Over the last couple of months there has been much blog-viating about what the models used in the IPCC 4th Assessment Report (AR4) do and do not predict about natural variability in the presence of a long-term greenhouse gas related trend. Unfortunately, much of the discussion has been based on graphics, energy-balance models and descriptions of what the forced component is, rather than the full ensemble from the coupled models. That has lead to some rather excitable but ill-informed buzz about very short time scale tendencies. We have already discussed how short term analysis of the data can be misleading, and we have previously commented on the use of the uncertainty in the ensemble mean being confused with the envelope of possible trajectories (here). The actual model outputs have been available for a long time, and it is somewhat surprising that no-one has looked specifically at it given the attention the subject has garnered. So in this post we will examine directly what the individual model simulations actually show.
By Stefan Rahmstorf, Michael Mann, Ray Bradley, William Connolley, David Archer, and Caspar Ammann
Global cooling appears to be the “flavour of the month”. First, a rather misguided media discussion erupted on whether global warming had stopped, based on the observed temperatures of the past 8 years or so (see our post). Now, an entirely new discussion is capturing the imagination, based on a group of scientists from Germany predicting a pause in global warming last week in the journal Nature (Keenlyside et al. 2008).
Specifically, they make two forecasts for global temperature, as discussed in the last paragraphs of their paper and shown in their Figure 4 (see below). The first forecast concerns the time interval 2000-2010, while the second concerns the interval 2005-2015 (*). For these two 10-year averages, the authors make the following prediction:
“… the initialised prediction indicates a slight cooling relative to 1994-2004 conditions”
Their graph shows this: temperatures in the two forecast intervals (green points shown at 2005 and 2010) are almost the same and are both lower than observed in 1994-2004 (the end of the red line in their graph).
New York Times, BBC News, Reuters, Bloomberg and so on), because of its seeming contradiction with global warming. The authors emphasise this aspect in their own media release, which was titled: Will Global Warming Take a Short Break?
That this cooling would just be a temporary blip and would change nothing about global warming goes without saying and has been amply discussed elsewhere (e.g. here). But another question has been rarely discussed: will this forecast turn out to be correct?
We think not – and we are prepared to bet serious money on this. We have double-checked with the authors: they say they really mean this as a serious forecast, not just as a methodological experiment. If the authors of the paper really believe that their forecast has a greater than 50% chance of being correct, then they should accept our offer of a bet; it should be easy money for them. If they do not accept our bet, then we must question how much faith they really have in their own forecast.
The bet we propose is very simple and concerns the specific global prediction in their Nature article. If the average temperature 2000-2010 (their first forecast) really turns out to be lower or equal to the average temperature 1994-2004 (*), we will pay them € 2500. If it turns out to be warmer, they pay us € 2500. This bet will be decided by the end of 2010. We offer the same for their second forecast: If 2005-2015 (*) turns out to be colder or equal compared to 1994-2004 (*), we will pay them € 2500 – if it turns out to be warmer, they pay us the same. The basis for the temperature comparison will be the HadCRUT3 global mean surface temperature data set used by the authors in their paper.
To be fair, the bet needs an escape clause in case a big volcano erupts or a big meteorite hits the Earth and causes cooling below the 1994-2004 level. In this eventuality, the forecast of Keenlyside et al. could not be verified any more, and the bet is off.
The bet would also need a neutral arbiter – we propose, for example, the director of the Hadley Centre, home of the data used by Keenlyside et al., or a committee of neutral colleagues. This neutral arbiter would also decide whether a volcano or meteorite impact event is large enough as to make the bet obsolete.
We will discuss the scientific reasons for our assessment here another time – first we want to hear from Keenlyside et al. whether they accept our bet. Our friendly challenge is out – we hope they will accept it in good sportsmanship.
(*) We adopt here the definition of the 10-year intervals as in their paper, which is from 1 November of the first year to 31 October of the last year. I.e.: 2000-2010 means 1 November 2000 until 31 October 2010.
Update: We have now published part 2 of this bet with our scientific arguments.