Monday, January 25, 2010

More Schwartz

I suppose I should say something about this, though I'm really rather bored by it. The sexy title "Why Hasn’t Earth Warmed as Much as Expected?" rather belies the content, which is just a simple energy balance of the earth system, with no new observational analyses or theoretical insights (contrast Murphy et al for a real job).

The "as Much as Expected" in the title is interpreted as the equilibrium response for GHG forcings alone, given an assumed sensitivity of 3C per CO2 doubling. The obvious answer is that (a) of course the system isn't in equilibrium, and (b) there are other forcings too, notably (but not solely) tropospheric aerosols. In fact, current GCMs which account for these factors do an excellent job of reproducing the observed warming, but the more honest and straightforward title of "Why Has Earth Warmed Just as We Expect" would not have been so catchy.

That picture is from Wikipedia (click for page with caption), based on data from the PCM ("Parallel Climate Model" of NCAR and collaborators, equilibrium sensitivity 2.1C). Just to show that this simulation does not indicate the IPCC's 3C sensitivity estimate is too high, here is HadCM3 (sensitivity 3.4C) doing the same thing 10 years ago in the IPCC TAR (again, click for the captioned version):

I've seen equivalent figures for MIROC3.2 (sensitivity 4.0C) too but don't have them to hand. Although one might argue that the models are over-tuned to the past, and thus this agreement doesn't prove anything, blah blah, one can hardly pretend that there is a puzzling discrepancy between the energy balance of the planet (which of course these models simulate) and the observed warming. So the premise of the paper is a big fat straw man at the outset.

The calculations presented are rather trivial zero-dimensional energy balance estimates. Reasonable enough for a blog post I suppose, I'm surprised that they alone are adequate content for publication in JClim. And anyway, when these two big missing factors are included (and set to reasonable values), the paper's conclusion is that the observed warming is entirely compatible with the simple energy balance theory:
Thus countervailing aerosol forcing could account for much or all of the discrepancy between the expected and observed increase in GMST over the industrial period.
There is no useful uncertainty analysis, just a repetition of the well-known observation that these data on 20th century warming do not tightly pin down the equilibrium sensitivity, and overall, it doesn't seem to me that this paper adds much to the debate. Climate scientists have been well aware of such simple energy balance calculations (and their limitations) for some decades now. So despite the misleading title which has predictably resulted in an equally misleading press release and coverage, there is really nothing for anyone (not even the sceptics!) to get excited about.


  1. Increasingly the press release IS the point. The question is who is the audience.

    This is the old Pielke two step, one version for the colleagues, another for the Watts up crowd

  2. This is more or less what I thought too... and I only read the abstract ;-)

  3. I'm not a scientist at all, but I had that same impression after reading the abstract and the conclusion: what is this supposed to add? Is there anything new here?

    IPCC 2007:
    "Greenhouse gas forcing alone during the past half century would likely have resulted in greater than the observed warming if there had not been an offsetting cooling effect from aerosol and other forcings"

    From Fig. 2.20 it is clear that the biggest uncertainty in forcings during the 20th Century is aerosol forcing. It seems obvious that the less the negative forcing from aerosols, the more the net forcing and (thus) the lower the climate sensitivity. And vice versa, the more the negative forcing, the less the net forcing and (thus) the higher the climate sensitivity. All this within the IPCC uncertainty boundaries.

    Denialists are, however, suggesting that this implies that models are wrong. It is clear that IPCC models took aerosol into account. Even their press release doesn't mention models at all.

    Regarding that this is not the equilibrium response, they at least seem to take into account the ocean heat content.

  4. Any hope of that graphic being brought up to current date? It's helpful but old.

  5. @Eli: 'Increasingly the press release IS the point. The question is who is the audience.'

    Indeed. And, alas, alarmist science is equally guilty. And press releases are the innocent end of it. Abstracts and titles are also tilted to further non-scientific points.

  6. By definition, if it is science it is not "alarmist".

    Somebody ought to go back and read Francis Bacon.

  7. Vinny: "And, alas, alarmist science is equally guilty."

    Only if you imagine that the climate isn't headed in a certain direction. IOW, Vinny, if you're not alarmed by now it's because you're pretty much crazy.

  8. Speculations on why JClim published this?

  9. Hank,

    I agree it's a bit dated - perhaps influenced by the 2000 cut-off of IPCC scenarios, but hopefully someone somewhere will have a go. Currently most people are panicking about getting their newest model version up and running for the new IPCC runs :-)

  10. David,

    Well someone somewhere must have thought it was a valuable addition to the literature :-) It only takes two friendly reviewer (or even one, with another who can be over-ruled).

    Although based on the hype and title/abstract I was initially sceptical about it, now having read it, it isn't actually that bad in scientific terms, just a bit pointless, and hyped with a misleading press release. Roe and Baker was probably worse, and that got into Science.

  11. Eli,

    Agreed. Once might be an accident, and eg others are more forgiving of the Latif cooling stuff, but when it becomes a habit, it is hard to avoid drawing conclusions...

  12. David B, it's the tilting of abstracts and titles that makes it alarmist. It's science in an alarmist wrapper, if you prefer.

    But it was rude and silly posting that comment here. It doesn't belong. (Besides, I don't have a convincing alarmist 'idola specus' example handy.)

  13. I'd hardly call this "alarmist", although I still think the "alarmist" tag is a bit much on my old project which mentioned "11C sensitivity" of some model results on my old project (although that was a pet fave to bash here & at RC et al). (disclaimer - I had nothing to do with the science on that -- and even another pet fave to bash here was against mentioning the upper-end of the models in the press release ;-).

    maybe it's just some problem with how institutions feel a need to feel "valuable" in spending on scientists. It seems it's not enough to get into Nature or JGR or whatever. It does seem to be that if 100 groups are all saying sensitivity is 3C +/- a degree, there is some need to "sex it up" to get a headline or good press release. it's sort of like collage (American) football, i.e. hyped up, but not necessary "bad" or "alarmist" science IMHO.

    stuff like the "UEA scandal" was far worse of a "hit" for the field, than any sexed-up title or press release or Jim Hansen crazy quote....

  14. James, do you remember offhand what the assumed aerosol forcing for MIROC3.2 is?

    From what Gavin has told me, I would assume that aerosols are pretty much a free parameter in the models, so you can dial in almost any number for climate sensitivity from maybe 2 to 11 that you want and still get agreement with the historical data. Eventually what is it that constraints the upper end up this range?

    Model E forcings is here for comparison:

    Like most models, it gives little or no net anthropogenic forcings prior to circa 1975.

  15. About -0.7W/m2 rings a bell (for recent compared to pre-industrial forcing) but I'm not sure. Note also that the GHG forcing is unusually low in miroc, about 3 or 3.1W for a doubling of CO2. It doesn't seem to quite add up, though there are also some other uncertainties like ocean heat uptake and also things like the sea ice extent in the current simulation may affect the transient response.

    Ultimately, there are no really rigid constraints, but it seems that most modelling efforts give similar (say within factor of 2) answers. It is worth emphasising that these numbers in the GCMs are not assumed or selected but arise as a consequence of peoples' attempts to simulate the physical processes as best they can. However, it is very hard to reconcile the range of observed responses - eg the short-term cooling for volcanoes, the gradual warming over the 20th century, and the historical periods such as the last glacial maximum, with either very high or very low values. Everything fits pretty well with about 3C climate sensitivity, with no need to invoke anything too extreme or implausible.

  16. Any news on this or Schwartz?

  17. In a Which is probably no bad thing :-) I hadn't seen that paper but based on the abstract it doesn't say anything interesting. The original was a rather pointless straw man.



    makes for a good and sad start :(

  19. Also see:

  20. Was on a symposium where Schwartz was one of the presenters. Have to say that he came across really well not saying silly things... defending the IPCC etc... Sounded like his publications comes from a genuine belief that overconfidence in today's models... the press release might as much have been done by university pr folks... who knows...