Thursday, 26 November 2009

The CRU 'hack'

Last week, someone posted a great deal of data (emails and programs) from the Climate Research Unit (CRU) at the University of East Anglia. This has caused much wailing and gnashing of teeth from within the relevant organisations. The leak has been picked up on by climate change researchers, sceptics and media organisations all over the world.

Firstly, there is the question about whether the data is really a 'hack'. Whenever confidential information gets out into the wild, organisations often respond immediately by saying that it was 'hacked' (another example was the email earlier this year that proved that people in the British Government were trying to smear members of the opposition). The reason for this is simple: a hack makes it look as though the organisation in question was a victim of a crime.

Yet in the case of the Smeargate emails, it was much more likely that someone passed them onto the person who publicised them. Likewise, it is eminently possible that the CRU data was collated and leaked (accidental or on purpose) by someone within the CRU or the University of East Anglia. Yet they automatically jump on the 'hacked' bandwagon as it makes them look like victims. This also diverts attention away from what the data leaked actually contains.

Much of the media comment has been on various small claims within the leaked data and emails - things like the hiding or deletion of data. This piecemeal information is easy to refute by claiming that the quotes were cherry-picked.

There is more important information coming out from these emails, however. Freedom of Information (FoI) requests to the CRU were made by several researchers, including some from Steve McIntyre, the Canadian man who runs the now (in)famous Climate Audit website. He is one of the individuals who has been trying to reproduce the results that the researchers have published.

Now, I am not a scientist (tm), but I do know a little about the scientific process (if you want a reader-friendly description of parts of the scientific process, see Ben Goldacre's excellent book 'Bad Science'). But I do know that reproducibility of results is vitally important. Any major research should be reproducible by others - this issue killed off Martin Fleischmann and Stanley Pons' cold fusion research.

So far, the scientists at the CRU appear to have been saying something like: 'our research papers have been peer-reviewed, so it is okay'). Well, that shows that they do not know, or do not care, about how the peer-review process works, or it's well-known deficiencies. Basically, peer-review is done before publication, where peers check the published paper for gross inaccuracies, unjustified or contradictory findings. What it does not do is try and reproduce the results given.

After a paper has been published, you can respond with comments to the journal that published it. This process can prove to be rather exasperating for the responder.

Therefore the peer-review process is just a sticking-plaster (albeit an important one) over the scientific process. What is also needed are other, independent teams of scientists to replicate, reproduce and validate important results. There have been many papers in the past which have been peer-reviewed and published, and yet the underlying research and science has subsequently been found to be incorrect.

So Steve McIntyre and others have been trying to reproduce the results. So far, there have been some notable discoveries; for instance Steve McIntyre discovered that some NASA data was incorrect. Although the people involved deny this effects the trends, surely discovery of *any* inaccuracy in science is a good thing, and McIntyre should be congratulated.

He has also been trying to reproduce some of the CRU results. For this, he needs the data they used and, ideally, the algorithms they applied to that data. Yet the CRU have been blocking him at every turn.

The attitude of the CRU can be seen in the following snippet from Professor Phil Jones, the director of the CRU:
Even if WMO agrees, I will still not pass on the data. We have 25 or so years invested in the work. Why should I make the data available to you, when your aim is to try and find something wrong with it.
The answer is simple: because it is science, and not your own personal fiefdom.

Now, the ClimateAudit website has Willis Eschenbach's email correspondance about FoI requests, along with some of the emails that they generated within the CRU. it is not a pretty read. At best, some people at the CRU appear to have broken the laws on FoI, and at worst have not been doing science.

What is obvious is how combative the whole thing has become. There is obviously no love lost between McIntyre and the people at the CRU, yet that does not mean that the CRU can ignore him and the other researchers.

Therefore we have are some famous results, one of the cornerstones of climate change research, which are unverified. We have the organisation that made the results doing everything possible to prevent others from verifying them. Frankly, this situation makes the people within the CRU look like terrible scientists, and brings scorn down on the University of East Anglia.

What is worse, they have been trying to subvert the FoI laws to prevent the data being released, including apparent deletion of emails that have been requested under the FoI. They have cultivated the people dealing with FoI requests within the University of East Anglia to ignore requests from people who post on the Climate Audit website.

Climate change is rightly seen as being a massively important issue; for this reason the science behind it should be thoroughly open for replication and study. Personal fiefdoms of information is intellectually and morally wrong, especially when the results from that data is so important.

The people at the CRU may not like McIntyre et al; but these emails show that their behaviour has been reprehensible. It makes any reasonable person wonder if they actually have (or ever did have) the data and algorithms that can be used to replicate their findings. This could mean anything of the options below:
  1. They never had the data and/or algorithms, and they do not want to release them;
  2. They have lost the data and/or algorithms, and cannot reproduce the results they got;
  3. They realise that the information will show embarrassingly large holes in their findings;
  4. They are just being petty and small-minded.
  5. They genuinely believe they cannot release the data and/or algorithms. This does not appear to be the case from their internal (leaked) emails, which detail continued obfuscation and delay. Besides, climate change is such an important issue that such data *should* be open, especially so may years after the original studies were performed.
This is not science. It is a travesty.

No comments: