Last year I wrote about the Monckton, Soon, Legates & Briggs paper, Why models run hot: results from an irreducibly simple climate model. Well, Mark Richardson, Zeke Hausfather, Dana Nuccitelli, John Abraham and myself, have managed to publish a response (Richardson et al. 2015) that came out yesterday. Greg Laden has already covered it as has Collin Maessen. I expect a couple of others articles out soon, that I shall link to when they’re live. I don’t think, therefore, that I really need to say too much.
Monckton et als. model was essentially
where is the non-feedback climate sensitivity, is the feedback fraction, and is the transient response at time . The one clear error in their formalism is that the transient fraction, , should really be convolved with the forcing timeseries since it is really the response at time , to a change in forcing from an earlier time. Arthur Smith explains this issue really well.
Monckton et al., in a sense, resolve this by assuming that . The problem with this is that it assumes that the response is instantaneous and, therefore, that the heat capacity of the system is zero, and that there should never be a planetary energy imbalance. Well, that’s clearly wrong. Monckton et al. also assume that the feedback fraction, , should be small () [Edit : As Zeke points out, this isn’t quite right. Monckton et al. actually assume that . The confusion is that I've gone and used an , while Monckton et al. used a .]. They base this on electronic circuit design and on the fact that the temperature variations over the last 800000 years have been small in absolute terms. Both of these arguments are wrong. The temperature variability over the last 800000 years is actually consistent with the feedback fraction not being small ().
Maybe the most amazing thing about Monckton et al. is that they test their model by extrapolating temperature trends to 2050. Sure, if you can make up future data you can probably make any model fit well. Anyway, that's all I was going to say. I'll link to the other articles if and when they appear. We're also trying to work out if there is a way to make the paper publicly available. We couldn't get anyone to sponsor the open-access fee. If anyone would like to know any more, feel free to ask in the comments.