Climate Science Humor: What if Your Model Predicts Wrongly

After noting that James Hansen’s 1988 climate model predicted too much warming in the subsequent 22 years, someone at Skeptical Science concluded:

The main reason Hansen’s 1988 warming projections were too high is that he used a climate model with a high climate sensitivity, and his results are actually evidence that the true climate sensitivity parameter is within the range accepted by the IPCC.

There is no consideration of the possibilities that (a) one or more other parameters were wrong or (b) the model — aside from parameter values — is wrong (e.g., it oversimplifies). Surely you are joking, Mr. Skeptical Science.

Thanks to Phil Price.

 

5 Responses to “Climate Science Humor: What if Your Model Predicts Wrongly”

  1. Tom Fid Says:

    Climate sensitivity is sometimes a parameter in simple energy balance models, like the Schneider-Thompson model later used by Nordhaus. But in GCMs, as used in Hansen’s work, climate sensitivity is an emergent property of other structure and parameters. So, stating that the model’s sensitivity was too high (as I recall, by about 1/3) subsumes both (a) and (b).

    Seth: Interesting. Not so much “subsumes” but “is an example of” both (a) and (b). Then I would ask: what about other examples? Why are they not considered as possible explanations for the wrong predictions?

  2. Phil Says:

    I can’t comment on the technical content of the Skeptical Science post — I haven’t even looked at Hansen’s model — but I have to say that you are mis-characterizing the post.

    The post says “Unfortunately, Dr. Christy decided not to investigate why the NASA climate model was too sensitive, or what that tells us. There are two main reasons for Hansen’s warming overestimates…”

    Obviously implicit in this statement is “Dr. Christy decided not to investigate…but I did.” You may be right that the guy didn’t consider that there are other possibilities, but you certainly can’t conclude that from what he says here!

    Seth: I also read the other answers (intermediate and advanced levels) to the question/objection. When a complex model with many assumptions predicts incorrectly, it is close to impossible to figure out why because there are so many possibilities. The Skeptical Science author says nothing about the difficulty. The reasoning seems to be: if we adjust Parameter X, the prediction becomes accurate, therefore the reason the model predicted incorrectly was that Parameter X was given the wrong value. If we could estimate Parameter X entirely independently, this would make sense but nothing is said about that and that too is very hard. To me it looks like there are two possibilities: (a) although the Skeptical Science writer wants to impress readers with the quality of climate science, an enormous and enormously careful and highly unusual and very important effort by climate scientists — figuring out why Hansen’s model was wrong — goes unmentioned or (b) the writer is overstating what is known.

  3. Phil Says:

    Seth, I think you’re missing Tom Fid’s point. The climate sensitivity is not an input parameter to the model; it’s an outcome that is determined from the model.

    Seth: When you put it like that, the Skeptical Scientist answer sounds circular. “Why is the room too hot?” “Because it got too warm.” “Why was the predicted warming too much?” “Because a measure of output closely correlated with amount of predicted warming was too low.” I do not disagree with you, however.

  4. Phil Says:

    It’s not circular.

    Suppose we agree that Hansen predicted more warming than actually occurred. There are three ways that could happen:
    1. Emissions were higher than predicted, but climate sensitivity to those emissions was much lower than predicted.
    2. Emissions were about the same as predicted, but climate sensitivity was lower than predicted.
    3. Emissions were lower than predicted, but climate sensitivity was about right or was slightly higher than predicted.

    Of course you could make finder distinctions — “much much lower,” “very slightly higher,” etc. — but these are the coarse ones.

    Any of those three possibilities would lead to less warming than predicted. Global warming skeptics assume the answer is 1 or 2. The Skeptical Science guy is claiming to have looked into it and found the answer is 3.

    It’s not circular.

  5. Phil Says:

    Note that I have not looked into it and _I_ am not making any claims here. I’m just explaining that you have mischaracterized the Skeptical Science blog post.