Winning A Climate Bet

  • Date: 13/01/12

Predictions, Neils Bohr once said, are difficult, especially about the future. They are even more interesting however, when there is money at stake.

In December 2007 I wrote what I thought was quite a straightforward article for the New Statesman pointing out that it was curious that when so many voices were telling us that global warming was out of control, and that the global warming effect dwarfed natural fluctuations, the global annual average temperature hadn’t increased for many years. I wasn’t promoting any particular point of view just describing the data. The New Statesman jumped at it.

It caused quite a storm resulting in an Internet record number of comments that were complimentary by a large majority, although there were some less than supportive remarks. It evidently also caused quite a fuss in the offices of the New Statesman. Realclimate.com responded with, in my view, an unsatisfactory knock-down of my piece based on trend lines, which I had expected. Trend lines, especially of indeterminate length in the presence of noise, can tell you almost anything, and nothing.

The New Statesman environment correspondent Mark Lynas chipped in eventually with, “I’ll be blunt. Whitehouse got it wrong – completely wrong,” after saying he was initially reluctant to comment. He reproduced Realclimate.com’s trendlines argument and accused me of deliberately or otherwise setting out to deceive. It was a scientifically ignorant article which subsequent events, and peer-reviewed literature, emphasise. Moreover, when I asked New Statesman for redress against such an unnecessary, and in my view unprofessional insult, they declined, and stopped answering my emails. In doing so they missed out on an important, though perhaps inconvenient, scientific story.

More or Less

To my surprise interest in my article was worldwide, and eventually the BBC’s radio programme “More or Less” got in touch. The programme is about numbers and statistics and they set up a series of interviews. You can hear the programme here.

Almost at the last minute the programme-makers came up with the idea of a bet. It was for £100 that, using the HadCrut3 data set, there would be no new record set by 2011. It was made between climatologist James Annan and myself. His work involves analysing climatic data and validating climate models. He accepted enthusiastically as he has a perchant for taking on ‘sceptics.’ The presenter said that if the global temperature didn’t go up in the next few years, “there would be some explaining to do.”

Later today, January 13th, “More or Less” returns to the bet, which I am pleased to say I won, though I note that this bet, or its conclusion, is not yet mentioned on Annan’s Wikipedia entry despite his other climate bet being discussed.

Writing shortly after the wager was placed James Annan said he believed it was a fairly safe bet, though not certain, as the trend since the current warming spell began, around 1980, was upward (showing those same trendlines!) He drew a straight line from 1980 to 2007 and projected it forwards concluding that sometime over the next few years HadCrut3 would rise above its highest point which was in 1998 (a strong El Nino year.)

The problem with this approach is that it destroys all information in the dataset save the gradient of the straight line. In climate terms 30 years is usually held to be the shortest period to deduce trends (though shorter periods are used often if the trend deduced is deemed acceptable) but that is not to say there is not important information on shorter periods such as volcanic depressions, El Nino rises and La Nina dips. Then there are the so-called, poorly understood decadal variations.

My view was that the information in the dataset was important, especially if projecting it forward just a few years when natural variations were clearly dominant. Looking at HadCrut3 it is clear that there isn’t much of an increase in the 1980s, more of an increase in the 1990s, then there is the big 1998 El Nino, followed by no increase in the past decade or so. It therefore seemed far more likely that the temperature would continue what it had been doing in the recent past than revert to an upward trend, in the next few years at least.

My approach was to listen to the data. The approach taken by James Annan was flawed because he didn’t. He imposed a straight line on the data due to theoretical considerations. I always wonder about the wisdom of the approach that uses straight lines in climatic data. Why should such a complex system follow a straight line? Indeed, the rise of HadCrut3 is not a straight line, but the past ten years is, and that in my view is very curious, and highly significant.

Why, I wonder start the linear increase in 1980? Obviously the temperature starts rising then, but why not start the straight line in 1970? The answer is that the temperature is flat between 1970 and 1980. It seems illogical to take notice of flat data at the start of a dataset but totally ignore it at the end!

When a record is not a record

During the recent interview for “More or Less” James Annan said that had other temperature databases been used he would have won. This is a moot point that also strongly reaffirms my stance. In NasaGiss 2010 is the warmest year, with a temperature anomaly of 0.63 deg C, only one hundredth of a degree warmer than 2005, and within a whisker of 2007, 2006, 2002, 2001 and 1998. Given the 0.1 deg C errors even Nasa did not claim 2010 as a record. Technically speaking 2010 was slightly hotter because of a strong El Nino. Otherwise, NasaGiss shows hardly any increase in the past decade.

During the “More or Less” interview the question arose of extending the bet to “double or quits” for the next five years. I was game for it with a proviso. Betting against a record for ten years raises a higher possibility that there might be a statistical fluctuation than betting for five years. Because of this I would like to see two annual datapoints, consecutively more than one sigma above the 2001 – date mean level. After all, that is the minimum statistical evidence one should accept as being an indication of warming. James Annan did not commit to such a bet during the programme.

It just has to start getting warmer soon.

Back in 2007 many commentators, activists and scientists, such as Lynas, said the halt in global temperatures wasn’t real. It is interesting that the Climategate emails showed that the certainty some scientists expressed about this issue in public was not mirrored in private. Indeed, one intemperate activist, determined to shoot my New Statesman article down but unable to muster the simple statistics required to tackle the statistical properties of only 30 data points, asked the University of East Anglia’s Climatic Research Unit and the Met Office, to provide reasons why I was wrong, which they couldn’t.

What was true in 2007 is even more so in 2012. Since 2007 the reality of the temperature standstill has been accepted and many explanations offered for it, more than can possibly be true! We have seen predictions that half of the years between 2009 and 2014 would be HadCrut3 records (a prediction that now can’t possibly come to pass) which was later modified to half of the years between 2010 and 2015 (likewise.) The Met Office predict that 2012 -16 will be on average 0.54 deg C above the HadCrut3 baseline level, and 2017 -2021 some 0.76 deg C higher. Temperatures must go up, and quickly.

So how long must this standstill go on until bigger questions are asked about the rate of global warming? When asked if he would be worried if there was no increase in the next five years James Annan would only say it would only indicate a lower rate of warming! Some say that 15 years is the period for serious questions.

We are already there

In a now famous (though even at the time obvious) interview in 2010 Prof Phil Jones of the University of East Anglia confirmed that there was no statistically significant warming since 1995. There was an upward trend, but it was statistically insignificant, which in scientific parlance equates to no trend at all. In 2011 Prof Jones told the BBC that due to the inclusion of the warmish 2010 there was now a statistically significant increase between 1995 and 2010. Since 2011 was cool it doesn’t take complicated statistics to show that the post 1995 trend by that method of calculation is now back to insignificant, though I don’t expect the BBC to update its story.

The lesson is that for the recent warming spell, the one that begins about 1980, the years of standstill now exceed those with a year-on-year increase. It is the standstill, not the increase, that is now this warm period’s defining characteristic.

The nature of the anthropogenic global warming signal is that, unlike natural fluctuations, it is always additive. Sooner or later, it is argued, it will emerge unambiguously, perhaps at different times in different parts of the world, but it must emerge. Some argue that by the time it does it will already be too late, but that is another debate.

James Annan is keen on a “money markets” approach to forecasting global warming, and bemoans the reticence of so-called climate sceptics to put their money where their mouth is! I hope that his early-stage financial loss won’t be too much of a setback and a deterrence for potential investors, not that I will be among them.

Now that I am joining the ranks of those who have made money out of global warming (or rather the lack of it) I wonder where the smart money will be placed in the future.

Feedback: david.whitehouse@thegwpf.com

 

Recent Popular Articles


We use cookies to help give you the best experience on our website. By continuing without changing your cookie settings, we assume you agree to this. Please read our privacy policy to find out more.