Long Term Certainty

Guest post by Stephan Lewandowsky

Who will win the AFL Grand Final in September?

No one knows, of course.

Human endeavours such as sports, markets, and marriages are notoriously difficult to predict.

Why, then, can climate scientists project global warming to the year 2100?

Part of the answer is that climate, unlike football, is a physical process. And physical processes can be predicted, sometimes with incredible accuracy. For example, the next total solar eclipse visible from Perth will be on the 31st of May 2068, and it will peak at 26.7 seconds past 11:54 local time. You can set your watch by that if you are still around in 58 years; the prediction is that accurate.

But if climate is a physical process that we can forecast, why can’t we now predict the weather during the Grand Final?

The answer is fascinating because it shows that short-term uncertainty can co-exist with long-term certainty. Suppose you visit the Burswood tonight and place a bet at roulette. Can anyone tell you what number is going to come up next?


Do you know what would happen if you kept betting for the next 10 years?


You’d lose a lot and the casino would win a lot. That’s why casino owners can sleep at night despite not being able to predict the next number at roulette—they can confidently expect to turn a handsome profit at the end of the year, which for the folks who own the Burswood was $280 million in 2009.

In much the same way we can be confident that the climate will be considerably warmer by 2100: That long-term virtual certainty is not affected by uncertainty about the weather during the Grand Final, in the same way that the Burswood’s profit is not affected by what happens during the next spin at the roulette table.

In confirmation, climate scientist Wally Broecker published a paper in 1975 that predicted the temperature rise by the end of the 20th century to within 1/10th of a degree. A quarter-century forecast that came within 1/10th of a degree! Today’s computer models are far more sophisticated and one would be ill-advised to ignore their forecasts for considerably more warming by the end of the century.

So is there no uncertainty about climate change?

There is uncertainty, but only in the way that there is uncertainty about what happens when you drive into a brick wall at 80 km/h. You might just get away with a few bruises and a concussion, but it is far more likely that you would break a leg or worse.

No one in their right mind would drive into a brick wall because the outcome is “uncertain.”

And no one in their right mind should delay action on climate change because we don’t know exactly how bad it is going to be. Whether it’ll be just bad or really bad or unbearably bad, we know that the climate is changing and that humans are causing it, with potentially serious consequences for us all.

So anyone who says that we shouldn’t act on climate change because of uncertainty is really inviting you to ride towards a brick wall at 80 km/h because it might not hurt. Are you feeling lucky? Or shouldn’t we better cut emissions in light of the uncertainty?

NOTE: this post is also being "climatecast" by Stephan Lewandowsky on RTR -FM 92.1 at 11.30 AM WAST today. It should air shortly after this post goes live so if you're reading this immediately (eg - you've subscribed to the SkS mailing list and just got this email), you can listen online via http://www.rtrfm.com.au/listen.

This 3-minute podcast was previously “blog reviewed” here on Skeptical Science, and I wish to thank those who contributed to improving this piece through their thoughtful and detailed comments. For new readers, please bear in mind that the podcasts are spoken and hence must be understandable by listening alone—though obvious, this is no trivial matter because it mandates simplifications that would be unnecessary in writing.

For those who want to know more about short-term uncertainty and long-term forecasting, I recommend “Chaos: A very short introduction” by Lenny Smith (Oxford University Press).

Posted by Stephan Lewandowsky on Wednesday, 18 August, 2010

Creative Commons License The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License.