As already said by Bob North, one need to compare emissions data against atmospheric increase data. That is much more interesting…

For the graphs, see
http://www.climateaudit.org/?p=2720#comment-215106

There is a big problem with looking at too short time frames, depending of the signal-to-noise ratio. The smaller the time frame, the bigger the noise (and the errors).

If you take e.g. the seasonal changes by month-to-month averages, then temperature is certainly leading CO2 levels (inversely!), due to the NH vegetation influence. Long-term trends are not visible in seasonal variations, as much smaller.

If you take year-to-year variations (dCO2/dt), then we see that there are two components: the temperature-influenced variability (which is about +/- 1.3 ppmv) and the average increase, which is about halve the emissions (for the Mauna Loa time frame).

If you take the long-term accumulated emissions vs. increase in the atmosphere (CO2/dt), you have a near fit (correlation: 0.999, R^2: 0.9988), with an increase of 55% of the emissions (Mauna Loa). For the ice core time frame, 1900-1960, the increase is about 58% of the emissions (R^2:0.911). Seems similar. The natural (temperature induced) variability is not visible anymore…

The 1940-1945 dip in the ice core CO2 seems an artifact of the measurements (1 ppmv drop within an accuracy of +/- 1.2 ppmv), as the drop preceeds the temperature drop with about 5 years…

Many of the objections of Jaworowski about the accuracy of ice core measurements were already refuted in 1996 by three core drillings of Law Dome. They used different drilling techniques (wet and dry) and avoided clathrates and cracks in the ice. See Etheridge ea. at:
http://www.agu.org/pubs/crossref/1996/95JD03410.shtml

The differences within the three cores were small, within +/- 1.2 ppmv (1 sigma). They also measured firn CO2 and noticed a lag of only 10 years between ice age and gas age at closing depth and an about 5 years smoothing for the two fastest accumulating cores (25-40 years for the third one, used for 1,000 years history). This refutes two of the allegations of Jaworowski: that the drilling method influences the CO2 levels and that the ice closing influences CO2 levels (closed and still open bubbles at closing depth had the same CO2 levels).

Further, ice core CO2 levels and South Pole direct atmospheric measurements have an overlap of about 20 years (1959-1978) and are in the same range. This refutes another allegation of Jaworowski, that the ice core data and the atmospheric data were “arbitrarely” shifted with 85 years (for another ice core and Mauna Loa). The data series are based on ice age (by counting layers) and gas age (by following the firn diffusion rate).

A further allegation by Jaworowski is that cracks and decompression leads to underestimated CO2 levels in the ice core. But as the surrounding CO2 levels are higher than the ice core levels, this should lead to overestimates, not underestimates, if outside contamination would have occured…

Thus please take Jaworowski’s sayings with a little more than a grain of salt…

This is about the fifth discussion blog/list about Beck’s and Allan’s data…

To start with: I totally agree with the comments that Bob North delivered here. Some additions follow:

Besides being a series of independent in time and place CO2 measurements, most of them are known to be at places with high local production (as well as human as vegetation). Some (Misra, India) were even taken intentionally in rice and soya fields, partly under the leaves, to measure CO2 activity during growing season.

As the variability in Diekirch (Luxemburg) shows, one can not even think about knowing the “background”/global CO2 levels, if one doesn’t know the local circumstances and/or wind speed in the neighbourhood of local sources. Most historical measurements, especially these of the last “peak” in Beck’s data (1935-1942-1959: up and down 100 ppmv), are at such places…

Keeling himself (see his bio at: http://scrippsco2.ucsd.edu/publications/keeling_autobiography.pdf ) measured at Big Sur state Park and other places. CO2 level changes with a difference of over 100 ppmv within 15 minutes were not uncommon. But he measured d13C levels at the same time, which showed that local vegetation decay/uptake was the cause of the variation. Unfortunately such measurements lack from the historical data…

Where can we find background data: near/at all oceans (the CO2 exchange rate with the oceans is slow), above the inversion layer, and everywhere with sufficient wind speed.
If we look at Beck’s data, only a few places are of interest (in the 1930-1950 period): ship’s surveys and near-coast in the North Atlantic (these data are around the ice core levels), near-coast in Scotland (370 ppmv, but no details known) and the Island of Wight (270 ppmv, no details known), Barrow (AK, USA) and the Antarctic coast. Unfortunately, Barrow used very inaccurate equipment (+/- 150 ppmv) and Antarctic data showed enormous variations in CO2, but also in O2 measurements.

If you don’t look at only averages of Beck’s data, then near all datasets show wide variations and all have minima which includes the ice core trend…

The biggest problem with Beck’s interpretation is the underlying assumption that a change of about +0.5/-0.1 degr.C caused a jump of 100 ppmv and back (the latter is physically near impossible), or about 200 ppmv/degr.C in about 7 years up and 1,000 ppmv/degr.C in 7 years down.

In all ice cores, some 10 ppmv/degr.C is found for natural variations on (very) long time scales. For very short (1-2 years) time scales, that is about 3 ppmv/degr.C change in CO2 increase speed, not a change in increase…

The ice core values are going to be precise than the direct readings we have these days. It is only to be expected.

http://cdiac.ornl.gov/trends/co2/csiro/csiro-cgrim.jpg

Thanks very much for the links. Interesting papers.

Briggs

Well, yes. I haven’t seen his data set yet, though I have seem some commentary that it couldn’t possibly be any good. A refusal to look at the entire data set, to me, anyway, seems more unreasonable than a refusal to look.

If I find there is nothing to it, or that it is somehow misleading, I will certainly say so. But I wonder: if instead it was found that his database was useful, not to measure the absolute value of CO2, say, but for measuring rates of change of CO2, would you be willing to examine that evidence?

Technical note: suppose each station had it’s own bias, in whatever direction; as long as this bias was more or less a constant, we could still infer periods of increase and decrease; at any one station, this would not be interesting, but if we found the same thing at widely disparate stations (over the same time period), this would be interesting.

Briggs

But I’m afraid you’ve signed on to a dying industry.