It's official: I'm not-boring! How do I know?! HvS has interviewed me as part of his ongoing anthropological studies of "interesting" scientists. You can read the interview here, and the more nicely formatted PDF for you to download and keep is here. It is an amazing honour, although I know I'm as nothing on the "interestingness" scale compared to the likes of Yamagata-sensei, a previous interviewee.
91 comments:
Wow.
Cheers!
"In order to publish in high impact journals the numbers must keep getting bigger and the outcomes more scary."
A little green in honor of St. Paddy's Day? :)
Evidently James has also been 'interviewed' for the UK's Daily Mail this weekend:
http://www.dailymail.co.uk/news/article-2294560/The-great-green-1-The-hard-proof-finally-shows-global-warming-forecasts-costing-billions-WRONG-along.html
Ho-hum...
In order to publish in high impact journals the numbers must keep getting bigger and the outcomes more scary.
Oh bullshit. Off the top of my head, Schmittner et al. 2011 in Science, Tung and Zhou 2013 in PNAS, Huntingford et al. 2013 in Nature Geoscience.
I hope you were quoted out of context, otherwise, shame on you.
Hello James,
Please tell me that Jules was either misquoted or being facetious when she allegedly said,
"In order to publish in high impact journals the numbers must keep getting bigger and the outcomes more scary."
My apologies for my candor, but anyone following the literature knows that unsubstantiatedassertion is demonstrably false and inane.
I'll wait to see if the quote is indeed correct before spending time demonstrating why that quote is false. What I find bizarre is that even your own publications in journals with high impact factors shows the statement to be blatantly false.
What gives?
By "high impact journals", I was referring to the career-making glossy magazines, Science and Nature. The shame of it is not that these magazines are looking for stories that sell, but that getting published in one is so good for scientists' careers.
A few counter-examples hardly proves the point, and what you are not seeing is the rejections of papers on the grounds that the results are not sufficiently extreme.
"Not sufficiently extreme" might overlap somewhat with "not very interesting" (to a broad audience). Although exact proportions aside, there does seem to be no lack of not terribly interesting stuff in those two pubs. And what about the less-glossy but also career-boosting PNAS?
A few counter-examples hardly proves the point
Assuming you mean "refute" the point, they obviously do. That's kind of what counterexamples to sweeping generalizations exist for. :)
If your general complaint is that Nature, Science (and possibly PNAS) tend toward eye-catching results, sure. Everyone probably agrees. This is something that people complain about in every discipline.
That's not what you said, though. The idea that you can't get published unless you have a 'big number' (i.e. a high end warming result) or 'scary' paper is demonstrably false.
The idea that papers must be increasingly more 'alarmist' over time is just... silly.
The big journals have a tendency towards surprising results. Sometimes that means a high ECS, sometimes it means a lower one. Sometimes it means a lot of SLR from Greenland, sometimes it means much less SLR from Greenland.
The common denominator isn't "big" numbers in terms of climate impacts or of "scary" outcomes, but of headline-generating results. Things that people will talk about, irrespective of which direction they happen to point.
I have to say that I am really surprised that you think this is even remotely debatable.
-------------------------------
Anyway, I enjoyed reading the profile overall, and I don't mean to be a killjoy. It's hard not to focus on the one part of something you find absurd rather than the 9/10ths that you don't...
Some scientists would appear to be taking an entirely different approach to getting recognition.
Others will doubtless continue having to be dragged kicking and screaming into the Century of the Fruitbat. :)
Thingsbreak et al: rest assured, you most definitely have killed my joy.
Obviously I should have written extreme rather than big, but I'm really shocked you've all tried to make this into such a big deal. All I was saying was that having these magazines out there who advertise the extreme results is rather unhelpful in an already highly politicised field.
I think I'll go back to just posting photos, and let James do all the writing from now on. He's so much more flame retardant.
Jules, the problem is that your statement is going to go to the top when we google Jules Hargreaves.
OTOH more flowers, bees and old cars are quite nice things and Eli don't give a damn, James is a stick.
PNAS does the best at selecting interesting science; AAAS's Science used to do fairly well. Nature? Well...
Jules, the difficulty is that there is a widespread view that there are as yet some pretty nasty unknown knowns and known/unknown unknowns lurking out there. Until that sense subsides (not real soon, and indeed the various model failures relating to key large-scale climate system features only tend to enhance the feeling), results that seem like they might point to one or another of those are just plain going to get more attention. Many such papers indeed will turn out to be wrong to whatever degree, but given the context (ref. the Tobis diagram) I don't think it's a mistake to promote them. Maybe your physical intuition tells you otherwise, but if so it's not news to you that many other scientists (and science journalists and activists) take the other view.
So as a general matter, papers that e.g. contribute to narrowing Charney sensitivity aren't going to qualify for the marquee (unless, of course, one is well-connected enough to overcome that tendency).
Probably all of this is a justification for bitterness, but please don't go that route.
Hi Jules,
You now say,
"All I was saying was that having these magazines out there who advertise the extreme results is rather unhelpful in an already highly politicised field."
No that is not what you said, and we all know that extremes are found on both ends of the spectrum. Regardless, the quoted text above may have been what you intended to say, but the words you actually said and that are now forever out there are:
"In order to publish in high impact journals the numbers must keep getting bigger and the outcomes more scary."
That is not the same as the first quote. To say what you originally stated was a poor choice of words would be a gross understatement.
You may be "shocked" about TB's response (I disagree, others not commenting are equally unimpressed), then please allow me to be equally shocked that you would be so careless in your choice of words on such an important issue. Your unsubstantiated comment is not helping with this polarizing issue.
I am curious, exactly what metrics did you have in mind that allegedly must keep getting bigger?
The news-worthiness filter pushes editors of the glossy magazines towards the extremes. In climate change, big numbers are considerably more exciting than small, so they are more readily published than the small.
On top of it all, these publishing houses are taking excessive amounts of taxpayers money through their sky-high paper charges and subscriptions.
Scientists get such a big career boost from these publications, that they think it is worthwhile.
All in all I think it is a very unhealthy situation.
Well I proof-read jules' interview (responses) before she sent them back, and didn't see anything particularly outrageous in them. I understood "high impact" to be basically synonymous with the glossies (not the proper mainstream science journals) but I can see how that might have been misunderstood by some. And while "exciting" does not necessarily always equate to "more scary", there's certainly quite a correlation there.
The Huntingford et al paper is a case in point, actually - the original Cox et al rainforest collapse, in Nature, was based on a model that was well-known to have a dry bias in the Amazon, such that the modelled forest was barely surviving under present day conditions. And yet it seems to have taken over a decade for a comprehensive and promiment debunking, and then in a lower impact journal...
"And yet it seems to have taken over a decade for a comprehensive and promiment debunking, and then in a lower impact journal."
James - it would be really reassuring if the rainforest itself were convinced of that debunking, but it isn't.
Exceptional Amazonian drought typically has occurred once-a-century. We've now had 3 in the last 15 years. And the region is now showing extraordinary seasonal variation in rainfall. Indeed, the exceptional 2005 drought impacted parts of the Amazon that persisted until the 2010 drought - which was even worse. See Saatchi (2013). The closing comment in the abstract of their study:
"The result suggests that the occurrence of droughts in Amazonia at 5–10 y frequency may lead to persistent alteration of the forest canopy"
Continued warming of the tropical Atlantic Ocean will draw moisture away from the Amazon over the all-too-important wet season (Southern Hemisphere summer), and the hydrological intensification of ENSO will likely give wetter wet seasons, but drier (and therefore hotter) dry seasons. Given the mass die-off of trees from the last two droughts, and the long-lingering after effects, the potential collapse of the Amazon rainforest is by no means debunked.
I certainly do hope that trees are mysteriously able to access more nitrogen and phosphorus to support the hypothesized carbon fertilization effect that the carbon cycle models project. That would be be very helpful.
I have to admit, I read the "high impact"-quote as it was intended - and I guess most will do so. Not least because I know a number of colleagues who would use very similar formulations. Reading it differently suggests ... a cultural filter? (I am tempted to write "malevolence" instead of cultural filter.)
However, I think Eli may have a point though not yet.
James, Cox et al. (2013) with Huntingford as co-author appeared in Nature just a few weeks before Huntingford et al. appeared in Nature Geoscience. They made similar points re the earlier Cox paper, so should both have been in Nature?
But re Huntingford et al., co-author Simon Lewis tweeted: "Important to remember what's not in the models - direct impacts of humans, logging, fires etc. Likely important feeds." And while it's probably asking too much of the models to nail those droughts, they're starting to seem like not a coincidence.
Thanks for that longer response, Rob. I was going to, but would have done a poorer job.
ob, what was most bothersome about the original was the assertion of an increasing trend in bigness and scariness, not just that things have to be big and scary. That's a major difference in meaning.
you most definitely have killed my joy.
Well, I sincerely apologize. In hindsight my criticism should probably have occurred in proportion to the amount of content in the interview I disagreed with.
A lot of agreement and praise mixed with a single sharp disagreement would have been both fairer and probably easier to take seriously on your end.
Apologies.
Hi Jules and James,
Like TB, I should have said up front that I agree with much of what you both say on the science, and there was probably only one statement in the interview that annoyed me and others. TB's previous posts explaining the reasons for the frustration are better than mine. Unfortunately, we are in a PR war and must choose our words carefully. This was a lapse, but of course nobody is perfect.
For the record, I suspected that you were referring to Nature and/or Science. Regardless, I still think, despite your latest posts, that the original statement was unjustified and in essence doing exactly what you are blaming Nature of doing.
Anyhow, I think we all agree with what Jules was trying to communicate, so time to move on :)
@JA:
The Huntingford et al paper is a case in point, actually - the original Cox et al rainforest collapse, in Nature, was based on a model that was well-known to have a dry bias in the Amazon, such that the modelled forest was barely surviving under present day conditions. And yet it seems to have taken over a decade for a comprehensive and promiment debunking, and then in a lower impact journal...
This was my brainfart, as I had meant to refer to Cox et al. 2013, which is published in Nature like the earlier paper you refer to. However, as I said, I was going off the top of my head, and couldn't recall who the lead author was of that paper (and having read the Nature Geoscience paper more recently, had it more clearly in mind).
Albatross et al:
Ah... I see what the problem is. You are misunderstanding the nature of these interviews.
They're appearing in the AGU atmosphere newsletter and as Hans states on his site, they are, "interviews with eminent atmospheric scientists on their subjective views about the state of science, its developments in recent years, and the perspectives for the future."
See that word "subjective"? Since you aren't me, and it is basically a character interview, that you agree or approve of the content is irrelevant. I gave my present thoughts on the matters I was asked about based on my own experiences. I think to have done anything else - to try and write some absolute truth - would have been dishonest, and even worse, boring. Furthermore I couldn't have done it in the time allotted, as I'm definitely no expert on male/female roles/ratios, so would have had to have done a lot of research to answer those questions "correctly".
You are misunderstanding the nature of these interviews.
They're appearing in the AGU atmosphere newsletter and as Hans states on his site, they are, "interviews with eminent atmospheric scientists on their subjective views about the state of science, its developments in recent years, and the perspectives for the future."
At least for my part, I can't plead ignorance there. In fact, the nature of the interview was a main source of my disbelief and annoyance at the one line that I seized upon (which was, admittedly, the only part out of a long interview I found troubling).
Climate science has its problems. One of those is not, contra a lot of denialist claims, a conspiracy to suppress all dissent. Rather, the field suffers from a dearth of serious criticism at the expense of a bizarre proliferation and propagation of obscenely laughable bullshit masquerading as dissent (that in turn gets rebutted, which fuels conspiracies of persecution).
One of the biggest offenders of propagating such bullshit at the expense of legitimate criticism is Von Storch. His latest Lüdecke turd left at the doorstep of CP speaks for itself.
The objection to adding demonstrably false fuel to the "conspiracy against skeptics" fire is, I hope, a little more understandable in this context.
---
The above is not in any way meant to excuse my needlessly confrontation and accusatory tone earlier. It was inartful and undeserved. I rather enjoyed the interview as a whole, and it was poor form on my part to let my problem with your comment, Von Storch, and the bullshit about suppressing dissent, get the better of my commenting manners.
I hope you aren't being serious about taking this as a reason to engage less rather than more on the blog here. I enjoyed hearing your perspective and, as much as enjoy JA's posts, would enjoy hearing more from you on the science end.
Again, my apologies.
Nice interview! Your perspective on Japanese culture is always interesting. (I spent a week in Japan once, enough to feel the foreignness, but not to get the more refined sense of it that I get from reading you.) I noticed your mention of a recent piece in nature about sex bias; do you have a link?
First, my interview series is meant to bring forward also those eminent atmospheric and climate scientists, who are to pushing forward for being in the limelight. So far 17 interviews, among them Alan Robock and Gabi Hegerl. Half of the questions are usually the same in all interviews, another half are specific to the interviewee.
I wonder what thingsbreak means with "Lüdecke turd left at the doorstep of CP". I have never done anything with Lüdicke, not written about him, had not any business with CP in recent years - but maybe thingnsbreak will explain what this supposedly is.
This person seems to be one of the crowd who know everything better than anybody else, scientifically and normatively. Sure, they should present their opinions, also in their appalling and bullying language hidden behind silly anonymity, but then - why taking them seriously? Let them bark.
But being serious, we should all thank Jules for having voiced her views and experiences in such a frank and open manner. Jules, you definitely have something to say.
After heaving read the various comments here, I am impressed and dismayed about such a comical group of shouting and posturing males confronting a female, who has supposedly said something stupid. And, unfortunately, Jules responded as the white-male peer group presumably hoped for: " think I'll go back to just posting photos, and let James do all the writing from now on. He's so much more flame retardant.".
Would it not make more sense to consider the possibility that she may have a point, which is better - albeit different - from yours?
Of my 17 AGU interviews, 7 were with females: I am trying to get the more quiet climate researchers, and almost all females are in this group, to the forefront, not only the "loudspeakers", who are usually white and male (like me). The interviews are all available from my web page - together with 12 other interviews with members of the Hamburg center of excellence for climate science plus a few more with individuals such as Klaus Hasselmann or Walter Munk. The idea is to let people tell their story
The printed interview is always authorized in detail by the interviewee partner.
Should read "NOT pushing forward" istead of "to pushing forward". Sorry
Albatross wrote
"Hello James,
Please tell me that Jules was either misquoted or being facetious when she allegedly said ...". Why did you not ask herself, Albatross - why did you ask her husband? This smells of an allegedly long gone attitude - husbands are responsible for their wifes. What type of group is this - 1950s?
Are you a budgie in a cage, not an albatross?
ob, thanks for mentioning the "cultural filter". Maybe Klimazwiebel has some impact.
HvS: You can't actually devine their gender or their skin colour, but pseudonymns are generally bad things as they enable everyone to say things in public that they otherwise would not. I'm quite sure I would find it very easy to be just as nasty, so I don't give myself the chance!
I'm reminded of the suggestion that someone made to me a while ago, that authors of papers should all be anonymous, but that the reviewers should always be named. Of course it wouldn't work well in some cases (where a wee postdoc is trying to point out the holes in the work of a powerful professor), but I think it's an interesting idea nevertheless.
@Von Storch:
a crowd of males has jumped on Jules for her frankly voiced opinions. The comments are worth to be read, being a good case of how determined males deal with a dissenting female, who in the end reacts with "ok, I will no longer disturb you boys in your playing in your kindergarden".
I find this characterization more than a bit puzzling. I can speak only for myself, but the issues I was discussing had nothing to do with gender. Moreover, I am not privy to the genders of all of the participants involved, though perhaps you are.
I actually agreed a great deal with the comments on gender in the interview.
If I have somehow been unknowingly engaging in a group exercise in misogyny, I of course apologize profusely.
Why did you not ask herself, Albatross - why did you ask her husband? This smells of an allegedly long gone attitude - husbands are responsible for their wifes. What type of group is this - 1950s?
While I can only speak for myself, there are a number of reasons I can think of why someone might address a comment here to James rather than Jules. It might be because they see the name of the blog and assume that James is responsible for the approval of comments and thus will be the first to see them. Or perhaps, as happens on other blogs all the time, people reply to the blog authors who comment the most (as occurs, for example, when people respond to posts at RealClimate by other authors by addressing Gavin Schmidt). Not being psychic, I can't speak to the actual reason. But I try not to assume bad faith without sufficient reason to do so.
Gender bias is a pervasive problem throughout society, and is especially troubling and counterproductive in some areas of science. I recommend some interesting discussions on the issue that occurred in the wake of some comments on the Ecolog-L listserve (this blog post is a good one to start with).
This issue is something I do my best to fight in my personal and professional life, and I am aghast that I might have contributed to it here, however unwittingly.
I wonder what thingsbreak means with "Lüdecke turd left at the doorstep of CP". I have never done anything with Lüdicke, not written about him, had not any business with CP in recent years - but maybe thingnsbreak will explain what this supposedly is.
I initially had several examples listed, and cut them down to one. This specific example is incorrect, as it should have been attributed to your co-blogger and not yourself. I apologize for the error.
Being incorrect on that specific will no doubt cause some to dismiss my broader argument out of hand. Which is understandable, though I would hope that a few people might be able to see the forest for the trees.
My contention is that climate science has an overabundance of nonsense criticisms and little substantive criticism, and we need more of the former and less of the latter. While there was much in the interview I agreed with (and I was absolutely at fault in not expressing that agreement first), the comments about publication bias contributed to the superficial sort of criticism that I wish we had less of.
Obviously I have made some mistakes in my discussion here, both in tone and in details, but I hope this big picture is what will be addressed.
Cheers.
Err.. "more of the latter, less of the former".
Hi Jules, I enjoyed your interview very much, and I agree we need to have a discussion about the role of 'high impact' journals (honestly academics, how often have you heard the sentence 'X is really good he/she has N nature/science papers'). On the other hand, papers there need to say something different or new - which can be 'scarier' or 'unusually unscary'. In both cases, its harder to have a careful discussion than it would be in a specialist journal. So I think Jules concern is spot on. High impact journals of course also do a great job communicating to a broader audience etc. But it is tricky and I am sure the trickyness isnt limited to climate science as a field.
HvS, Jules, thingsbreak, et al.: there's an excellent word for this phenomenon -- "mansplaining."
@JBL
there's an excellent word for this phenomenon -- "mansplaining."
I'd appreciate it greatly if you could point out specifically where you see this happening in my conversation.
Thanks in advance!
(By the way, I don't think anonymity has anything to do with it -- people will be assholes on the Internet whether or not their names are attached to it. In the case of formal professional activities like Jules mentions though it might work better, logistical issues aside.)
thingsbreak, check your own posts; you explain to Jules the difference between what she said and what she meant! Your opening post ends "shame on you!" Both are terribly condescending. Would you like more?
@JBL:
you explain to Jules the difference between what she said and what she meant!
Can you please point out specifically where I did that? Much appreciated.
Your opening post ends "shame on you!"
Yes, it certainly does. And?
Perhaps you and I have greatly divergent understandings of the concept of mainsplaining?
I was not assuming "poor Jules, her ladybrain is inferior and she doesn't understand the dynamics of publication bias." I assumed that she fully understands them, and the "shame on you" is predicated on that assumption. It was not a criticism of failure to "get" something, it was a criticism about pandering.
But again, perhaps you have a different definition of mainsplaining that I am unfamiliar with?
Would you like more?
Yes, please. I am genuinely seeking to rectify any and all errors on my part, so I am thankful for any corrections.
@JBL:
Just to avoid any potential miscommunication arising from the lack of tone conveyance, I'm not being facetious, and I thank you for your constructive criticism.
I am aghast at the thought that I have been engaging in gender discrimination of any kind in my attempt at a conversation here.
von Storch,
You are engaging in some serious and unfounded innuendo. I do not know whether to laugh out loud at the inanity of your suggestion or accuse you of slander in what appears like an attempt to fabricate a strawman argument.
That was a silly mistake on my part that is easy to explain. Jules does not post very frequently on "James' empty blog". So I tend to associate posts on "James' Empty blog" with, well, James. Additionally, I simply did not see the small footnote saying that Jules had posted the content. It is that simple. Regardless, I sincerely apologize to Jules if she was offended by my mistake.
Please note that once Jules made a post that all subsequent posts were addressed to her when relevant.
I'm terribly sorry to disappoint you von Storch, but I can guarantee you that you really are barking up the wrong tree here. In fact, I would very much appreciate an apology from you for insinuating that I am a male chauvinist.
Thank you.
Hello Jules,
Just in case you missed it above, I sincerely apologize if you were offended by my mistake.
For the record, I would like to note that I took issue with one statement that you made in that interview. A statement that has nothing to do whatsoever with the misguided accusations that von Storch is making in the comment thread.
I just noticed the text on the wall to the left saying something like 'MOTOMACHINES'. Is that related?
Ooops, wrong thread. And boy is it the wrong thread.
Paul S: Ooops, wrong thread. And boy is it the wrong thread.
Yes it is. Please keep it down.
You're making their work more difficult for the caricatures-of-human-beings hair pulling crowd.
The behavior of these shit flinging monkeys on this comment thread could so easily be turned into a faux Monte Python skit.
You bet, as they're only caricatures of human beings we can feel free to put them on the drone list or whatever, right?
Do try to keep a lid on that nastiness.
JBL: Nature had a set of features recently, see here for an intro.
One point that seems worth mentioning: of three independent climate scientists commenting (and that's excluding both myself and jules) none of them seem to have a big problem with the content of the interview. I'm sure there is room for reasonable difference in opinion about the magnitude of the effect, but its existence is common currency among people who are actually working in the area!
The fact that some denialist memes (that it's all a conspiracy of exaggeration and lies) bear a passing resemblance to some aspects of climate science, does not mean that these aspects of climate science should be beyond discussion. Indeed, there are certainly some of us who think a bit more healthy dissent on many fronts would be a good thing. It's a truism to point out that science progresses through people disagreeing.
Thanks, James!
Just to be clear, my problem (and IANAS I hasten to note for any lurkers who don't know me) wasn't about the wholly uncontroversial point that the big glossies tend toward the exciting in their choice of what to publish. Rather it was with the claim that there's an increasing trend in same (and note that the claim was not about an increasing proportion of such papers, but rather a qualitative change in them). As far as I can see, none of the three climate scientists defended that point. That said, maybe it's true. But what is the evidence?
Steve: I based my interview responses on my personal experiences. It could be, however, that the effect is somewhat short-lived. In my experience, journalists don't have very long attention-spans. So it may not be that the hype is growing uncontrollably. It just feels like it is sometimes!
What tends to happen is that an extreme estimate of something appears in Nature or Science which irritate everyone, who then write corrections, which get published in much less high profile publications. It would all be good fun and frolicks if only the rewards for the extreme papers and the corrections were equivalent - after all, both have their place in the scientific process.
I think perhaps we should remove Nature and Science from the citations indices, and include them in the "outreach" or "PUS" parts of our CVs. Then Nature and Science could go back to being glossy magazines and stop pretending to be scientific journals. That is actually how they were treated when I did my PhD in Astrophysics. You broke the story in Nature and then afterwards wrote a detailed "proper paper", which is what the other astrophysicists actually read. These days people never write the proper paper. Instead you get this hashed up mess of supplementary information added to the Nature or Science summary article.
"hype is growing"
I meant...
"hype isn't growing"
re: "proper paper" vs. glossy letter
But this is also true for general "letter"-journals (e.g. GRL) which were conceived to provide room for such "announcements". ?
Well that's possibly true to some extent, though I think a basic GRL article is already about 50% longer than a Nature letter. Also, they tend to be simpler pieces of work that don't necessarily need a longer version (that's surely the case for our GRL articles) whereas it seems to me that Nature papers are often a really substantial undertaking, condensed to little more than an extended abstract for the purposes of publication.
But, that's just my impression.
I hope Jules won't be put off posting by some of the more hysterical comments posted here! I enjoy both her photos and her thoughts!
Re 'the glossies' I read this very much as Jules intended - I am a scientist (though not a climate scientist) so the shortcomings of Nature/Science are familiar to me.
I recently noticed that one climate scientist lists those of his papers appearing in 'Top three' publications (Nature/Science/PNAS) in a special category at the top of his pubs list. This did make me cringe a bit but perhaps that is just sour grapes :-)
Oh oops. I didn't realise Penis was a journal I was supposed to be reading. I thought it was just a pontification outlet for some pompous American old boyz club? I looked at it once but it didn't seem to have anything very relevant in it, so I didn't subscribe (to the RSS feed). Am I missing out terribly?
If nothing else, you have to give them credit for being up-front about their willy-waving nature...
An immunologist of my acquaintance says PNAS is where you go if you need to be published in a hurry but is not particularly high quality, FWIW. No idea whether that would vary from field to field.
James Annan: Well that's possibly true to some extent, though I think a basic GRL article is already about 50% longer than a Nature letter.
I think that paper-length bloat has been going on for a while:
There are "old school" papers that are one page long. (See e.g., John Nash's game theory paper.)
On the other hand, one of the most influential journals in physics is Physical Review Letters, which limits papers to 4 pages in length.
I suppose, in those days, you didn't have a lot of people trying to replicate your work, so providing enough detail so that could be done was a bit unnecessary.
One could also argue that science advanced slowly enough in those days, that text books could more or less keep up with the growth of a field (so less internal detail is necessary).
My personal soap box is the decline in the overall quality of the scientific literature. There are so many journals and so much stuff gets published that is mostly not new. The academic system which tends to count publications is part of the problem. It makes it hard for even specialists to wade through it all and find the important points.
And there is a tendency for the ambitious to try to get scarier and scarier numbers in some fields.
I too hope the critics of Jules will just relax and get a life. Her interview was just fine. This obsession over single sentences that "might give aid and comfort to vague and DARK forces (EVIL oil company lobbyist perhaps?)" is just so lame and stupid. Things scientists say will always be used by some inappropriately. Get used to it thingsbreak. The alternative is much worse. You have to have some confidence that the truth will win in the end. If you don't, you degenerate into some rather strange conspiracy theory tin foil hat Keepers Of Odd Knowledge Society (KOOKS) arena where you will find the company quite unpleasant.
Speaking of PNAS & quality of publications, this seems relevant:
http://www.pnas.org/content/109/42/17028.short
If you think things aren't great in the physical science, visit the biological sciences (where a substantial portion of the funding decisions are profit driven).
Anyway, here's the abstract.
A detailed review of all 2,047 biomedical and life-science research articles indexed by PubMed as retracted on May 3, 2012 revealed that only 21.3% of retractions were attributable to error. In contrast, 67.4% of retractions were attributable to misconduct, including fraud or suspected fraud (43.4%), duplicate publication (14.2%), and plagiarism (9.8%). Incomplete, uninformative or misleading retraction announcements have led to a previous underestimation of the role of fraud in the ongoing retraction epidemic. The percentage of scientific articles retracted because of fraud has increased ∼10-fold since 1975. Retractions exhibit distinctive temporal and geographic patterns that may reveal underlying causes.
Yes, Carrick, there have been several editorials on the growing incidence of outright fraud and retractions. There was an article in the New York Times last spring interviewing an editor for an infectious disease journal who had some disturbing observations about the way science has been evolving recently.
This issue of positive results bias I don't think gets enough attention. Anytime one is doing computational studies, you will almost always make many simulations. It's tempting to throw away the ones that "didn't match the data" and only report the good ones. The problem here is that many simulation methods are very sensitive to input parameters and model parameters, so the main scientific content of many computational studies is the negative results. Given enough time and a complex enough model, you can always match a given set of data. That tends to skew the literature and make it appear that modeling is a lot better than in fact it is.
I could quote you chapter and verse but I already did that at the Blackboard for you and I'm not sure James is interested in it.
Actually, for those who are interested in this issue of numerical uncertainty and positive results bias in numerical simulation, there are several interesting threads at Judith Curry's blog on this subject. The one on Lindzen's talk at the House of Commons is a pretty good one. I know James thinks Judith is irrelevant, and there are a lot of cranks of all stripes there, but some of the threads are really very good and some of the citizen scientists are also very informative.
David, if you want to acquire some minor credibility with climate scientists you might consider spending some time with aspects of the science that might actually matter. McI and the lukewarmers just keep rehashing the same uninteresting stuff year after year. If you just follow their lead, so will you, as we have seen here.
So why not look at something interesting and important, e.g. the failure of the models to track current Arctic amplification and, more significantly, that of the mid-Piacenzian?
My pet theory about this stuff is that people like you are interested only in the politics growing out of climate science and so pay attention only to aspects of it that they think can be used to support their political POV. But by all means go out and be an exception to the rule.
You see Steve, your final paragraph is what is so discouraging about the field of climate science. The default assumption is that anyone who says anything critical is politically motivated. I'm actually more interested in the numerical aspects of numerical simulations of PDE's. My interest in this arose from the realization that in fact uncertainty was higher than generally realized in engineering fluid dynamics and structural mechanics calculations. I also believe that the literature is not very good on this. There are some recent examples of negative results being reported, but its still unusual. A lot of the literature can be summarized as follows: We ran the code, results were initially bad. We then spent months tuning the computational grid and the subgrid model parameters (and adjusted the data in ways not fully reported) and then we agreed with the data almost perfectly. All intervening disagreeing results are not reported. Of course, from a scientific point of view, the intervening results are the interesting ones.
I also was pointed to some work of Paul Williams that pointed out that a lot of climate models still used the leapfrog method, a method that has well known nonlinear instability problems. This was known when I was in graduate school, circa 1975. Williams did an excellent job of proposing an improvement. But there are lots of vastly better methods available than his minor tweak to leapfrog. But it led me to recall the state of the codes at that time as I had some first hand knowledge of them and it was pretty bad. Anyway, numerical methods have advanced tremendously since then and I have a sense that climate models have not kept up. I am concerned that in fact climate models do not discretely conserve mass, momentum, and energy. Lacis says they do, but there is evidence they do not. This is DISCRETE conservation, not whether the contimuum equations are conservative as they must surely be. What I wanted to do was to get climate modelers to look at these newer methods and I rapidly found that the field was highly politicized. Paul Williams I think has some of the same frustrations as does Gerry Browning. At this point, my attitude is that I'm not sure climate scientists are really interested for a variety of reasons. I have learned a lot from citizen scientists and not as much from climate scientists about the fundamental mathematics involved. That in itself has been very useful to me in my professional life.
I'm not very interested in the problem you mention. My instinct would be to rewrite the climate models to actually address numerical error control and insert modern nonlinear stability methods. That's an impossible task for an individual. You need a team. I've seen a lot of legacy codes in engineering that are virtually impossible to adapt to better methods. It's often better to start over.
If you are really intersted I can send you some references. One break through seems to be the multiscale methods of T. J. R. Hughes and collaborators. These have very interesting potential to do subgrid modeling in a more rigorous way and are superior as discretization methods for the Navier-Stokes equations. Another breakthrough was variable time step methods with modern error control. This is actually very old and was developed in the 1970's. A third breakthrough was solution adaptive methods even though how applicable to climate models they are I'm not sure of. These are all quite easy to find in the mathematical literature or the engineering simulation literature. I would also suggest you revisit Gerry Browning's work which is fully rigorous and could be a breakthrough in controlling numerical dissipation. Based on Gerry's experience, I'm not holding my breath.
Steve, this second bit was beyond the character limit of James site.
I also am concerned as a citizen by the politization of the field, the bad press coverage as seems to have happened with this latest Marcott paper, the seeming closed nature of the literature to outsiders, the conflation of communication with science, and other issues of transparency. Where large issues of life and death are concerned, we expect better in other fields. I see little of that in climate science. There has been an interesting reexamination of science going on recently because of what Carrick talks about above. The medical literature has similar problems but there are better controls and people acknowledge that there is a problem. What seems to me to be worthy of censure is that climate scientists have circled the wagons and associated legitimate questions with "denial," "evil fossil fuel" interests and other fraudulent descriptions. I've been called a denialist and worse, but I don't care about that personally, but it is a symptom of a deeper problem. Some in the field are rewarded for this behaviour. So I try to harvest useful information from the debate, but am overall very frustrated with the whole affair and in a way just enjoy the give and take with those willing to engage honestly and even enjoy trading barbs sometimes with the usual anonymous trolls and activists and KOOKS.
Surely, in your private moments, you wish the field was more transparent and less political. If not, you and i would strenuously disagree.
Sadly, David, events are likely to outpace the development of the perfect climate model code. I'm not really surprised that you're uninterested in other aspects of the science.
Jim Hansen, who I seem to recall is some sort of modeler, noted recently that the evidence that we have a problem is based on paleoclimate, observations and models, in that order. IOW we don't need the models to draw some important conclusions.
But good luck finding nails for that hammer.
BTW, I interact with plenty of climate scientists, and on the whole I find them to be friendly and helpful. Possibly that's because I don't have a chip on my shoulder.
Yea, Steve, I'm not surprised that you seem to be uninterested in the details of what I said, except to repeat what everyone has heard many time from the "communicators" of climate science. BTW, I never said I wasn't interested in other aspects of the science, merely that the problem you stated probably required addressing the numerical aspects of climate models and that there was no hope of doing that without a team.
Cllimate scientists are I'm sure friendly and helpful as individuals even though some are sarcastic and egotistical on the web. My point was something else entirely, as I'm sure you know.
I guess that's what I get for assuming you would give someone who disagreed the benefit of the doubt. Odd that people you have never met can determine your motivations. :-) I guess that's the new frontier in climate science, mind reading.
Relating to one of Steve Bloom's comments, it's not clear that models are doing all that poorly in tracking polar amplification. People who are claiming the models are may be guilty of what they accuse others of--namely conflating natural variability about a mean with a secular trend.
It may of course be that the models are getting it wrong in this case, but, I think, the next ten years should tell the story.
My money is on the models having "issues" with properly characterizing natural variability for periods much less than 30-years, but are getting the secular trends "about right" (possibly running "a bit hot", I'm still sticking with my 2.5°C/doubling number though).
We can see a similar issue in the failure of most models to adequately capture the interrelations between ENSO and global climate), but as you increase the period that you are observing over, my expectation is, reality should conform better with the models.
David Young: I guess that's what I get for assuming you would give someone who disagreed the benefit of the doubt. Odd that people you have never met can determine your motivations. :-) I guess that's the new frontier in climate science, mind reading.
I don't think climate scientists as a rule assume motives or think they can mind read.
I think we're seeing "poisoning the well" activist tactics rather than scientific conduct.
David, I took your gratuitous insulting of Mike Mann as a hint. You may wish to review what you wrote in that regard. Such rhetoric will cause people to want to pigeonhole you.
I'll see that poisoning the well, Carrick, and raise you an ad hominem.
Steve, Steve, Steve, so your last comment indicates that your guise as an honest guy just trying to help me was a ruse and you had already judged me long ago. And you wonder why there is a problem here. Strange to say, my humorous comment about The Mann is a lot less judgmental than his fellow climate scientists at the time who judged his work in rather scatalogical terms. Mann's work is I think still rather questionable as discussed in Annals of Statistics. Why wasn't this done in a climate journal? I found the Team's response there rather unconvincing.
But my real question is did you actually read my detailed comments in which I outlined the areas of concern and do you have any response to them besides the talking points of climate "communicators." If not, then why did you come here in the first place? It obviously wasn't to have an open and honest exchange of views. But, its never too late to start an honest and open dialogue. I remain hopefully that you will do so and remain sincerely yours,
One who still thinks people should be honest and direct. And by the way, I'm not a coal lobbyist!! :-)
Ah, so doubtless this recent award takes the form of a bronzed pile of poop. Who knew. (Note that this is from the paleo section, and Yurpeens to boot, although I suppose that just makes the conspiracy even vaster.)
Re me coming here, that was something on the order of ten years back when it really was nearly empty (not that that gives me dibs or anything).
Re your proposal for climate modelers to straighten up and fly right, as modeling experts without attachment to a particular model J+J are the perfect audience. Make the pitch. I won't get in the way.
True, the case you're trying to make implies that they've basically been wasting their time these last ten years, but you don't need a team of modelers to show that the models are fundamentally defective as they stand. To borrow Einstein's phrase, it just takes one, and if as you say you know they've been proven wrong, that one is you. So get to it.
Although one drawback I would point out to J+J is that no sooner would one group of self-appointed auditors/experts be satisfied, however provisionally, than another would crop up. But I expect they already know that.
As to why I don't find your arguments very interesting or worth engaging, it's because I tracked their development in real time over the years and never could find much substance. Not to put words in J+J's mouths, but perhaps they have the same view for the same reason, albeit a far more technically-informed one than my own.
Almost forgot:
"I never said I wasn't interested in other aspects of the science, merely that the problem you stated probably required addressing the numerical aspects of climate models and that there was no hope of doing that without a team."
Oddly I made exactly the opposite point, or tried to. This is a matter of missing physics, which should be right up your alley. There are teams standing by to take it from there.
David,
When I first got involved in climate (well, at first it was just ocean) modelling, I had just previously been involved in some CFD research, and I was surprised at the apparently primitive methods involved in climate science.
However, although I'd agree there is certainly some potential for improvement, it soon became clear that the numerical methods aren't generally a limiting factor, or even close to it. The important controls on model performance are the uncertainties in sub-gridscale parameterisations.
There are plenty of people working on more advanced numerical methods. When they can demonstrate any tangible benefit (or perhaps, sufficient benefits to overcome the added complexity), I'm sure their methods will be adopted.
Note also that these primitive methods and models are widely used in the highly competitive and commercially important area of weather prediction. And they do pretty well (at least, the limitations are not due to technical details of the numerics.
Yea, James, that is what I have been told too. Paul Williams however showed quite convincingly that numerics can make a difference in the limit and give you different "climates."
I'm skeptical that numerics don't make a difference. As you know if you worked in fluid dynamics, numerical dissipation is a deadly enemy of accuracy. And Browning makes some rather convincing arguments that there is a lot of extra dissipation in the models. That could lead to "false" stability and spurious infinite time attractors. In fact, it seems that dissipation is added precisely to keep the simulation from "blowing up." That is very unsatisfying.
What I've heard is something along the lines of "Every time I run the nodel, I get a reasonable looking climate." So the doctrine I guess is that the attractor is strong enough that you eventually get "sucked in." Not very rigorous or very satisfying. Of course, excessive dissipation would lead to exactly this result. So I trust Browning whose results seem to rigorous and provable.
Stere, Have you seen Paul Williams material? It's pretty convincing and he did really improve model performance. I appreciate your more honest last message though!
As to your last point, its a very old dogma, trite and wrong. Adding more "physics" can actually increase the uncertainty. Look at Nature from last spring roughly. It's also clear in simpler systems. That's why Reynolds' stress models give inferior results compared to eddy viscosity models, even though they include a lot more "physics". There are just too many parameters to tune. If you are really interested its easy to find the information. But you must exercise skepticism and not just accept the results. Every prefect reulst is the result of "tuning" often for a particular type of flow. It will take some work however. But I trust you are sufficiently ingeneous to do it.
But please, patronizing statements about your concern for my influence in climate modeling and how to make an impact are not very genuine, are they? If what I'm saying is correct, then its worth pursuing even if climate scientists are "friendly and nice people." I really don't care on a personal level about this. My concern is about the science and the way its been politicized and the way people like you seem to just parrot the talking points.
Hope that helps.
Steve, In terms of The Mann, you should read the papers in Annals of Statistics. It's a pretty lopsided matchup. And there are those pesky emails where his co-workers are truthful about his work. How did that happen? You know Al Gore was awarded the Nobel Peace Prize even though his movie contained serial misrepresentations. Muller is excellent on this by the way. Have you seen his presentation on it? Just for grins, you should go a little bit out of your comfort zone and view it. It's an hour long lecture and also discusses the "hide the decline" controversy. You can find it on youtube.
James, One more point. Sorry to be so long winded! Modern methods such as the SUPG method of Hughes et al have additional advantages such as exact discrete conservation including of some of the moments.
There is some evidence that the models do not discretely conserve mass, momentum, and energy. It was shown roughly 30 years ago that this discrete conservation gives dramatically superior results. And it affects the limit of long times too.
But the most important concern is just this dogma of the attractor as I described above. That's a recipe for missing tipping points and abrupt climate shifts. You know of course that the n body gravitational problem is also chaotic at long enough time scales. But no one in his right mind would say that numerics are unimportant because "a million years from now the statistics might look reasonable." Numerics are critical for what is important, such as whether an asteroid is going to terminate our joyful lives here on earth next year.
Subgrid models are also an interesting issue. Once again, SUPG offers the opportunity to do a more rigorous job here. Basically, the method gives you a way to use the Green's function to accurately get the subgrid influence on the resolved scales. It's still in its infancy, but it could be a breakthrough. Once again, however, a lot of subgrid models introduce extra dissipation and always too much dissipation which degrades the accuracy of the simulation. Eddy viscosity models are a good example. They are far too dissipative in vortex sheets, a little understood but deadly effect.
David,
It's not like anyone is preventing enterprising researchers from developing better methods - Paul Williams being a case in point, also in a rather different direction, the stochastic approach that Tim Palmer and others are investigating. But the idea that they are either (a) a magic bullet that will solve all the problems or (b) a necessary step before models can be considered useful, are entirely misguided IMO and IME.
James, I agree that better numerics is not a magic bullett that will fix all problems. You and I agree on that. However, I could cite an interesting case study where they enable things like stability analysis that are not possible with older methods. That it seems to me is what we are really interested in, can we say anything meaningful about stability of climate states and when will there be rapid changes. I'd argue that dissipation especially if its as large as Browning says would make these things much more difficult.
Has anyone really seriously tried more modern methods. I think Williams work shows that even with a little bit of effort improvements are possible.
Steve Bloom: I'll see that poisoning the well, Carrick, and raise you an ad hominem.
Yeah, I am sure you would see me an ad hominem. :-P
Saying this is an example of poisoning the well, though, isn't an attack on your character (so it's not an ad hominem), but a criticism of a class of behavior by activists (not just those in your camp either), but it does apply to your comments on this thread, and as far as I can tell involves:
1) if a researcher is also an advocate and on "my side" in the policy debate, adopt them like a lost puppy.
2) If a researcher is agnostic or hostile to your views, cowl him/her with bullying tactics.
3) If that doesn't work, try to undermine their credibility ("poison the well"), in this case, by dissing their arguments as being politically motivated.
Nice try, Carrick, but no cookie.
"can we say anything meaningful about stability of climate states and when will there be rapid changes?"
So long as we still have missing or substantially incomplete processes (per obs and paleo), no, of course not. Which is yet another reason why stopping and doing the sort of extensive re-working you'd like doesn't make much sense. And also why your unwillingness to think about what is knowable without recourse to modeling is so strange.
At a certain point this stuff from engineers, programmers and statisticians just starts to sound like rent-seeking.
James, I hope you are still paying attention to this thread, because I think I have as close as one can come to a proof that modern methods can be a powerful tool in determining stability of solutions, finding multiple solutions, and generally shedding light on a field where 20 years of work had led to unreliable codes whose results had to constantly be tuned and were often unreliable.
The example is the high Reynolds' number Reynolds Averaged Navier-Stokes equations (RANS). As I'm sure you know it has a lot of characteristics in common with the atmosphere such as compressiblity, nonlinearity, vorticity, turbulence characterized by a subgrid model (with 20 years and 100s of man years investment), and Rayleigh Taylor instability. RANS itself has perhaps 10000 man years in "research" much of it spent "running" codes and puzzling over strange results. 15 years ago a team led by my Forester Johnson set out to build a more reliable code and used the latest methods. One result of this work is AIAA paper 2013-0663. Just to clarify, most of this work was done by the authors and not by me personally. I can send you a pdf if your email is available. Basically, all other codes use weak iterative methods that are pseudo time accurate methods to attempt to converge to steady state. What we have found is that in fact there are many different machine zero steady state solutions for RANS with common turbulence models. These solutions are sometimes stable to small perturbations, sometimes not. What is more disturbing but not covered in the paper is the existence of pseudo-solutions, flow fields for which the residuals are converged 5 orders of magnitude but ARE NOT machine zero solutions. Often the machine zero solutions are far away but in some cases they are rather close. More disturbing the pseudo-solutions seem in many cases to be closer to test data than the machine zero solutions. These multiple solutions can be "found" by changes to the transient time stepping procedure, but nothing that changes the steady state solution. The method used is Newton's method with sophisticated domain decomposition preconditioning.
You might say steady state doesn't interest you, however, this is significant I think for time dependent methods as well, because these solutions can be considered stationary points arrived at by some form of time marching, albeit one involving strongly implicit methods. In any case, one could demonstrate what appear to be stable orbits and attractors using a weakened version of the time marching procedure. This happens all the time in older codes and the code is said to be "unconverged" or to have found a time dependent flow. All of these are imprecise euphemisms.
This has broad implications:
(Continued in next comment)
This has broad implications:
1. Results from other codes are suspect since they rarely converge to more than 4 digits in the residual. In fact, there is no way to reliably determine what their solutions mean. Results can vary significantly with small changes in the thousands of choices needed to run a complex code (often based just like climate models on the best numerical methods of the 1960's). This explains the huge industry that tunes and runs these codes. Expertise is required to get "reasonable" results. Are they "repeatable" or merely tuned using "expert judgement?" I would claim there is no rigorous criteria to judge the difference, until now.
2. We suspect that turbulence models are calibrated incorrectly, having been calibrated using the old codes. Pseudo-solutions are in some cases the ones that were used. However, no one had any idea that was the case (just as in climate where parameters are tuned based on model runs that are compared to data, albeit I understand some are tuned based on first principles).
3. It calls into question the whole notion of Reynolds averaging, its relationship to direct simulation and Large Eddy Simulations.
4. On a practical level, it gives us a tool to estimate the uncertainty in our results in much more rigorous terms.
In short, even though not yet fully recognized in the outside world, this work is a huge advantage for the next decades in CFD.
And EVERYONE in CFD in 2000 said just what you said about numerical methods. Hey, the results are reasonable (admittedly mostly on easy cases after expert tuning). The literature shows the results are pretty good. (Read positive results bias). We could spend a decade developing a better code and might get nothing out of it. In any case, our current codes are "useful." In short, everything I have heard from climate scientists about models was standard dogma in CFD for 20 years, so much so that at first the team did not believe the results, even though I had expected them. Gradually it became obvious that people had overlooked something.
I would be happy to follow up with you and provide more material. I hope you appreciate that this cannot be fully shared for reasons you surely know very well. But, we can share what I have shared here as well as some examples that might whet your appetite.
It's a long road and prolonged investment is required. But it may be the only way to provide a new level of understanding required to move a field forward.
Regards,
A coal industry lobbyist as unmasked by BBD through extensive research (if only it were true I would probably retired by now sitting on my own private beach)!!! ;-)
Steve Bloom can read it too provided he's not a French industrial espionage agent (Just wanted to confirm your prejudice that its all due to conspiratorial thinking) ;-)))
PS, I will provide the extra material to you personally at your private email address if you find a way to get it to me. I think my email will appear to you as the site administrator with these posts.
"I think I'll go back to just posting photos, and let James do all the writing from now on. He's so much more flame retardant."
I love your photos. And I enjoy your science. I suppose you could call this a lesson in the thin green line.
There are certain things you cannot say out loud. They have nothing to do with actual science, but rather the sociology of science and the politics of science.
So, you basically have two choices.
A. Silence of the lambs
B. Judy's path.
I vote for this
http://www.youtube.com/watch?v=VDiybWkyfgY
Ah, the Curry Comparison. You really know how to flatter the ladies....
The last imbecile who did the 'woman scientist - Judith Curry' thing in front of me would have had my beer emptied over his head, had I not already drunk it... I was so angry I went silent - for a moment!
So, you basically have two choices.
A. Silence of the lambs
B. Judy's path.
Those are hardly the only choices.
Post a Comment