20100318_EC

Source: The Economist

URL: http://audiovideo.economist.com/?fr_story=337236bd018b5e33ce1c9b7c695ae8780000fcdc

Date: 18/03/2010

Event: Sir Brian Hoskins being interviewed about climate models by The Economist.

People:

• Sir Brian Hoskins: The head of the Grantham Institute for Climate Change

The head of the Grantham Institute for Climate Change on the dangerous influence of headlines on science.

Introduction: In the past 40 years, studying the climate has gone from being a slightly eccentric backwater of physics, to being possibly the most politically engaged science in the world. Brian Hoskins watched that all happen, first of all as an academic meteorologist, now in a broader role as the head of the Grantham Institute at Imperial College, where he's responsible for bringing together ideas about climate and energy solutions to climate problems from across the campus.

We wanted to talk about the uncertainties in climate science, and what those mean in the political arena, so we had him in to tea.

Interviewer: What's striking is the degree to which uncertainty about the whole question of climate change has come back onto the agenda, as a result of a number of things. I can't remember, have you actually been an author on the IPCC?

Brian Hoskins: I was a review editor in the last IPCC, so that's the first time I've been involved.

Interviewer: What do you make of what's come out recently about the problems with the other working groups?

Brian Hoskins: In terms of the science, it's amazing how we've taken steps backwards, really on the basis of, perhaps a couple of paragraphs in 3,000 pages. And that just seems amazing - that a scientific argument based on evidence, and a couple of bits of that evidence may or may not have been expressed rather too strongly, or not checked enough, and the fact that that's cast uncertainty on the whole seems amazing to me. And the - we have such confidence that we are doing something dangerous to the climate system, that by adding greenhouse gases, it's going to warm - purely the question is: how much? I think every climate scientist of whatever persuasion, would go along with that.

Interviewer: But you talk about this great confidence, which I've observed in climate scientists. Do you worry that that does actually lead to a certain form of bias? Bob Watson, who was once the chairman of the IPCC, pointed out recently that he thinks that one thing that worries people is that all the possibly relatively minor errors that people have been pointing out in the last report, have all pointed in the way of making things look worse, not better. And there might be some sort of subconscious level, in which that is actually a process, there's a sort of like group think within climate scientists, that does see it as always seeing the worst thing, and also possibly always thinking that the worst thing is probably the best thing to highlight.

Brian Hoskins: There is a danger, and we must be aware of that. I think in general, IPCC has done a remarkably good job of actually giving a range of uncertainty - and this is full of uncertainty. I think when it has been used, usually not by the scientists involved, then certainty comes over in a manner which I don't go along with, and I've often squirmed over, and I, for instance, think it could be a good thing to have people only involved in two IPCCs, maybe. Because then, that would - like US President, it gets over the danger of someone getting too involved and too stuck in there, and this danger of an in-group, which is a real danger. I'm not saying it's happened, but I think there can be a danger.

Interviewer: But - the perception has been given, I mean, people have said at times, the science is settled.

Brian Hoskins: Living science in the headlines, and in the policy - and the headlines and in the politics as well, is actually a very difficult - and can be dangerous - game. We talked - we have said that going for the headlines by actually perhaps taking the more extreme rather than the middle of the range. And it can mean that people find it difficult to highlight the uncertainties because those uncertainties will be taken by others to say "You know nothing. And therefore we can't base policy on what you're talking about." I sometimes wish economists were a little more open about their uncertainties, and yet we base an awful lot on their predictions. So we should be open about our uncertainties but also that should not be a basis for saying "Well, we can't use what you're talking about."

Interviewer: Tell us the war story. How bad were the climate models when you started out?

Brian Hoskins: Ah, they were pretty lousy, and they're still pretty lousy, really. They were terrible. They had difficulty representing the general sort of high pressure over the poles, and then the general westerly winds in middle latitudes, say. They were approximately right, but -

Interviewer: - and this was how far back?

Brian Hoskins: You go back to the - well, I'm afraid I go back to the '70s, and that's when - yeah. They were really developing and they didn't have an interactive ocean. And so they were really a very limited system and a very limited ability. So now, compared with that, they're fantastic. But still, it's a very young science, as I say, and it's got a long way to go. And I think we know the sort of course we've got to go down. It's always a mixture, really, of what is predictable and what isn't.

Interviewer: Do you have an experience of, from your own research, of a time when a model did something that you really couldn't understand?

Brian Hoskins: In general, I don't think we found those models are wrong, although when we were developing high-resolution models at one time, the westerly winds were getting better and better represented up to a certain resolution. Then started to get worse again as we went to even higher resolution. And that suggested we didn't understand what we were doing.

Interviewer: Did you find a way around it?

Brian Hoskins: Well, we found that, perhaps, that some of the processes that we thought we were representing properly, were not quite so well represented. You learn something - you learn more when a model goes wrong, probably, than when it actually seems to behave properly.

Interviewer: Isn't that the exact problem of climate science being so political, that as scientists, yes, you do want things to go wrong? But for climate scientists to stand up and say "Our models are wrong" is a problem politically? The model leads to a political solution - leads to a political response that many climate scientists wouldn't like.

Brian Hoskins: Climate science has got to, like other sciences, proceed by things - by things going wrong. By not understanding things, and then that proceeds provide a stronger basis for the future, so we have to be able to say we don't understand this, or this model doesn't seem to be behaving properly. Because that's how we improve. It's never going to be that a model is right or wrong, it's going to be aspects of the model are working well or not working so well. And all the time, we will be looking at those models to find out what needs to be improved for the next generation of models. And in some ways, climate scientists thought we have to show that we're always right and we're going to solve the problem. But actually we'd be doing ourselves out of a business if we said we'd solved the problem, so we've got to highlight uncertainties and things we're not doing very well, in order to say, well this is where we go next, to actually improve our understanding.

Interviewer: Brian, thanks very much, it was very nice having tea with you.

Brian Hoskins: I enjoyed the tea and the conversation, thanks.

Afterword: Despite the trouble some of his colleagues have been having, Brian Hoskins is clearly still excited by his climate science, and enthusiastic about what he can do to make that science relevant to policymakers. He'll need both those qualities if he, and the rest of climate science, are to keep the role of honest brokers that they need.