The debate is just beginning — on the Cretaceous!

Will wonders never cease? Evidently not. Just when the hothouse starts looking really, really hot, along comes a new Science article by Bornemann et al, dealing with climatic conditions in the Turonian (93.5 to 89.3 million years ago). The principal result of this paper is that there appears to have been a 200,000 year period right smack in the middle of one of the warmest periods of the past half billion years, when there were ice sheets (presumably in Antarctica) that were up to 60% the volume of today’s Antarctic ice sheets. How in the world do you get such large ice sheets in a high CO2 climate warm enough for crocodiles to survive in the Arctic at the other side of the planet? And this apparent glaciation is not the result of a global cold snap. As in the Eocene results quoted earlier, the tropical ocean surface temperatures are again on the order of 35C — courtesy once more of the wondrous Tex86 proxy.

How was the ice volume inferred? Primarily by an especially meticulous application of an old technique. When a glacier forms on land, the water it is made of is enriched in the lighter form of oxygen, 16O, which leaves the ocean enriched in the heavy form, 18O. Single-celled shelly creatures called foraminifera (“forams” for short) record this composition, but they are very subject to diagenesis. The key to the new estimate was to take samples from pristine glassy portions of exceptionally well-preserved foram shells. The sample was taken from a core in the Tropical Western Atlantic, so the investigators are able to determine tropical surface temperatures, making use of Tex86 proxies from organisms living near the surface. The ocean water isotopic composition is estimated using both surface-dwelling and deep-dwelling forams.

Since the oxygen isotope composition of forams depends on temperature as well as ocean water composition, the Tex86 proxy was used to correct for the temperature effect in forams living near the surface. There is no independent temperature proxy for the deep ocean, but the investigators assumed (a bit questionably) that deep ocean temperature did not change much over the time period. Be that as it may, the deep ocean oxygen isotope shift (uncorrected for any temperature effect) was similar in magnitude to the estimate from surface forams. Once you have the oxygen isotopic composition of sea water, you can translate that into ice volume by making an estimate of the isotopic composition of glacier ice. All this is easier said than done, but they did it. The glacial interval corresponds to the excursion of delta-18O toward positive values in the figure below, taken from the paper:

There is a useful commentary by Richard Kerr One must exercise the usual caution we urge in connection with radical results, and await confirmation before taking it to the bank. As Kerr points out, there is other data from this time period that doesn’t show the isotope shift.

There are two additional things I myself noticed, which seem inconsistent. The first is that in order to get reasonable numbers for ice volume, the investigators needed to assume that Antarctic ice in the Cretaceous period had the same isotopic lightness as Antarctic ice today. Most theories of fractionation would have Antarctic ice being less fractionated in a warm climate, however. Perhaps the high Equator to Antarctic gradient helps keep the Antarctic ice light, but this is something that needs to be checked. What’s even more troubling to me is that the bottom-dwelling forams (uncorrected for temperature) indicate the same ocean water isotopic shift as the temperature-corrected surface dwelling forams. However, if Antarctica glaciates, the deep ocean should be filled with cold Antarctic bottom water, which should produce an additional positive isotopic shift in the uncorrected bottom dwelling forams. That this shift isn’t seen suggests that something is amiss to me.

Still, this paper already has a lot of modelers scratching their heads. To give an example of the magnitude of the problem, I reproduce below a figure from one of Rob DeConto’s old simulations (Nature 421, (2003) ), showing the glacier distribution in Antarctica as a function of CO2, as CO2 is steadily decreased. These are done for orbital parameters favorable to Antarctic glaciation; the simulations don’t use Late Cretaceous geography, but they do give a good idea of how hard it is to get a big glacier in Antarctica with anything much above twice the preindustrial CO2.

Page 3 of 4 | Previous page | Next page