A talk by Macquarie University’s Murry Salby where he opines that the increase in atmospheric CO2 is natural is gaining some attention. (See, for example, Gavin Schmidt, Judith Curry, and Things Break). Unfortunately, we just have the audio and Salby has not responded to my request to provide the slides from his talk, so I’ve used a graph from Open Mind to illustrate this post.
A graph like the one below is Salby’s key argument. It compares the annual change of CO2 concentration in the atmosphere (black line) with anthropogenic emissions (the red line is 55% of emissions). Salby observes that the black line goes up and down while the red line is steady and concludes (at 15:45):
“Net global emission of CO2 changes independently of CO2 [emissions by humans]”.
Salby goes on to calculate the correlation between temperature and annual net emissions of CO2 and concludes that it is temperature causing the increase in CO2 rather than the other way around. The trouble with his argument is that by looking at the annual emissions he is looking at the first difference of the concentration and has removed any long term trend. Yes, since human emissions vary little from year to year they don’t cause short-term variations in net emissions. But what matters is the long-term trend and Salby’s analysis tells us nothing about that. If all this seems familiar, that’s because it’s the same mistake John McLean made. If anything, Salby’s error is worse, since he also ignores the conservation of mass: The increase is atmospheric CO2 mass is less than the mass of anthropogenic emissions, so what does Salby think happened to all that CO2?
Nor is Salby’s argument novel: you can see it debunked at Skeptical Science last year after someone at Watt’s blog advanced it.
I was lucky enough to attend Murry Salby’s talk at the IUGG conference in Melbourne. The thesis is not quite so simple as a correlation between CO2 rise and short-term temperature variations, because he found corroborating evidence in the change of CO2 slope over time. This made the argument not so easy to dismiss out of hand, although Salby was extremely careful not to draw any conclusions in his public presentation.
It was quite good sport to play “spot the flaw” in real time. Fortunately, the talk was the last of the session, and both Alan Plumb and myself chatted with him right afterwards. Aside from whether a statistical argument makes physical sense, it also must hold water statistically by being applicable beyond the time frame of model development. In discussing what his model would mean for past variations of temperature and CO2, it eventually became clear that he believed all paleoclimate data that supported his statistical analysis and disregarded all paleoclimate data that countered his statistical analysis, even though the latter collection was much larger than the former.
Eventually I realized that if 0.8 C of warming is sufficient to produce an increase of 120ppm CO2, as Salby asserted, then the converse would also have to be true. During the last glacial maximum, when global temperatures were indisputably several degrees cooler than today, the atmospheric CO2 concentration must have been negative.
That was enough for me.