I was supposed to be running my next post on correlations today (my goal is to run one such post a week until the series is finished), but recent political developments have left me in a foul mood. If you don't know what I'm referring to, you can see my previous post here. I don't trust myself to put the final touches on a technical post while in a foul mood, so instead, I'd like to discuss a different, but somewhat related, matter.
I am interested in paleoclimatology. I am no Steve McIntyre (who runs Climate Audit), as has been pointed out by several of my critics, but I like to think I am reasonably knowledgeable on the field. I certainly know more about it than the average person, and when it comes to fine technical details of individual proxies/reconstructions, I might even be more knowledgeable than a number of "experts." SO today, I'd like to discuss a problem in the field.
Years back, a man named Warwick Hughes requested data regarding the modern temperature record from a scientist named Phil Jones. Jones (in)famously responded:
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.
Which is about as unscientific a response as one could imagine. This e-mail came up again recently on Twitter when a person said:
He was referencing this (in)famous response in regard to a request for data I made. However, this request I made was not of a climate scientist. In fact, it was not of a mainstream scientist at all. I had requested data from one Jennifer Marohasy. She had recently published a paper which exmained various paleoclimatoloogical records and drew conclusions based upon a type of computer modeling known as artificial neural networks.
As a person who has long been interested in paleoclimatology, my first reaction to the paper was to look for the dat aused in the paper. I won't name names, but a number of people who quickly jumped to criticize this paper didn't bother to do this simple step, making a variety of strange and unfounded claims because they had no idea what the paper actually examined. I took my time. I read the paper, and I looked for a data archive which would contain the data used for the paper.
I didn't find one. As such, I went to the blog post Marohasy wrote about the paper in an attempt to ask for the data. The post had been closed for comments. Undeterred, I wrote this comment on a more recent post Marohasy had published:
I apologize for this being off-topic, but comments on the last post closed quickly and I didn’t get a chance to ask a question there: Has the data for your recent paper been published? If not, would you mind uploading it so people have access to it when considering/replicating your paper?
Marohasy responded with an inline comment:
In our recent paper in GeoResJ we took temperature series already published in the peer-reviewed literature: this is ‘the data’ underpinning our work.
We then decomposed these series into their component parts, before inputted them into an artificial neural network (ANN), which is a form of machine learning.
This is ‘the data’ underpinning our study: it is already published as part of the six studies that we cite/that we used/that form the foundation of our work.
The results from our ANN models are subsequently clearly displayed as charts in our new paper in GeoResJ.
I don't believe telling people the data you used it out there somewhere for them to find is an adeqaute answer. As such, I wrote a response. I would like to link to my follow-up comment and copy its text, but I cannot, The response i posted was:
This comment was on Marohasy's blog for over 12 hours, during which time I received a response from a user who goes by the name mack. Amusingly, mack went on at length about how I should be making complaints about the University of Queensland (which he called the Queensland University) refusing to share data. If you don't know what that is ironic, well, let's just say there is some history between me and that university (including threats of having me arrested).
However, when I went to respond to mack, I was surprised to find out his comment had been removed. So too was my response to him, as well as my comment which triggered his response. That's correct. The comment you see in the screenshot above was deleted. Even worse, it was deleted without any remark by Marohasy or explanation of what was supposedly wrong with it.
While that was going on, there was also a matter of private communications. I thought it might help if I contacted Marohasy directly, so I sent an e-mail which said:
My name is Brandon Shollenberger, and I've been a critic of paleoclimatic reconstructions in the past. If I need to prove any credentials, I think my short eBookk explaining why it is reasonable to call Michael Mann's original hockey stick a "fraud" should suffice:
I recently came across your latest paper, which was kindly shared for free, and I wanted to look into the analysis it performs to try to better understand what it says. At part of the process, I would like to ensure I am working with the same data you are. As such, I was wondering if you published the data as you used it somewhere, and if not, wuld it be possible to obtain a copy?
My personal preference is to have data and code that exactly replicates authors' results so their can be no confusion/uncertainty, but at a minimum, I like to make sure I am working with the *exact* same data set as other people. Many different data sets for the same data exist in the paleoclimate world, and I see no reason to have obstacles based on miscommunication/uncertainty.
Thank you for time,
I sent this e-mail at about the same time I posted my second comment on Marohasy's blog. I would normally wait longer before making anything of the fact I haven't received a response (and most likely send a follow-up e-mail to be safe), but combined with the fact Marohasy deleted my public comment responding to her statement regarding her data, it seems Marohasy is unwilling to publish the data she used for her paper.
Now, there are a lot of things which could be said about a "skeptic" refusing to share data while secretly deleting comments which ask for it. I'm not going to say them today. While fundamentals of beliefs and policies regarding the access of data are important, I'd like to focus on a more direct issue. To do so, I'll quote Marohasy's paper:
Published graphical temperature proxy reconstructions were digitized using UN-SCAN-IT software. Table 1 gives a summary of the temperature proxy reconstructions used for analysis. The digitized time-series were then examined by spectral analysis using AutoSignal software, applying the Parametric Interpretation and Prediction tools with Fourier Transform analysis.
The paper relied on the results of six previous studies. Normally, I would call into question the idea of "skeptics" relying on the work of studies "skeptics" had previously criticized, For today, I'll ignore that issue. Instead, I'd like to focus on a much simpler issue.
Marohasy's paper says the data for these previous studies was "digitized using UN-SCAN-IT software." That is, the authors didn't use the numerical data for the studies it examined. Instead, it looked at figures in those studies and used computer software to generate numerical values based off the lines in the charts.
That seems like a less than ideal process to me, but the issue which stands out to me is... why won't Marohasy share her data? The one response she gave to me before deleting my comments indicated the data she used was available in other studies. However, she estimated data points for those studies via a digitization process. Even if I could obtain the actual data for those studies, how could I possibly be expected to come up with the estimations Marohasy came up with for her paper?
The answer is I can't. For whatever reason, Marohasy has decided not to share the data she uses for her study. When asked for the data, she refused to publish it. When a follow-up request was made via e-mail, she ignored it. When a follow-up request was made via a public forum, she deleted it. For whatever reason, Marohasy, a supposed global warming "skeptic," is refusing to share her data so it can be examined by other people.