Watt about the CFCs?

Watts Up With That (WUWT) has a new post about a new study suggesting that global warming is caused by CFCs not carbon dioxide. It refers to a paper by Q.-B. Lu with the title Cosmic-ray-driven reaction and greenhouse effect of halogenated molecules: culprits for atmospheric ozone depletion and global climate change. I’ve had a look at this paper. It’s quite long, so can’t claim to have gone through it in detail but below is my summary of what is claimed. As usual, happy to be corrected by those who know better.

The paper starts by attempting to remove the solar influence on global surface temperature anomalies. The paper produces the figure below. The figure shows the measured global surface temperatures (orange dots), the solar irradiance (pink line), and the global surface temperatures with the solar effect removed. The net result of this correction is that the surface temperature anomalies appear almost flat up until 1970 and then only start to rise after that. The basic claim seems to be that pre-1970, the global surface temperature rise was driven almost entirely by solar variation and that CO₂ did not play much of a role. The implication is that this makes it hard to reconcile with CO₂ being an important greenhouse gas because the change in CO₂ concentration from 1880-1970 should have had the same effect as the change from 1970-2013. If it didn’t contribute prior to 1970, how can it be dominant after 1970?

Measured global surface temperatures (orange dots), solar irradiance (pink line), solar-corrected surface temperatures (red dashed line) (credit : Lu 2013).

One of the reasons this claim is confusing is that many climate models have successfully reproduced the global surface temperatures from 1880-2013 with all the normal forcings (CO₂, solar, halocarbons). The paper by Q.-B. Lu states that

the observed temperature increase by 0.6^oC in the late half of last century has forced climate models to use a very low climate sensitivity – 0.8 K/Wm^-2.

I find this a rather odd thing to state. In the absence of feedbacks a doubling of CO₂ produces a total increase in net forcing of 3.7 Wm^-2 which results in a 1K increase in global surface temperatures. Therefore, in the absence of feedbacks, climate sensitivity would be 0.27 K/Wm^-2. The supposedly low climate sensitivity used by climate models of 0.8 K/Wm^-2 is therefore associated with a 3K increase in global surface temperatures. Doesn’t seem “very low” to me. Seems right in the middle of the expected range.

I suspect what has happened is the following. Even climate models show that solar forcing was quite significant in the first half of the 20th century and is responsible for most of the variations in the global surface temperatures. As far as I can tell, the paper hasn’t done any modelling and has simply assumed a linear relationship between solar forcing and surface temperature changes. The paper has simply chosen a relationship that essentially removes the trend in the global surface temperatures prior to 1970. This doesn’t mean that CO₂ was not playing any role. It is simply consistent with solar forcing being significant and being responsible for most of the variability prior to 1970.

What else does the paper go on to say. It includes the following plot which shows a linear relationship between the global surface temperature anomalies and Halocarbon concentrations (in ppb) between 1970 and 2012. This leads the author to conclude that the relationship between Halocarbon concentration and radiative forcing is ΔF = χC. The paper then goes on to suggest that one can determine the change in surface temperatures due to Halocarbons by using
ΔT = α β ΔF,
where α is the climate sensitivity factor and β is the climate amplification factor. Using this they produce the following figure which shows the global surface temperature anomaly (with solar effect removed) compared with temperature changes due to CO₂ forcing (dashed line) and Halocarbon forcing (green line). The claim – as should be fairly clear from the paper – is that there is a much better match to changes due to Halocarbon forcing than due to CO₂ forcing.

Comparison of global surface temperatures (with solar effect removed) with forcings due to CO₂ (dashed line) and Halocarbons (green line) (credit : Lu 2013).

Firstly, there is something fundamentally odd about this. If you consider the very first figure – which shows the global surface temperature anomalies with the solar effect removed – it shows that the temperature anomaly is essentially zero between 1880 and 1970 and then rises rapidly from 1970 till 2012 (and is essentially unchanged,between 1970 and 2012, from the uncorrected values). This is primarily because Lu (2013) claims that he’s removed the solar effect. The conclusion is therefore essentially that CO₂ provides no forcing from 1880 to 1970 and then all the forcing from 1970 to 2012. I suspect that what this analysis has done is over-represented the significance of CO₂ from 1970 to 2012 and consequently enhanced the disparity between 1880 and 1970.

There is, however, a more fundamental problem. The paper assumes that the relationship between the temperature anomaly and Halocarbon forcing is given by the equation above. The figure below (from Lu Q.-B., Journal of Cosmology, 8, 1846-1862, 2010) shows the variation of surface temperature anomaly with time for different climate sensitivity factors, α, and climate feedback factors, β. The best fit seems to be for α = 0.9 and β = 2.0. One can use the equation above to calculate that this suggests that the change in forcing due to Halocarbons is ΔF = 0.333 (this is actually quite similar to the value used by the IPCC).

Figure showing temperature anomalies against time for different Halocarbon climate sensitivities and Halocarbon climate feedbacks.

This is what I think is the major issue with this work. If one assumes that most of the forcing is due to some atmospheric constituent (either Halocarbons or CO₂) then the climate sensitivity in the absence of feedback is a surface temperature change of 1K for a change in radiative forcing of 3.7 Wm^-2. This gives a climate sensitivity of 0.27 K/Wm^-2. If you want to use the equation ΔT = αβΔF then α = 0.27 K Wm^-2 and β is the feedback factor. To reproduce the results shown in the figure above would then require that β = 6.666. In other words, for Halocarbons to be the dominant driver of global warming since 1970 would require a feedback factor of almost 7. Using α = 0.9 K/Wm^-2 suggests that in the absence of feedback a change in forcing of 1.1 Wm^-2 would produce a change in surface temperature of 1K, which seems rather unphysical to me. One can do a basic blackbody flux calculation to show why this doesn’t really make sense, but I’ve done this in previous posts so won’t repeat it here (happy to respond in the comments if anyone would like clarification).

This post has become rather long and somewhat convoluted, so I hope it makes sense. One of points I was trying to make is that calculation to remove the solar effect appears to be rather simplistic and essentially assumes that Sun provided almost all the forcing prior to 1970 and virtually nothing after 1970. The other point was that the forcing calculation used to represent the influence of Halocarbons would seem to require a feedback factor of almost 7, much bigger than the feedback factor needed for CO₂. Quite what provides this extremely large feedback is not made clear (as far as I can tell). This seems like a fairly classic example of reasonable complicated data manipulation and calculations that show correlations, but that dooesn’t really include much in way of the physics that would be necessary to show that these results actually make sense.