That is Not a Correction

In my last post, I pointed out an untrue statement made in an article by a climate researcher which the author obviously knew was untrue. I also predicted in a comment:

It should be interesting to see if he directly acknowledges what he wrote was incorrect and tries to fix it. I'm cynical so I expect he won't. I expect neither he nor anyone else who recognizes what he wrote was incorrect/misleading to do a thing about this. That's been the result in most cases in my experience. My experience is people usually refuse to squarely deal with errors/mistakes.

Sometimes I think my cynicism is excessive, but the researcher posted a "correction" to his article which proved my expectations correct. I'm going to show how the "correction" he posted didn't actually correct anything.

The claim in was:

Broecker also did not take other greenhouse gases into account in his model. However, as the warming impact from methane, nitrous oxide and halocarbons has been conveniently cancelled out by the overall cooling influence of aerosols since 1970, this does not make that large a difference.

Normally I would link you to where you can find this claim, but as it happens, the "correction" isn't documented or disclosed anywhere. The article was secretly changed. You can see the current version here (though who knows if it might change again). I don't want to harp on how it is inappropriate to secretly change articles after they've been published, but... come on! How can you possibly think that's okay?

In any event, the problem with this claim is obvious to anyone who studies temperature change and climate forcings - the effect of aerosols on the planet's temperature is highly uncertain. As documented in my last post, its effect is estimate to be anywhere from 10% to 190% of the effect of greenhouse gases other than carbon dioxide (in the opposite direction). The author of that article knows this.

If you know an effect could be as small as 10% the size of another, why would you say the one effect "conveniently cancelled out" the other? You have to know that will mislead people into believing the uncertainty in size of the effect is small enough to justify what you say even though it is not. The author responded to me on Twitter to say:

Obviously, the point he says he was trying to make does not match what he actually wrote. I pointed this out and said:

Which annoys me because I left the word "out" out of the phrase "called out." The lack of an edit/preview feature on Twitter can hurt sometimes. Fortunately, this tweet still conveyed my meaning clearly enough and the author responded to inform me the article had been updated to say:

Broecker also did not take other greenhouse gases into account in his model. However, as the warming impact from methane, nitrous oxide and halocarbons has been largely cancelled out by the overall cooling influence of aerosols since 1970, this does not make that large a difference (though estimates of aerosol forcings have large uncertainties).

That is not a correction. The author doesn't dispute my point about the uncertainties in the impacts of aerosols. He can't because he knows I am right. However, despite knowing I am correct in pointing out the uncertainties are so large the effect of aerosols might only be 10% that of "methane, nitrous oxide and halocarbons," he still says the one effect "largely cancelled out" the other.

That is complete nonsense. The fact he adds a parenthetical at the end of the paragraph to say there are "large uncertainties" in the impact of aerosols doesn't fix anything. I can't imagine people reading this parenthetical will interpret it as saying, "You know when I said the impact of aerosols largely cancelled out that other impact? Yeah, we don't actually know that's true." Even if they would, why write a paragraph with such contradictory messages?

According to the author, it's okay to state as fact that aerosols have largely cancelled out the impact of greenhouse gases other than carbon dioxide because "the mean estimate of" the effect of aerosols say so. That there are enormous uncertainties doesn't matter. He apparently feels comfortable stating as fact something he knows he does not know to be true.

It would have been easy for this author to use a phrase like "believed to have largely cancelled out." That would have been an accurate statement of what the current body of reserach indicates. Instead, the author "corrected" his piece by changing the text in a way which still states as fact something he knows not to be a fact.

Nobody is going to speak out about this. No climate researcher is going to come out and say, "Hey, that's an error. You should fix it." None of them are going to say, "Hey, don't make secret changes to article you publish. If you're going to make a change, document it." And certainly, none of them are going to say, "Hey, dude, that link you provide to support your claim says nothing about the issue."

Yeah, that's right. The phrase "have large uncertainties" in the article is given with this hyperlink. The blog post it takes you to is from 2007, which I don't understand. I would think if we want to talk about the uncertainties we face nowadays we would want a source from sometime in the last decade. Even if not, surely we'd want something a bit more detailed than a blog post which says, without providing any references or details:

The forcing for CO2, CH4 (including indirect effects), N2O and CFCs is 1.66+0.48+0.07+0.16+0.34=2.71 W/m2 (with around 0.3 W/m2 uncertainty). Using the formula above, that gives CO2_e (Kyoto) = 460 ppmv. However, including all the forcings (some of which are negative), you get a net forcing of around 1.6 W/m2

At a minimum, we would want a source that actually says something, anything, about the uncertainties in the estimates of aerosol impacts.

But hey, global warming is just the most serious threat to humanity. It's not like honesty or accuracy in one's communications about it could actually matter.


  1. Since 1970 forcing from non-CO2 GHGs has been about 0.55w/m2, plus about 0.1w/m2 for ozone (tropospheric minus stratospheric). That's about 0.65w/m2, versus 0.18w/m2 for aerosols. The increase in aerosol forcing may be even smaller, as the IPCC estimate ends in 2011 and this forcing was declining. In any case, ignoring aerosols and non-CO2 GHGs means ignoring about 0.5w/m2 or forcing.

    See page 1408. Hausfather's statement may be true if one talks about total forcing since 1750, but not since 1970. Furthermore, CO2 forcing since 1970 has only been about 1.1w/m2; including ozone and other GHGs would increase this by about 50%. So the 30% overshoot in the Broecker model probably would have been a 100% overshoot.

  2. Please notice my previous comment uses rough measures. The forcing missing from the Broecker model may be not an additional 50%, but an additional 45% or whatever. Nevertheless, recent estimates suggest methane forcing (and its increase since 1970) have been underestimated

    And aerosol forcing if anything is being revised downwards, which probably would make the increase since 1970 smaller as well.

  3. Thanks for your comments Reader. When I wrote my previous post about this, I intentionally buried the lede by putting this issue at the end because I didn't want to try to make a detailed, technical argument. It's been a couple years since I last looked at the time evolution of climate forcings so I knew an issue like the one you raised might exist but didn't have the knowledge to make a case for such. I think the IPCC table you reference makes it clear Hausfather's claim was even more wrong than I indicated.

    I would be curious to see what, if anything, Hausfather has to say about this, but given the "correction" he made in response to my initial remarks, I don't have high expectations.

  4. I just read the article. Without knowing exactly what forcing values went into each model, it's little more than a curiosity. Of course one can always 'explain' things after the fact: oh, it turns out we over-estimated X and under-estimated Z; after correcting for this it turns out the models are correct! --> get your paper on Nature and congratulate yourself.

    He does mention that actual CO2 concentrations were similar, or lower or higher than used in the models, but what about methane, ozone, etc.

    The author also states most of the models' sensitivity, but really, two models with rather different sensitivities may still give an almost identical temperature projection a few decades out. What he should have reported is the models' TCR. It's about 30% higher in CMIP5 models than in recent sensitivity papers, roughly 1.4C vs 1.8C.

Leave a Reply

Your email address will not be published. Required fields are marked *