Tuesday, May 24, 2005

Yet more betting on climate with World Climate Report

(previous installments can be found here and here)

World Climate Report (apparently another fossil-fuel-funded lobby group) is a blog run by Cato Institute member Pat Michaels and a couple of others.

Back in December 1998, just as temperatures were coming off the peak of the hottest year on record, WCR offered the following bet:

That 1998’s extremely warm temperatures were largely confined to one calendar year makes the annual record high temperature 1998 has established quite a difficult one to break.

If we were of a betting sort (and there are some nasty rumors going around that we are), we would be willing to wager that the 10-year period beginning in January 1998 and extending through December 2007 will show a statistically significant downward trend in the monthly satellite record of global temperatures.

Surely such a wager should sound interesting to those who think the planetary temperature will increase several tenths of a degree during that period.

No reasonable offers refused...

Unfortunately, I only found out about this recently. Nevertheless, I asked them what their current position was.

Chip Knappenburger replied. He admitted that at present, they stood to lose the bet (there is a minscule negative trend from Jan 1998 to April 2005, but certainly not a significant one - note that significant was their qualifier, not mine). He tried to sound bullish about how he expected the MSU temperature to drop in the future....but was not prepared to bet either on the terms of the original wager, or propose a new bet starting from the present. "No reasonable offers refused" indeed. Have they learnt their lesson? Time will tell.

Here I've plotted the MSU data from Jan 1996 to the present (April 2005 at time of writing), together with the overall trend in black starting in Jan 1998 (-0.007 degrees per year) and how this trend has evolved over time (the red lines show trend after 1, 2, 3, 4, 5, 6, 7 full years from Jan 1998). I guess they might have been feeling confident for the first year or two. I included the two previous years to indicate to what exent they cherry-picked their starting point - the first data point is the one at +0.5. Of course it doesn' t matter what bet they offer if they won't honour it!


Geoffrey Allan Plauche said...

It's odd that the Exxon Secrets website doesn't say how much money they get from Exxon and other corporations. I've seen environmentalists make the same ad hominem attacks on Pat for his affiliation with Cato, but Cato apparently only gets a mere 2% of its funding from corporations. That Exxon Secrets site is useless and disingenuous.

If your case is so strong, why do so many enviromentalists (including yourself) rely so heavily on ad hominem attacks against skeptics' source of funding? It certainly isn't an example of rational discourse.

James Annan said...


About 95% of my post is focussed on climate science, only 1.5 lines refer to the fact that WCR is a coal industry front. The latter fact is the best explanation I can think of as to why they got the climate science so spectacularly wrong, but the wrongness of the climate science itself is the main point.

I was thinking of updating this graph. Any guesses as to whether the trend will be more "statistically significant downward"?

Chip said...


Just a quick update on how things would have worked out for WCR’s proposed bet. The RSS TLT data show a non-significant (p=0.28) overall downward trend of about -0.006C/yr from Jan 1998-Dec 2007. The UAH TLT apparently has a spurious warming during recent months that will likely be corrected for in some fashion in the future, so I don't think that I can yet accurately assess the trend from that record. But, I would guess that it too, will prove non-significant.

So, we would have lost our bet, but, actually, not by all that much...another 3-4 months with anomalies like the last two months would push the downward trend to being significant at the p<.05 level.

Not that we know whether that will happen or not (although the tendency of the monthly data over the past 3 years suggests that it is not out of the question) just that (as luck would have it?) it didn't turn out that we were ludicrously off base.


James Annan said...

Hi Chip,

Thanks for the update. I must admit I am a little bit surprised that the trend stayed so low/negative (depending on the choice of data) - although only slightly, cos I never did any serious calculation on that detail. Perhaps you now wish you had tried to negotiate a bet with me back in 2005 :-)

Roger said...

Interesting exchange, and it is nice to see the follow up. FYI, more somewhat related here:


Chip said...


Now comes an update to my previous update:

It turns out there likely is NOT a warm bias at the end of the UAH MSU TLT data (instead there may be a cold bias in the RSS data). This means we were much further from winning our bet than I initially thought. Using the most recent UAH TLT data, the January 1998 to December 2007 (linear) trend is a (non significant) postive trend of 0.04C/decade. This trend is actually not far from what you may have expected given a continued warming following the relaxation from the El Nino highs coupled with the 1998 start date. Perhaps, the post 2002 period is somewhat interesting, but it is still too early to tell.



Tom said...

Your ad hominem attacks suggesting that those who disagree with you do so for corrupt reasons, suggests an emotional approach which diminishes your credibility. Why not rely on your data analysis. The fact that your ad hominem attacks are 1 or 2 percent of your comments does not diminish their significance.

Tom said...

The ad hominem attacks suggesting that those who disagree with you are corrupt suggests an emotional level to your approach which diminishes your credibility. Why not rely on your data analysis. The fact that personal attacks are a small percentage of your approach does not take away from its significance.

Tom Connors