“But you said the ice was going to disappear in 10 years!”

Nearly two decades ago, some scientists predicted that summer sea ice in the Arctic would disappear “soon.” These predictions were mentioned by Al Gore and received a lot of press. However, they did not gain wide acceptance in the scientific community and were quickly refuted. Unsurprisingly, this still comes up a lot. It’s time to dig deeper into what happened and why…

Not surprisingly, climate opponents bring up “failed predictions” from the past to bolster their argument that nothing needs to be done about climate change. [It is equally unsurprising that they don’t bother to mention the predictions that were skillful, but let’s not dwell on that!]. For a long time, his favorite so-called “failed prediction” was that there was a consensus that a new ice age was imminent in the 1970s (a topic we have covered many times), but more recently he has turned to Al Gore’s supposed prediction that “Arctic summer sea ice would disappear” within a short number of years. This has it all: the message ‘But Al Gore!’ Instinctively, a combination of Al Gore and the scientific community, it sounds suitably apocalyptic, and of course, Arctic summer sea ice hasn’t gone away (it’s only down about 40%):

What did Al Gore really say?

Going back to December 2007, immediately after the shocking decline in sea ice that summer, Gore gave his acceptance speech for the Nobel Peace Prize he had received jointly with the IPCC. In it he said:

Last September 21, as the northern hemisphere moved away from the sun, scientists reported with unprecedented anguish that the North Pole ice sheet was “falling off a cliff.” One study estimated that it could disappear completely during the summer in less than 22 years. Another new study, to be presented by US Navy researchers later this week, warns that this could happen in as little as seven years.

What was he reporting on?

This was a truthful report. The first study (I think) refers to a commentary piece in EOS (or perhaps a preprint thereof), which pointed out the poor performance of climate models in tracking Arctic sea ice loss, and made an expert estimate that summer sea ice would disappear around 2030. The second (forthcoming) refers to a fall 2007 presentation at AGU that would be given by Wieslaw Maslowski, who at the time ran one of the highest resolution ice models available. However, their prediction was not based directly on their ice model, but rather on a linear extrapolation of the ice volume from their model:

One might sensibly ask why a prediction made in 2007 only made it into a review article in 2012, despite having been very advertised At the moment? We’ll get to that.

Gore continued to reference Maslowski’s prediction at least until 2009.

In the years that followed, a few other people jumped into the sea ice prediction game using similarly unorthodox methodologies. Chief among them was Peter Wadhams, emeritus professor at the University of Cambridge. Wadhams (and a group calling itself the “Arctic Methane Emergency Group” (AMEG)) began showing graphs of ice thickness extrapolated from the University of Washington study. PIOMASA model:

Even without being a sea ice expert, one might question some of these methods: naive adjustments to noisy data that extrapolate out of range, the strange fact that the same methods applied to the extent or area data gave very different times of ice-free conditions, and, most obviously, the lack of any physical model for the future state. Sure, the standard climate models (CMIP3 at the time) used in the scenarios were behaving too conservatively, but ignoring them completely…?

I don’t remember if I was at Maslowski’s talk at AGU 2007, but I remember seeing him present similar results at least a couple of times. And even if he was not present, his results were widely discussed among relevant scientists in multiple workshops. From what I remember, the reviews were quite negative.

[Update 9/30: Axel Schwieger in the comments reminds me we had a guest post from his group making this very clear in 2012!]

What is the physics behind your prediction?

In 2014, the Royal Society organized a workshop on Reducing sea ice in the Arctic. I was invited to give a talk on paleoclimatic perspectives on sea ice change, modeling and methane. It should be noted that Peter Wadhams was present and presented a graph very similar to the previous one. If you look carefully in the wayback machine you can find some of the audio recordings of the meeting and, specifically, if hear to the question and answer period of his talk, you can hear me ask [43:00] whether there was any physical basis for such extrapolation. The answer was no. [As an aside, this was one of the first climate workshops that really embraced Twitter (as it was then) as a means of broader dissemination, though this wasn’t appreciated by this particular speaker!]. Interestingly, Wadhams remained confident that 2015 (at this point is less than a year away) would be ice-free in summer.

To be clear, I am not claiming any specific brilliance by being skeptical of these predictions. Almost no one in the field was unconvinced by these extrapolations from the initial presentation at the 2007 AGU meeting onwards. The reason these predictions never made it into a peer-reviewed publication? I imagine it was the difficulty in finding reviewers who found these methods credible.

Lessons learned?

Science is very competitive and scientists fiercely protect their independence. For them, agreeing on even one thing requires a great effort. Therefore, there will always be a variety of opinions and methods on any topic and people who will hold firmly to them. The desire and culture of assessments (like the IPCC) arose specifically to synthesize that broad range of individual scientists into a more coherent and better-balanced assessment that a large majority of experts agree with.

In retrospect, it’s clear that some people were fooled by chance, giving too much weight to the movements and not the long-term trend (which, to be honest, is a pervasive problem):

One could look back at this episode and what has been made of it since and declare that scientists should have somehow prevented Maslowski and Wadhams from presenting their ideas or speaking to recovering journalists or politicians. But that is absurd: no scientist or group of scientists has that power, nor would they even want it. Alternatively, other scientists could have loudly voiced their skepticism of these results and made better evaluations. But both things happened. Some even went further and started betting against extreme predictions (quite successful in retrospect). For serious people, interested in serious screenings, that might be enough. However, all this will be (and is) ignored when someone wants to obtain a laugh line on Fox News.

If people are really interested in what the scientific community thinks, the evaluated projections of the IPCC and the like are their best bet. It can be useful to examine the range of individual projections or opinions, especially in rapidly changing situations, but it is very difficult to discuss them in public and be immune to subsequent distortions.

References

1. W. Maslowski, J. Clement Kinney, M. Higgins and A. Roberts, “The future of Arctic sea ice”, Annual Review of Earth and Planetary Sciencesvol. 40, pp. 625-654, 2012. http://dx.doi.org/10.1146/annurev-earth-042711-105345