The latest news is from a new blog post by Susan Crockford that includes two points of considerable interest – her letter of complaint to the journal, and a release of emails (Polar-Bear-Gate?) between polar bear scientists. Here I’ll discuss these two issues separately.
“Grab your popcorn”, Susan says. This brand might be appropriate.
Crockford’s complaint letter
Susan Crockford’s letter was sent to Scott Collins and James Verdier, editor in chief and senior editor of the journal Bioscience. The journal is published by Oxford Academic, which is part of OUP. They have an ethics policy here, which states that
“Whilst striving to promote freedom of expression wherever possible, OUP aims to avoid publishing anything that harms the reputation of an individual, business, group, or organization unless it can be proven to be true. We take all possible measures to ensure that published work is free of any text that is, or may be considered to be libellous, slanderous, or defamatory.”
It will be interesting to see whether Collins and Verdier adhere to this policy.
The letter is quite detailed and carefully written. It deals with the false and misleading claims of her lack of expertise in the paper, and some specific lies in the paper, such as the unsubstantiated claim that “Crockford vigorously criticizes, without supporting evidence, the findings of several leading researchers who have studied polar bears in the field for decades”. She points to her preprint paper, which is full of evidence. Look at her other posts, for example this recent one, and you can see there is plenty of supporting evidence. She gives several other examples where her criticism is backed up by detailed evidence.
The letter ends with
These allegations are untrue, defamatory and malicious, but in addition, the failure to mention my Ph.D. and my recent scientific critique constitute a falsification of my expertise and work output. In addition, the purported scientific analysis is shoddy and the language used is reprehensible.
I formally request that you retract this paper.
The Polar-Bear-Gate emails
In her letter, Crockford discusses a set of emails from 2012 and 2014 between polar bear scientists. These were acquired via a FOIA request (submitted by someone else, and then passed on to her). The emails have been transcribed and the email addresses removed.
A key point in the emails is that they show another lie in the Harvey et al story. We’ve already established that the claim of a bunch of AGW denier blogs is a lie. But the emails show that the other side of the picture, of a happy family of scientists all in agreement, is also a lie.
Some of the emails involve Dr. Resit Akçakaya, one of the scientists responsible for the IUCN Red List of endangered species. Other participants include polar bear researcher Øystein Wiig. Steven Amstrup (“Steve”) is the polar bear academic who is regularly criticised by Crockford and who lied about her in a recent press interview. Kristin Laidre works on polar bears and other arctic mammals and has worked with Wiig. “BN” models are Bayesian Network models, if that helps. SPSC is a committee within IUCN. PBSG is the Polar Bear Specialist Group, including Dag Vongraven, Amstrup, Laidre and Wiig.
Here are some excerpts from the emails, in date order.
Wiig to Akcakaya, 12/3/2012
I understood from Dag that you at the outset are a bit skeptical to the use of Bayesian methods in such an assessment.
What, skeptical? Surely not? Is he a denier?
Akcakaya to Wiig, 20/3/2012
I did voice some concerns. But these were not about Bayesian methods in general, but rather about the specific paper (Amstrup et al. 2008). After reading the paper again, and soliciting other opinions, here is a summary of my concerns…
So my first concern is that this approach will make the assessment too much dependent on expert opinion…
Furthermore, I am concerned about the degree to which uncertainty would be compounded through the model with so many uncertain parameters…
Thus, it seems like this is mostly a probability based on expert opinion, and not what is meant by “quantitative analysis” in criterion E.
I think that this paper is Amstrup et al 2008. It makes dramatic claims of ‘extirpation’ of polar bears by mid-century, which is one of the claims that Crockford is critical of on her blog. Criterion E is presumably one of the conditions that must be fulfilled for red-listing. This guy Akcakaya is sounding a bit like Crockford, in his scepticism of Amstrup’s work!
Akcakaya to Wiig, 20/10/2012, replying to an email enquiring about the red-listing procedure:
Then, the crucial issue is the population reduction that will result from that amount of habitat loss. If there are not data on this, and therefore it has to be based on expert opinion, I would think that it would be more appropriate to call the population reduction “suspected” instead of “inferred.”
Akcakaya to Wiig again, 20/10/2012
I wanted to remind you that the SPSC had a large number of reservations about the BN modeling of Amstrup et al.
I would therefore recommend against relying heavily on that paper.
Wiig to Laidre, 7/5/2014
I discussed all this with Steve and he was not very happy about Resit’s view.
Laidre to Wiig, 8/5/2014
I wanted you to know I had lunch with Resit today. It was very informative and we had good talks about the polar bear Red List issues…
– Overall he is very negative to the Amstrup et al. model and says it has very little value for anything, or the Red List.
– He said it is difficult to fill out just one conditional probability table, but having thousands all merged together in the BN is just crazy and tells you nothing (he actually said “only Bayesians would accept ‘a belief’ to be a “probability!”)
– He said he felt if PBSG used the BN model again this time for the new Red List it would very likely NOT be accepted and the polar bear would be listed as “Data Deficient”, which he felt would be a shame and also not a good thing politically.
Ouch! Amstrup’s model has very little value for anything. Just crazy. And if we admitted that data was deficient that would not be “a good thing politically”.
Between the 3 of us, this is a difficult thing to navigate in the PBSG because Amstrup plays a big role and is very sensitive about his model.
He’s very sensitive about his model. Is this starting to sound familiar at all?
We as a group will have to decide what we put forward for polar bears .. even if some are not happy.
But surely, all the top scientists in the field are in agreement aren’t they? Fig 2 of Harvey et al showed that. There’s a 97% consensus! Surely it couldn’t be the case that climate scientists are telling us in public that they all agree with each other when in reality they disagree? Again, does this seem familiar?
There’s more. Resit suggests to Kirstin, over a second lunch, that if they want to show that polar bears are declining, they need to get some data. This almost seems to be a new concept to Kristin and Øystein.
We need to find some population rate data that can be related to change in population size.
What? Really? We need to find some data? Why can’t we just rely on our expert opinion and computer models?