U.S. and Global Temperatures: a correction

Dr Gavin Schmidt, a climate modeler at NASA and a principal author of the Real Climate weblog, has emailed me to point out an error (mine) in my review of Ian Plimer's Heaven + Earth.

I said that I had learned from Ian Plimer that NASA had reversed it's claim that the ten years following 1995 were the hottest ten years of the century when Steven McIntyre showed that the record belonged to 1934. But, as Dr Schmidt points out, 1934 was the hottest year in the GISS records only for the United States, not in the global GISS records. He goes on to accuse Prof. Plimer of 'lying' to his readers about this, and other, matters.

I am grateful to Dr Schmidt for pointing out this error, which I have corrected. But I still share Ian Plimer's amply-documented conclusion that there is no reason for alarm about the slight warming that undoubtedly took place over a fifty year period from the 1940s.

First, Dr Schmidt's email:

I tried to post this on your blog, but got a "You are not authorized to
perform this action" response. Regardless of whether this gets posted or
not, you should know that Pilmer is not a trustworthy source of
information.

Gavin Schmidt

=====================================


Many of the things that you "didn't previously know", are actually just
not true, as a moment's research would have shown. 

For instance, the GISTEMP analysis is freely available online: 
http://data.nasa.giss.gov/gistemp 
and shows very clearly that, indeed, that the top ten warmest years have
occurred since 1995, and that 1934 is nowhere near the warmest (anomaly
-0.05, compared to the last ten years range of +0.32 to +0.62).
http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt

What you are reading instead is a grotesque misquotation and distortion
of an extremely minor correction involving only a set of US temperatures
that had less than 0.002 degC impact on the global mean.
http://data.giss.nasa.gov/gistemp/2007/

You are being lied to - and you can easily do some 'due diligence' of
your own to prove it to yourself. 
http://www.realclimate.org/wiki/index.php?title=Ian_Plimer#Books

============

In fact, I had checked Ian Plimer's assertion at the NASA GISS site before posting my review and had realized that the data was U.S. data only. So I am guilty of the same elision that is in Plimer's book (page 99). In his defense, Prof. Plimer was discussing the different prominence given to findings that support the case for alarming warming and those that contradict it. My elision was the worse one. I can only plead that I mentioned the fact in passing and did not pay sufficient attention to the implication.

I do not agree with Dr Schmidt that I am "being lied to" by Prof Plimer. I am aware, of course, that a number of other reviewers have also pointed to statements in Heaven + Earth that they say are wrong. Some have turned these (claimed) errors of fact into hysterical ad hominem claims.

I would have been surprised if, in such a long book as Plimer's, with such a variety of argument (and a bit of invective as David Stockwell notes), there were not some errors. I cannot agree, however, that this means Plimer is 'lying' to me or anyone else. I conceded in my review that his purpose was rhetorical (that is, to persuade). But his approach is thoroughly respectable in the sense that he punctiliously cites sources, even to the point of exaggeration. If he's wrong in fact or interpretation, he'll be found out as Dr Schmidt has shown. That's what scholarship is all about. It's not at all what we expect of a liar.

But Dr Schmidt has a second good point, too, about doing the 'due dilgence'. The impact of the U.S. record on the global temperature record was debated at length in 2007. I was unaware of this at the time I wrote my review. I've caught up now:

• Here's Steven McIntyre's report and a series of comments discussing the impact of the changes
• Here's the response on RealClimate.
• Here's a round-up of the 2007 debate by Warren Meyer (check the 'update' links at the bottom of his post, too).

My conclusion

The U.S. surface station temperature record, from which this debate about 1934 springs, now seems to be, for the most part, questionable. No doubt it will be improved. But where does this leave the more interesting question of the impact on the global temperature record? I agree again, here, with Dr Schmidt that the U.S. temperature record should make no more difference than the United States' 2% coverage of the surface of the earth implies. But it's very likely that the U.S. temperature record does have a disproportionate influence on the global record.

The prortion of GISS stations outside the United States was estimated to be only about 50% (for the other 98% of the earth's surface) in 2005. Recent data indicates that the proportion has dropped dramatically over the past thirty years as non-US stations drop out of the network. This is said to be part of the explanation for the divergence between the GISS records and the global satellite records that show still more clearly than GISS the fall in temperatures since the start of this century (GISS shows this trend, too).

At the least, the disproportionate representation of the U.S. stations in the GISS global database is a reason for legitimate concern that the errors in (and adjustments of) the U.S. data would permeate to the rest of the GISS global averages temperatures.

Footnote: I do not normally reproduce emails. But Dr Schmidt makes clear that this email transmits comments that he intended, but was unable, to make on the story itself. So I have presumed that he would not object to it being published.

Tags for this entry: climate data temperature statistics