Open Source Science

I have had enough brushes with eccentrics who think I am going to destroy the world with the Large Hadron Collider to appreciate that in scientific fields where many lives and dollars are genuinely at stake, pressures can be enormous and cool scientific discourse is hard to find and maintain. In such fields, pressures don’t just come from eccentrics and obsessives, but from smart and well-funded lobby groups. I have enormous respect and sympathy for those scientists who work in such areas, under intense commercial, political and moral pressures.

One such area is climate science. I’m writing this basically because the Institute of Physics, of which I am a Fellow, has waded in to the so-called “climate wars” recently; once with a statement before Copenhagen and more recently in “evidence” to the Science and Technology Committee.

The first IoP statement is a call to arms to physicists to help deal with anthropogenic global warming (AGW). The second advocates openness in science, and contains strongly worded criticisms of some scientists at the Climate Research Unit at the university of East Anglia, who famously had their emails stolen.

The reaction to the IoP statement was interesting, to say the least. Lobby groups who claim AGW is not happening rushed to publicise it, lauding the IoP’s credentials as a respectable and representative scientific organisation, and conveniently ignoring the IoP’s previous and subsequent statements that the scientific evidence for AGW is sound. Meanwhile, some of those on the other side in the climate wars leapt in to abuse the IoP. (Follow some of the links and threads from here to get a flavour, along with some decent criticism of what the IoP said.)

My level of expertise here is “scientist” not “climate scientist”, but in my opinion the IoP evidence betrays a pretty naive attitude; and in point 5 they overstep the mark by claiming “suppression” of data which the CRU have in fact published. To me (I don’t know who wrote it) it has the feel of being authored by someone who is firmly in the “anti” camp, then imperfectly moderated by others*. In general it was not of the quality I would expect from the IoP, which usually I find to be rather professional, measured and sensible. I much prefer the submission by the Royal Statistical Society.

Nevertheless, the IoP statement cannot be dismissed simply by pointing out that the response may have been influenced by energy consultants and the like.

Climate scientists bear a massive responsibility, which all scientists share in principle, but are not used to carrying day-by-day under huge pressure. Scientific data should be open as open as possible, science should be published in such a way that the analysis, and if possible the experiment, can be repeated by a third party. We don’t all meet these standards all the time, but the more at stake on our results, the greater the imperative. It’s pretty clear from this exchange that at least one early climate paper fell short of this. It looks to me like there was unreproducible analysis, and a deep reluctance to admit this**.

I can’t see anything in the emails (and I have read quite a lot of them) which discredits the main AGW results. But I do think that if scientists expect the world to change its economy and behaviour on the basis of their science, they really ought to have stable, reproducible databases and code. If they don’t have the funding to do this, more fool us all, and that needs fixing.

I was struck in particular by the “Harry README file” for some analysis code which came out. One thing we should have learned over the past years (e.g. from the continued success of linux in a world dominated by big software companies) is that open-source coding is a very effective way to produce robust and relatively bug-free code. If a “skeptic” finds a mistake, that strengthens the analysis. This actually applies to science in general.

Another area of science which operates under pressure is medicine. The level of irrationality shown by some conspiracy theorists (e.g. some of the anti-vaccine brigade), and the financial muscle of  “big pharma” and “big quackery” make this a highly politicised environment. Scientific critics face legal challenges (just as some try to apply to the climate scientists). The answer in general to the challenge seems to be more openness and more rigour. Ad hominem attacks on people who disagree with you do not advance science, and do not win over neutrals, at least not fellow scientists. Climate scientists should listen to their fellows with experience in this.

Since major global decisions depend on the climate data and the analyses, surely they should all be out there to benefit from the criticism of all. Valid criticisms lead to improvement, and openness is also the best way to despatch “zombie” invalid criticisms which continue to be touted by the unscientific.

Climate scientists are doing their work on behalf of us all. Is the human race smart enough to correctly evaluate, collectively, the threats we face? The level of emotion and abuse on both sides of the climate wars doesn’t bode well. To have a chance, climate scientists, and their colleagues across all disciplines, must publicly claim openness and true skepticism as a key element of the scientific process.

* Footnote: Some people whose views I respect disagree more strongly with the IoP evidence than I do. See here and here for example.

** Footnote: Before I call down hordes on my head let me say immediately that while in principle I think I could reproduce the analysis in all my papers, in practice I don’t have the time, and if I tried I might well find big problems blocking me from doing so. Actually I have published a paper which had a mistake in it, which I have owned up to (go on – guess). It did not change the conclusions of the paper (though in principle it could have done). I didn’t react in a hostile manner to the error being found, and I have not (so far) been vilified in the press or by congressmen for it. Science often doesn’t live up to the ideal, but it is usually “good enough”.


About Jon Butterworth

UCL Physics prof, works on LHC, writes (books, Cosmic Shambles and elsewhere). Citizen of England, UK, Europe & Nowhere, apparently.
This entry was posted in Climate Change, Politics, Science, Science Policy and tagged . Bookmark the permalink.

6 Responses to Open Source Science

  1. Pingback: Dreamt of in your philosophy « Life and Physics

  2. spotthelemon says:

    I think you’re pretty much right, The Deniers main tool is picking away at any weakness and the CRU people should have understood this and been working to the highest standards all the time. I consider it possible that one or more IOP people were pretty annoyed with them, I was, for effectively supplying the Deniers with easy ammunition and therefore had no qualms about a submission with “skeptical” undertones.

  3. mrmcgoo says:

    One of the problems in the history of science is it must always go begging for funds.

    Anyone with money can buy results, not always correct results.

  4. Pingback: Materials: Engineering, Science, Processing and Design – Study … | Materials Engineering Addict

  5. Pingback: Open Source Science « Life and Physics | Open Hacking

  6. I agree broadly with what you say here, and I too quite like the Royal Statistical Society submission.

    I do feel the IOP submission unfairly criticised the CRU for not holding to a standard which pretty much no one else keeps, casting general doubt on the whole field of climate science (without justification and in contradiction to it’s stated policy) and providing an “analysis” of the leaked e-mails which is simply wrong.

    I think your analysis of how the IOP submission came to be made is largely right, I do wonder why the Environmental Physics Group wasn’t asked to contribute though (since they seemed to have led on the earlier statement).

    It’s interesting to note that in this months Physics World envelope comes an admission that the IOP submission on STFC may not have been entirely to everyone’s satisfaction and that methods for compiling submissions need to be improved.

    I say a bit more about my view of climate science here:

Comments are closed.