I had a brief Twitter exchange with Roger Pielke Sr. yesterday. I had commented that I thought the AGU position statement on human-induced climate change was very good, and he wanted to know what was wrong with his alternative. Well, I’m not sure how Roger Pielke Sr. can produce an alternative statement. Surely, there can only be one AGU position statement. It seems a bit presumptious for an individual to think that they could produce an alternative.
Anyway, after the discussion I came across Roger Pielke Sr.’s, now retired, blog. I noticed that one of the last things he discusses is a paper called Ocean heat content and Earthʼs radiation imbalance. II. Relation to climate shifts by D.H. Douglass and R.S. Knox. In an earlier post Roger Pielke Sr. lauds this papers and finishes his post by saying
This paper is an important much-needed assessment of the climate system, and illustrates, for example, the inadequacies of using linear trends over multi-decadal time periods, as well as the value of using ocean heat data as the metric to assess global average radiative imbalance. I recommend the entire paper be read.
So, I’ve read the paper. The basic idea is that there are two energy terms to consider, the top-of-the-atmosphere (TOA) flux and the input from geothermal energy (taken to be 0.087 Wm-2). Together these tell you the rate of change of the ocean heat content (dOHC/dt). The basic equation they use is
FTOA + FGeo = 0.62 [dOHC/dt].
The 0.62 on the right-hand side is because the rate of change of ocean heat content is normally in units of 1022J/yr, while the TOA and geothermal fluxes are in Wm-2 (i.e., convert 1022J/yr to J/s and then divide by the area of the Earth). They then take the ocean heat content data, which I show in the figure below, and use it to determine dOHC/dt.
From the rate of change of ocean heat content, they then use the equation above to determine the TOA flux (FTOA) assuming that Fgeo = 0.087 Wm-2. Their result is shown in the figure below. From this they conclude that there are “climate shifts” where the TOA flux switches from being positive to being negative and then back again sometime later. They are therefore concluding that there are lengthy periods where this is no warming (FTOA negative) and that these are associated with some kind of, as yet unexplained, climate shift.
Issues with this paper have already been addressed in a comment by Nuccitelli et al., and you can read more at Skeptical Science. One fairly basic problem is that they only used ocean heat content down to 700m, while we now have data down to 2000m. Including this extra data, changes the result considerably.
I, however, have a much more fundamental problem with the paper. This seems to be mentioned in Dana Nuccitelli’s comment paper, but doesn’t seem to be addressed fully (as far as I can tell). Essentially, the fundamental problem is that their assumptions and their equation (that I show above) is wrong. Let me see if I can explain this.
The Earth’s climate has two sources of energy input, there’s energy from the Sun and there’s geothermal energy. We primarily lose energy through radiating energy back into space. If the system was in balance then we would have
FSun + Fgeo = Fout.
If it wasn’t in balance the TOA energy imbalance would be
FTOA = FSun + Fgeo – Fout.
If this was positive, then the system would be gaining energy; if negative, the system would be losing energy. Given that the oceans absorb most of any energy excess (at least 90%), then one could estimate the TOA imbalance by using the gradient of the ocean heat content data. However, you don’t subtract the geothermal flux from this gradient (as Douglass & Knox appear to have done) because the geothermal flux is already included in the TOA flux imbalance. The gradient of the ocean heat content data, alone, is the estimate of the TOA energy imbalance (or about 90% of the imbalance).
What I think Douglass & Knox have done is the following. If you have a satellite that attempts to measure the TOA energy imbalance, it can measure (directly) the energy from the Sun, and can measure directly the energy being radiated back into space. However, it can’t measure the geothermal energy entering the system. Therefore a satellite measurement of the TOA energy imbalance is essentially
FTOA = FSun – Fout.
In other words, a satellite measurement will underestimate the flux imbalance by an amount equal to the geothermal flux. I don’t know if this is corrected for or not, but the errors in such measurement are quite large and the geothermal flux is quite small, so worrying about it may not really be necessary. So, at best, what Figure 2 from Douglass and Knox is showing is what a satellite would directly measure the TOA flux imbalance to be. It’s not the actual flux imbalance because to get that you need to add back the geothermal flux. Hence, even with only data down to 700m, the TOA flux imbalance would be positive for all times after 1970 (which should be obvious from Figure 1 anyway). If you include (as Nuccitelli et al. correctly suggest) data down to 2000m, it is positive and significant (~0.5 Wm-2).
So, why am I writing this? Well, I know Roger Pielke Sr. is a professional climate scientist with a fantastic publication record and lots of experience. How, then, can he think this paper is an important much-needed assessment of the climate system? As far as I can tell, they not only ignore a big chunk of important data, their basic assumption and calculation is simply wrong. There’s a chance that I’ve misunderstood something so, if I have, feel free to tell me what and I’d be happy to be corrected. However, until such time, I certainly won’t be seriously considering an alternative to the AGU position statement from someone who thinks that Douglass & Knox is an important piece of work.