Mistakes made by Trenberth (and many others) when claiming veracity for climate models

Six months ago I posted, “Sensational statements by Dr Kevin Trenberth, IPCC supporter”, June 23rd, 2007 by Warwick Hughes.

Pat Frank says:
December 18th, 2007 at 3:40 pm

The central point Trenberth makes about GCM predictive reliability is this one, “The current projection method works to the extent it does because it utilizes differences from one time to another and the main model bias and systematic errors are thereby subtracted out. This assumes linearity.”

What he’s saying is that one can reliably project future forced global temperature differences because in subtracting forced GCM runs from control GCM runs all the errors subtract away, leaving a reliable trend in anomalies.

But this is assuming much more than mere “linearity.” It assumes the GCMs procduce global climate projections with completely accurate slopes. Trenberth is claiming that GCM runs are only linearly offset by some standard vertical magnitude from being fully correct — and the same standard vertical magnitude is present in both control runs and forcing runs. Therefore, subtracting the latter from the former produces accurate anomalies.

But this is very different from assuming mere linearity. It supposes that the physical representation of the global climate itself — the physical theory in the GCMs — is complete and accurate. The errors are merely from imperfect measurements and a too-coarse resolution because of computer limitations. This assumption — assertion, really — is entirely unjustifiable.

For example, all GCMs include a hyperviscosity because the Navier-Stokes equations can’t be solved at all the necessary levels of resolution. The hyperviscosity, which is completely unphysical, is the only thing that makes the GCMs integrable — they would catastrophically diverge otherwise. Because there is an unphysical hyperviscosity, the parameterizations in GCMs must also be unphysical in order to compensate. Consequently, GCMs inherently cannot be physically correct. Trenberth’s claim includes an implicit but absolutely central assertion that cannot be true. The physical theory in GCMs is neither correct nor complete. Trenberth is wrong, the reliability of temperature predictions can not be claimed accurate through taking differences, and the whole of AGW so-called science is powered by this sort of tendentious delusionalism.

9 comments to Mistakes made by Trenberth (and many others) when claiming veracity for climate models

  • all GCMs include a hyperviscosity

    Have they studied Physics or are they only geeks with very beautiful toys?

  • Pat Frank

    Warwick, Skeptic magazine is going to publish my manuscript assessing the reliability of GCM projections of future global average temperatures in their next issue — I think January or February 2008. In that article, I show that the uncertainty in future global temperatures has been vastly under-reported by the IPCC. In fact, the usual SRES scenario plots of CO2-induced temperature don’t show a physical undertainty at all. It turns out that the reliability of GCM projected future temperatures is zero. And not just for 100-year projections, but for 1-year projections.

  • Matt

    Lots and lots of anti-GCM model papers coming out recently.

    Here’s one: www.uah.edu/News/pdf/climatemodel.pdf

  • Michael

    So the argument is that because the GCMs are not exactly like the real world they cannot be true?

    So exactly how different from the real world are these GCMs, and is this difference enough to matter?

  • Matt

    Did you even click the paper I poseted? They aren;t even remotely close. If they cannot predict what we know has ALREADY HAPPENED, how in the f@#k can they predict anything?

  • SteveSadlov

    The GCMs are to the understanding of climate as neolithic people scratching pictures into the mud with sticks were to the understanding of astronomy. Those who are current in the areas of space weather, cloud physics and other similar highly advanced and complex fields will know that I am not in any way overstating my case.

  • SteveSadlov

    RE: #3 – Contrast this paper with the general meme embraced and propagated by Iassac Held at NOAA. A keystone in the alarmist edifice is the notion that 2 x CO2 leads to a specific manifestation of energy imbalance in the tropics that is “solved” by:
    1) A higher tropopause
    2) More expansive Hadley Cells and Subtropical Highs.
    3) Poleward movement of the general climate bands

    The problems with this are many. I will not at this time argue whether or not 2 x CO2 increases the vertical temperature gradient at the equator in a transient manner. Let’s assume for the sake of argument it does (momentarily forgetting about how the ITCZ and clouds actually behave). A larger thermal gradient will mean more instances of tropopause breaching. That is not accounted for in the models. Also, it is assumed that a higher tropopause automatically means a more expansive Hadley Cell. That would reaquire that the Ferrel Cells actually yield. Do they? Why would they? Why wouldn’t Subtropical Highs simply become more energetic without actually expanding?

    This is only scratching the surface. The assumption of a “moveable” tropopause is driven by the notion that the lapse rate at a given latitude is stable and static. This we know is not true, due to both lower atmosphere inversions and convective transport and mass transport by water. Space weather is another unknown. Given what we are learning about electrical interactions between the ionosphere and clouds, there may be boundary conditions in the Ionosphere and higher, that serve to constrain the atmophere in ways we do not understand. The whole energetics picture is only barely starting to be understood.

  • Reid of America

    Trenberth speaks with a forked tongue. Check out the Climate Science post on Trenberth.


    Sometimes he believes his models and sometimes he doesn’t.

  • Brooks Hurd

    Anyone who believes that GCM models have some reasonable predictive accuracy, must be able to give some examples of one or more models which (last year) predicted this year’s record cold temperatures in China, the Middle East, and North America.

    If yesteday’s weather forecast can not reliably predict today’s temperature to an accuracy of +/-0.5 C, how can we expect a GCM model to predict temperature 50 or 100 years in the future to the same accuracy? Errors in predictions do not diminish over time.

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>