From: farrar@128.160.13.2 (Paul Farrar)
Newsgroups: sci.environment
Subject: Re: T Moore's CO2 Level Claim
Date: 27 Jun 1995 15:09:04 GMT

My news server has not gotten Moore's reply yet, but I assume Tom Gray's
citation is a correct copy.

Tom Moore writes:
> farrar@128.160.13.2 (Paul Farrar) writes:
> >In his essay, "Global Warming, A Boon for Humans and Other Animals"
> >revision of 20 Mar 1995, URL
> >http://www-formal.stanford.edu/jmc/progress/globalwarming2), Thomas
> >Moore, The Hoover Institution, makes the following claim: 
> >   Studies of carbon (CO and CO2) in the atmosphere show that at
> >   times in the last 8,000 years the level has been substantially higher
> >   than it is today and greater than it is likely to reach any time
> >   soon. [line 93]
> >
> >This claim has no footnote in the online copy. The statement is, by the
> >best evidence, completely false, and so probably cannot be supported by
> >any citation. This is not a minor quibble, because the whole theme of the
> >essay is that any possible CO2-induced warming has historical precedent,
> >and that the precedent was generally favorable.
> 
> The published copy cites Crowley and North 1991, p. 70.

The problem is that p. 70 has nothing at all on CO2 levels. It appears
that Moore has mistaken a plot of the proportion of atmospheric C
which is in the isotope 14C for a plot of total amount. The plot seems
to be the source because its time axis covers 8000yr.  The y-axis is
Delta in parts per thousand, ppt, (The Delta is capitalized. A symbol that 
looks like %o is used for ppt.) and the data decline from values around 
80ppt to about 0, with a sudden dip at the end due to the Suess effect,
the release of old C by fossil fuel burning.

Crowley and North never really say that Delta is a proportion, but that
is because they expect that the reader knows that already. (They do this
a lot.) The use of delta (lower case this time) is customary for isotope
ratio work. For instance for oxygen isotope work, one sees delta 18O (ppt),
ie Fig 6.22 of C&N. delta is defined as 

           [14C]s/[12C]s
delta 14C=(-------------------- - 1)*1000
           0.95*[14C]st/[12C]st

where [...]s is sample concentration and [...]st is standard concentration.
The 0.95 is due to the standard used: most deltas, such as for 18O, don't
have it. Delta 14C (now capitalized) is the delta 14C projected back to
its assumed age and corrected for fractionation effects (using delta 13C) and
new versus old standards. The resulting Delta 14C is a best estimate
of the delta when the sample was formed. Delta is used to estimate the
changes in the atmospheric 14C reservoir for 14C dating work. (Broecker, 
Olson and Bird, 1959, Nature 183, 1582-1584.) It appears to be largely a 
proxy for the earth's magnetic field strength, and possibly for solar output, 
although this is far less certain. 10Be can be used to check for reservoir 
effects, such as the Suess effect seen in the graph.

This error illustrates the danger of bypassing peer review and technical
specialists and going straight to a technically naive audience. One of the 
purposes of peer review is to catch things like this. (And a moderately
competent reviewer would have caught it immediately.)

Paul Farrar
not an official spokesman