Yin Yang Tree

This paper was recently pointed out to me as absolute proof of a huge level of CO2 in the atmosphere around 1940, much greater than the current level. I hadn't seen this paper before, but apparently it has gained much credence with global warming skeptics that don't bother to look at actual scientific papers. So, I wanted to put up a quick review in case anyone else runs across this, giving some quick ways in which it and the underlying data are flawed.

First let me note that this paper does nothing except take previous measurements by other scientists, combine them, smooth them, and present them. There is no discussion or hypothesis on what could lead to such giant rises in CO2 or their subsequent precipitous falls. Naturally, since neither of those topics is discussed, there is no discussion of why, since the CO2 measurements have been continuously recorded since 1958, we don't see either of those phenomenons since that time.

Given that, let's take a look at some of the underlying data and see what we can note from it. First, the sites of these measurements not chosen to be accurate representations of the general atmospheric CO2 as the current sites are. Secondly, looking at the data, we can see that, in many cases, the large fluctuations in measurement weren't year to year or even month to month, but day to day, and even hour to hour. Let's take the Misra data set as an example. Look at the first line of measured CO2 percentage, taken at 06:00:00 on 12/14/1941. They show .052, .063, and .087 at the respective measurement heights. at 10:00:00 4 hours later the measurements were .031, .030, and .037. approximately 1/2 of the CO2 concentration measured just 4 hours earlier. Of course, that's just one data set. Let's look at some more. Duerst, Kreutz, Hock-Schollander, and to a lesser extent, Haldane all show tremendous fluctuations over very, very short periods of time. Ones that don't show these incredible fluctuations, such as Buch and Steinhauser all show measurements much more in line with the ice core record and well below the current CO2 level.

Please also note, the Misra data is especially important, since it is the only titrimetric data measurement between 1939 and 1950 that Beck uses.

One last note. Ever since the modern measurements have been taken, we have seen very regular annual cycles of CO2, peaking in April/May/June and bottoming out in Sept/Oct, causing 10-15 ppm difference. Looking at Kreutz, the volumetric measurement from 1939-1941, we see April/May/June hovering around 340-370 ppm whereas August/Sept shoots up to over 500 ppm.

So, in summary, let me say that wildly varying data (in ways we know that it shouldn’t vary) from 2 scientists in specific locations that are not controlled for any kind of CO2 contamination (Misra's location even seems to be chosen because of the agricultural CO2 contamination) with no even suggested theory as to how these spikes came about or why they dissipated isn't much. I guess that's why this paper was published here, instead of an accepted peer-reviewed journal.

On the other side, we have a method of measuring historic levels of CO2 at a location we know has no contamination issues and that current measurements match almost exactly with the continuous measurements of other locations chosen to avoid contamination. This doesn't seem like much of a contest, but somehow, there seems to be lots of grasping at this particular straw.

Update - 4/3/2007: I wanted to elaborate on two points:
1. The fact that the data is varying so much in the samples above could have one of two explanations. First, it's possible that there is a close source of CO2 or CO2 absorption. In any event, it is clear that what is being measured is not a stable condition of ambient CO2 in the atmosphere. The other explanation is, of course, that the measurements are insufficiently precise.
2. On the subject of precision, the volumetric method used by many of the papers referenced by Beck is discussed in the Erikson paper:

The average CO2 found in the Point Barrow samples is near to 0.040 per cent, and hence is seemingly 0.01 per cent higher than the standard value. This discrepancy is, however, within the analytical uncertainty of the method and may be unreal...Within the accuracy of the method (+ or - 0.015 per cent) the carbon dioxide and the oxygen concentrations were found to be constant through the seasons and within the standard values for temperate regions.

So, the volumetric method, trying to measure something that is approximately 0.030% of the volume of the sample has a error of 0.015%, or half of what it's trying to measure. Great...