasy.”  If that is the explanation, though, the fluctuations in global sea level were very large: up to 200m – much larger than the inches or feet that alarmists of global warming warn us about today.  Can sea level changes of that magnitude be forced by orbital cycles?  That’s what Horton and Poulsen wanted to find out.  Remember, the plausibility of a hypothesis has nothing to do with its credibility, and vice versa.  Science wants to run the numbers.

The next section of the paper described the methods and results of their model.  A global model, of course, requires dependence on other models, like “at atmospheric general circulation model (GCM) … coupled to a three-dimensional ice sheet model.”  They also had to rely on models of topographic elevations of a continent – Gondwana – that no longer exists.4  And, they played with carbon dioxide levels “consistent with proxy estimates.”  The growth and extent of the resulting ice sheets reflected the level of atmospheric CO2 as expected, but they couldn’t get it to melt except at the higher concentrations.  They said “increases in excess of 2000 ppm were required to cause substantial melting of Gondwanan ice sheets.”  (Note: current levels are about 384 ppm amidst all the hubbub about global warming.)  In addition, the causative factors seemed inadequate.  “The dynamic response of continental ice sheet volume to our prescribed transient orbital insolation variations is modest,” they said.  That means that orbital forcing does not have that great an effect.  It certainly did not produce global sea level rises of 100 meters or more.  And it contradicts other studies that claim to find correlations between recent ice ages and orbital periods, like a recent paper in Science.5  (See summary on Science Daily).

In their discussion and conclusions, they evaluated possible resolutions of the paradox.  Perhaps “Orbital changes were linked to the carbon cycle through a positive feedback” like they believe was operative during the greener Pleistocene epoch.  Even so, the model required extreme concentrations of carbon dioxide to get the ice sheets to retreat completely between cycles (ablation).  But what in nature would generate such high concentrations in the Paleozoic?

Our simulation of late Paleozoic glacial conditions presents a paradox.  While our simulation of large (>100 m sea-level equivalent) continental ice sheets is in good agreement with sedimentological evidence of Gondwanan glaciation, our orbitally driven ice-volume changes are ~10 m, much smaller than the late Paleozoic glacioeustatic variations implied by both cyclothems and isotopic analyses.  The absence of significant continental-scale ice sheet ablation in the face of changing orbital insolation poses a significant challenge to our current understanding of late Paleozoic ice sheet dynamics.

More bad news.  What does this do to the more recent Pleistocene glacioeustasy theory?  It was supposed to be better understood because ice core data admittedly shows some agreement with orbital and atmospheric factors.6  But even here, “The cause of Pleistocene glacial-interglacial cycles is still debated,” they said, “but is generally thought to be due to a combination of orbitally controlled insolation forcing and greenhouse gas fluctuations.”  But those are the same factors used for the Paleozoic model.  They didn’t work there.  How do we know the same factors would produce Pleistocene cycles?

To save the Paleozoic paradigm, they appealed to ignorance.  “Unlike the Pleistocene, late Paleozoic pCO2 levels are not well constrained.”  Maybe somehow the levels fluctuated wildly back then; “however, the temporal resolution of this record remains coarse,” so there is no way to know with any accuracy.  “Furthermore, a mechanism by which atmospheric pCO2 concentrations could repeatedly fluctuate by ~2000 ppm over orbital time scales is not known.”  Methane clathrates under the sea could discharge a great deal of greenhouse gas quickly; “However, the long recharge rate of clathrates would prevent repeated discharges on orbital time scales.”  So that idea is no help.  Is it possible that the sea levels rose and fell independent of carbon dioxide concentrations?  An “energy balance model” (EBM) reproduced the cycles, they noted.  But then they discounted it:

Simulations using an energy balance model (EBM) coupled to an ice sheet model indicate that orbital insolation variations alone can produce repeated ~100 m sea-level fluctuations (Hyde et al., 1999).  We cannot say with certainty why our results differ from those using an EBM; however, we suspect that differences in the paleoboundary conditions and/or the treatment of ablation and precipitation rates in the calculation of mass balance over the ice sheet might be responsible.  For example, unlike our model where precipitation over Gondwana is explicitly calculated, EBM precipitation rates are based on prescribed modern precipitation rates (Hyde et al., 1999). Predictions of equilibrium ice sheets made using GCM ice sheet models with fixed (nontransient) orbital conditions have also been used to infer large late Paleozoic glacioeustatic fluctuations (of as much as 245 m; Horton et al., 2007).  However, our new results indicate that these estimates are too large.  The reason is straightforward: in the fixed-orbit experiments, there is no preexisting ice sheet to influence the final mass balance.  In contrast, in our transient experiments, the preexisting ice sheet (simulated during the previous orbital step) has a substantial influence on local conditions due to temperature-elevation and ice-albedo feedbacks.  Orbitally driven insolation changes are not large enough to overcome these local ice sheet effects; consequently, orbital changes produce only small ice-volume fluctuations.

Time to assess the situation.  Horton and Poulsen believe they experimented with a reasonable model, but they could not replicate the paradigm.  This may “have potential implications for the late Paleozoic climate system and for ice sheet dynamics in general.”  Maybe the model is the problem.  Any model has limitations; it is only a simulation of a process in which multiple factors interact in complex ways.  Still, if there were an orbital signature, it should appear in the simulation.  Their main obstacle was getting rid of the ice sheets once they formed (in some runs, the ice sheets reached the latitudes of modern-day Buenos Aires).  The Pleistocene paradigm invokes additional processes to get rid of the ice quickly between cycles: including “subglacial sediment destabilization (MacAyeal, 1993; Clark and Pollard, 1998), reorganization of the ocean’s thermohaline circulation (Maslin et al., 2001), and the removal of coastal sea-ice buttressing (Rignot and Thomas, 2002).”  Future runs including those factors may have better luck.  But there’s a disturbing alternative interpretation: “Alternatively, our lack of large glacioeustatic change could also indicate that the Pleistocene Northern Hemisphere glacial-interglacial cycles may not be a good analogue for late Paleozoic glaciation.”  The two epochs are not comparable, in other words.  This has a more disturbing side effect: “in which case non-uniformitarian processes (e.g., very large perturbations of the carbon cycle) may have driven late Paleozoic glacioeustatic fluctuations.”  To rework a phrase: the (almost-present) Pleistocene was not the key to the past Paleozoic.

Here’s the upshot: they found a paradox, and couldn’t resolve it.  This undercuts a paradigm that was thought to be fairly well understood.  They only explanation left was non-uniformitarian – something that runs against the grain of the whole science of geology.  It reduces to the Stuff Happens Law (09/15/2008).  The consequences are enormous, but the solution lies in the nebulous future: “The resolution of this late Paleozoic paradox is fundamental for understanding the processes that drive glacioeustatic cyclicity and late Paleozoic climate change (Poulsen et al., 2007; Peyser and Poulsen, 2008), and is relevant to our current understanding of the climate-cryosphere system.

This can only imply one thing: jargon and math and computer skills notwithstanding, geologists, planetary scientists and atmospheric scientists understand very little of this at all.  Politicians should take note.


1.  Daniel E. Horton and Christopher J. Poulsen, “Paradox of late Paleozoic glacioeustasy,” Geology, August 2009, v. 37 no. 8 p. 715-718, doi: 10.1130/G30016A.1.
2.  These include rates of precession, obliquity, and eccentricity that can combine into long-term cyclic fluctuations.  Presumably, this affects insolation (the amount and angle of sunlight hitting the land masses) and therefore the climate, although other non-orbital factors could be involved (e.g., volcanic outgassing).
3.  These glacial cycles are different from those in the Pleistocene when mountains were much higher.
4.  Skeptics might ask if this hypothetical continent really did exist, if its existence is model-dependent on the interpretation of the very cyclic deposits examined in this paper.
5.  Clark et al, “The Last Glacial Maximum,” Science, 7 August 2009: Vol. 325. no. 5941, pp. 710-714, DOI: 10.1126/science.1172873. 6.  See 02/05/2008 and 08/08/2006 about whether orbital forcing mechanisms match the geological record.

Time for a refresher course on our Guide to Evolution (e.g., a list of Murphy’s Laws).  The First Law of Scientific Progress and Maier’s Law are, as you can see, not just jokes, but standard ways of doing business in the halls of academia.  Getting grants, writing computer models, and publishing them in scientific journals is all fine and good, but when the ultimate explanation is Stuff Happens, are we any better off?  When the authors throw up their hands and task future researchers with the obligation to figure this out, has our understanding of Planet Earth increased?  When the answer is Stuff Happens, and we don’t know why it happens, what is the probability the answer is outside the box?

Unbeknown to these two well-meaning modelers, they just handed a gift to the young-earth creationists (may their memory be forever Expelled – obligatory curse).  They admitted that uniformitarianism might be inadequate.  And they revealed that no known combination of present processes can account for the field data.  Maybe it’s time to resurrect the forbidden theory of Directed Catastrophism.

Categories: Geology, Physics

Leave a Reply