A recent study of climate models concludes that they are badly distorted. Which is pretty clear to everyone who’s not inside the climate echo chamber, which turns out to be a crowded spot. Linnea Lueken notes that the alarmist outlet Inside Climate News actually believes in the crop failures taking place inside a computer despite the bountiful harvests outside. But taking the longer perspective, Mike Jonas observes on Watts Up With That, “I find it difficult to believe that any rational person could still believe that climate models can work, given the IPCC’s statement that “long-term prediction of future climate states is not possible” because climate is non-linear (“chaotic” in the technical mathematical sense), so no matter how sophisticated the software or powerful the hardware, such processes cannot be reduced to equations that given approximately accurate inputs will create approximately accurate forecasts. And he cites as an instance of that limitation, their incapacity to solve the so-called 100,000 year problem. As he explains in his paper “The ‘100,000-year problem’ refers to an apparent unexplained change in the frequency of inter-glacial periods which occurred about a million years ago. Before that, inter-glacial periods seemed to occur about every 41,000 years, in line with the obliquity Milankovich cycle. But after that, they seemed to occur about every 100,000 years, in line with the orbital inclination Milankovich cycle.” But in his view what’s really going on is a 21,000-year cycle that sometimes skips a beat because climate is non-linear. And because climate is non-linear, no computer can say why.

The problem of mathematical chaos theory is more serious for climate studies than virtually anyone involved in the debate seems to believe. Thus a group of scientists claim to have overcome it and to be able to control the weather… inside their computer. Which to us sounds like a perpetual motion machine except less probable. And we’d like to see computers predict the known past before they tackle the unknown future instead of the other way round.

Which brings us back to Jonas’s paper in particular. CO2-obsessed computer models have surprisingly little to say about paleoclimate, including why the Earth hadn’t seen an ice age since before the dinosaurs and then the Pleistocene started 2.6 million years ago. But if you challenged an alarmist to explain why, since then, the Earth has seen alternating glaciations on a 100,000 year cycle for the last million years, but every 41,000 or so before Victor Mature made that absurd film, you’d reliably get a pat on the head and say it’s all well-known Milankovitch Cycles, regular wobbles in Earth’s rotation, or more exactly in the eccentricity, axial tilt and precession of its rotations around its axis and the sun, that affect how much incoming radiation hits the land-intensive Northern Hemisphere rather than the watery Southern. Evidently they drove temperature until 1970, after which man-made CO2 took over and messed the place up. And the only issue in this glib response was why 41k yielded suddenly to 100k. Weird, huh?

Jonas says a lot of cleverness was expended to account for that transition. But actually it never happened. There is no 100k cycle. And no 41k cycle either. He says it’s a 21,000 year “precession” cycle. In which case Milankovitch was, basically, right about wobbles. But here’s the problem. Based on Antarctic temperature reconstructions, the 21k cycles come along cyclically. And every single interglacial corresponds with one. But the reverse is not true. And it’s not that the 21k cycle isn’t there. Rather, the reconstructions show, “some missing cycles did have a noticeable effect, but they failed to start an inter-glacial period.”

Now it bears repeating here that if you want to determine precisely how many interglacial periods the planet has seen since Megalodon bit the seabed, cooling being bad for life, you’re working with indirect proxies whose testimony is often unclear. And especially if you got an interglacial that wasn’t, like the Eemian, recent and dramatic, considerably warmer than the current Holocene “hottest ever man-made there go the corals and polar bears aaaaaaaah!”, the traces will be especially faint and ambiguous.

In Jonas’ words, “Clearly, the sequence of number of missed cycles is not precise, because the data is so open to interpretation, but the data clearly shows a pattern of missed cycles over time, with a generally greater frequency of missed cycles within the last 900k years.” Which he thinks might be due to the fact that, he says, “Earth's temperature has gradually declined over the last 5 million years” to which we reply “And good luck getting the computer to tell you why it happened, especially if you insist in advance that it blame CO2”.

Or to tell you why anything happened. Even if it can tell you with apodictic certainty that the crops will fail and taste bad, Miami Beach will become Miami Seabed and poison ivy will grow legs and a long poisonous tongue and eat us all up if RCP8.5 comes true.