GRL paper on climate shifts

This post is a diversion from the usual subject of this blog into the more technical aspects of climate change and global warming. I do this in reaction to a recently published paper in Geophysical Research Letters: A new dynamical mechanism for major climate shifts by Tsonis et al. The press release which announced this paper is here (which is also how my attention was drawn to it).

There have been blog reactions to this (for example, here and here). Most of them are by skeptics and Deniers who claim that this is somehow 'proof' that CO2 has no effect on radiative transfer in the atmosphere and does not cause global warming or some such nonsense. Perhaps, when reading the somewhat misleading press release you can parse those words to mean something like that. However, reading the actual paper (always the authoritative source!) tells a different story.

To begin, the authors of the paper are not addressing the issue of anthropogenic climate change, but rather that of interdecadal variability. They aren't the same thing. No vaguely respectable climate scientist would say that there is not interdecadal variability, and it is certainly a possible explanation for some of the observed warming in recent decades. Nobody has claimed otherwise, including the IPCC. Unfortunately as interdecadal variability is something that is poorly understood at this time, the uncertainty is likely to remain.

Second, the paper does show some interesting shifts in the observations of the global temperature trends in the early 1910s, the early 1940s and in the mid-70s. The correspond with the 'synchronization' of the behaviour of some of the primary modes of climate variability, the main thrust of the paper. Between 1910 and 1940, the temperature increases. From the 40s to the 70s, the trend is around zero, or even slightly negative. It again begins to increase from then until now. The timings of the synchronization and the trend changes are very close, which provides strong evidence in favor of their hypothesis.

The authors then use 'state-of-the-art' climate model to see if they can detect similar signals there. Two simulations are used: one with constant pre-industrial CO2 levels, the other a business-as-usual 21^st century simulation. The vertical axes on the plots are a bit misleading (they change between observations and simulations...), but the models show an effect that is much smaller than seen in the observations by at least a factor of 2. Also noted in the text: the authors removed a 2^oC/century trend due to radiative effects in the 21^st century simulation for clarity in the figures. This radiative trend would swamp the relatively small signal of the natural climate shift.

All in all, the paper provides a nice hypothesis for understanding the mechanisms of natural interdecadal climate variability, but the work is far from complete. For instance, there is no discussion of what it means physically when the indices synchronize. Can we observe it as it happens, or is it something that can only be determined after the fact? However, this hypothesis in no way refutes or invalidates the hypothesis of anthropogenic climate change through CO2, nor do the authors intend to do so. Here is the closing line from the paper:

...the climate shifted after the 1970s event to a different state of a warmer climate, which may be superimposed on an anthropogenic warming trend.

I think that is a pretty clear statement of the authors meaning.