Friday, November 05, 2010

Should we be skeptical about climate change?

I’m not going to argue that our global climate is changing (for – as far as Humans are concerned – the worst) or that humans are causing much of this change.  It is.  They are.  That’s all I want to say on the matter here. 

Some people say we should be skeptical about this conclusion and they’re quite right, we should.  But our skepticism should be well-targeted.  It should be at the root of climate science.  Does a slice of data support a hypothesis?  Well, that’s what we need to check.  It’s because this skepticism is present at the root of climate science that it’s safe for those of us who don’t know much about that science to believe what the consensus is telling us. 

Skepticism is less useful at the level of whether AGW is happening or not.  This is not to say we should be discouraged from asking questions (we shouldn’t) but that any skepticism at this level has to refer to the lower levels anyway.  Eventually, you’d need to present an argument that showed lots of data doesn’t support the hypotheses of AGW. 

In other words, just saying “I’m not convinced by these scientists and their data and their models” is not enough.  It’s no better than just picking a conclusion and manufacturing arguments to support it out of thin air.  In that sense, it is the very opposite of skepticism.

Some such methods and attitudes can be seen in the comments here (an excellent video by the way):

notably from correspondent Pag.  Pag doesn’t understand how science is actually done and this ignorance colours his views.  It’s important to note that Pag doesn’t deny AGW, but is skeptical.  Unfortunately, he targets his skepticism incorrectly as described above.  I’ll pick out a few points to show what I mean.

Probably a bad idea to throw that out there, but here's why I'm skeptical about the whole anthropomorphic global warming thing [….] The gist of it is that the science of global warming is attempting to predict the future decades in advance, something that nobody has ever been able to do with any kind of reliability, in any field, as far as I know. What's worse, the whole thing is two parts politics for one part science, which doesn't help its reliability.

Of course there are fields where predictions can be made over such timescales and much larger ones.  Astronomy is the most obvious example, but there are many others.  However, the point is about what is being predicted, with what accuracy etc.  It’s not necessary to predict the temperature on July 14th 2085 in order to predict increasing heat or in large weather events or whatever.  Predictions always have error bars and much of the work being done in climate science is concerned with reducing those error bars.  We can already predict more or less what’s going to happen, but more data and better ways of analysing it can predict it more accurately.  And can more accurately predict the effects of any intervention that might be attempted, such as reducing CO2 production.  As for the 2 parts politics charge, to begin with, this is a flat, unsupported assertion and the further charge that politics inevitably casts doubt on the conclusions of climate science is not explained.  However, if we look at the situation more carefully, we can see that there’s another category error: Pag is confusing the level of uncertainty banded about in the media (Global warming is happening!  No it isn’t! etc.) with the daily work of climate science (These data support hypothesis X!  My algorithm shows that the results are probably not significant!).  But more on this later.

First, the scientists themselves. Are they unbiased? I'm unconvinced. On one side, some scientists are paid by the oil industry who gains by having oil be perceived as safe. On the other side, some scientists are paid by environmental organizations who get their money from people who believe that the environment is in great peril (let's face it, if people stopped believing in global warming, Green Peace and others like them would lose a huge amount of donations). Neither side is particularly neutral or trustworthy.

This might be the case if the greater number of climate scientists were directly funded by organisations with an agenda.  In fact, most are funded through government.  If government have a bias on the issue, I’d expect it to anti-AGW since this stance would mean they didn’t have to make unpopular decisions.  But again, this broad argument doesn’t really get to the heart of the misunderstanding.  Pag rightly points out that individual scientists – regardless of funding – are subject to bias and likely to look more favourably on results that confirm that bias.

This is true.  We are all subject to bias.  Science is organised such that this bias is reduced, but it would be naive to accept that this removes it entirely.  We should expect to see some bias in the scientific record.  However, this where Pag gets it wrong, because – perhaps counter intuitively – this doesn’t make the conclusions of climate science suspect.

This is because of what the daily business of science actually is.  What we see of it is for the most part what appears in the media.  This view is badly skewed at best.  We tend to see headlines announcing a ‘breakthrough’ or which claims to ‘cast doubt’ on some important or supposedly cherished previous finding.

But what we see in practice isn’t like this at all.  A scientific paper typically documents something relatively unimportant in the general scheme of things, but which supports or refutes a minor hypothesis.  It’s a tiny piece of a big picture.  It’s certainly possible that a researcher might prefer methods that support their preferred hypothesis rather than refute it, but that hypothesis will be at the level of “this method is good for analysing this kind of data” rather than “AGW does/doesn’t exist”.  It’s by no means clear that a slight bias of this kind will have a noticeable effect on the question of AGW.

Pag goes on to explain that climate scientists are more biased than others because they are emotionally involved:they are passionate environmentalists and have a vested interest in maintaining the status quo.  His basis for this assertion is unclear.

Second, is the model reliable? I'm sure scientists are doing their best to make it as reliable as possible, but is it good enough? We're talking about simulating the whole earth for 100 years -- no small challenge.

This isn’t what climate scientists are trying to do at all.  They don’t try to simulate the whole earth, which would indeed be a tall order.  Instead, there are lots of models of different aspects of the climate, which show different kinds of trend, with differently-sized error bars with different types of input and which are effective over different time spans, with differing levels of precision etc.  These generate evidence that purports to support or refute some hypothesis.  This is far from the vision of some grand model of the universe with scientists feeding data in one end, cranking the handle and a perfect prediction of the climate’s behaviour comes out the other end. 

We can’t, practically speaking, answer questions like that.  At least, not yet.  But we can ask questions like “if X has this kind of relationship with Y, what will it mean for Z?” and then we can look for data that such a relationship exists and that our model does indeed predict the effect on Z with some accuracy.

Pag makes a similar point about data being unreliable.  Pag has a poor understanding of what data is and where it comes from.  He or she doesn’t understand that climate scientists know that their data is imperfect and know ways to correct for it.  People like Pag always think they’re the only ones who understand any of this and that scientists are somehow unaware of the difficulty.  He prates on about how the models don’t admit to any uncertainty in results, which is just plain wrong and accuses the models of not taking into account future perturbations (the example he cites is the possibility that everyone will start using electric cars).

“I'm well aware that various scenarios are considered for global warming simulations,” he admits, “but they're very limited.”

Really?  He/she manages to conclude this without knowing a single thing about the models or how they are applied.

So, as I see it, global warming predictions come from biased scientists using flawed models based on partial and imprecise data to predict the future. And they ask us to trust them with billion-dollar decisions based on these predictions. I'm sorry, but I remain skeptical.

And this is at the heart of why skepticism is more properly levelled at the low levels of climate science rather than the broad strokes.  Pag’s objections to models and data aren’t valid.  Pag’s ignorance of these things and how they are used results in badly-targeted skepticism and wasted effort.

So be skeptical about science, but make sure you focus your efforts where they can do the most good.  Are there people out there faking data?  It’s not unlikely.  Is there a set of data or some conclusions that are relied upon too highly by many researchers?  Could be.  Skepticism about climate science are more properly applied at this level.  If we get that right, our conclusions can only become more accurate over time.

No comments:

Post a Comment