Jones and CRU exonerated

[BenBurch]

Trouble is, they DO understand the papers you speak of. And that is WHY they are not referred to by other papers.

 

[CoolSceptic]

Trouble is, they DO understand the papers you speak of. And that is WHY they are not referred to by other papers.

Ah, so 1900 citations is a low number then is it?

Ben, I think there is still someone in the world who can't sense your desperation. Keep at it.

 

[BenBurch]

You're the desperate one. Sorry, but I know who has the upper hand here, and its not the AGW deniers.

And every day that passes, your cover gets weaker and weaker.

 

[BenBurch]

Its not getting any cooler...

 

[uke2se]

Ah, so 1900 citations is a low number then is it?

Ben, I think there is still someone in the world who can't sense your desperation. Keep at it.

Just to be clear, these 1900 citations are on a paper that is relevant to this discussion, i.e, critical of AGW and not mentioned in the AR4, and further, the 1900 citations aren't 1900



from climate scientists?

 

[uke2se]

No, seriously, stop it. Back to real science.

In science, the value in a paper is entirely based on the correctness of the paper, based on reasoning and evidence. It has *nothing* to do with popularity.

Of course, if you do not know enough to assess the merits of a paper yourself, then you may be limited to merely assessing what other people think of it. Please do not project that position onto others, though.

I think you're quite deep in a hole right now, but just to help you along, I didn't say popularity increased the value of a paper. I said popularity is indicative of the value of a paper. I.e, if a paper is popular, that is likely because it is seen as a good paper by the people who would know.

This is not always true, however. Sometimes a paper is cited just to show how wrong it was.

I'll leave you to keep digging your hole now.

 

[DogB]

and further, the 1900 citations aren't 1900



from climate scientists?

And thus we see the problem with using popularity as a measure of significance.

I'm sure Benveniste's Nature paper was cited a fair few times as well.

 

[uke2se]

And thus we see the problem with using popularity as a measure of significance.

I'm sure Benveniste's Nature paper was cited a fair few times as well.

It's as good a place to start as any. Of course, if you settle with the amount of citations you aren't doing your argument any favors.

 

[DogB]

I'm confused. Are you suggesting there's a way to judge a papers popularity without using citation data?

 

[lomiller]

And thus we see the problem with using popularity as a measure of significance.

So can we put you down as believing citations and peer acceptance are irrelevant and once a paper is published it's fact no matter what the scientific community thinks of it?

 

[uke2se]

I'm confused. Are you suggesting there's a way to judge a papers popularity without using citation data?

That depends on how you define popularity. If you mean "popular to cite", then it is the way to go. If you mean "popular because the paper is good" than it is just a one tool. You can never settle for citation data only. If you do, you'll run into a paper sooner or later that is simply cited because it's wrong.

I don't know if you're trying to be purposely obtuse in order to trick me, but it's really not that hard to understand. Citation data can indicate if a paper is popular or not, but it doesn't prove that a paper is popular because it's good. Alright now?

 

[lomiller]

I don't know if you're trying to be purposely obtuse in order to trick me, but it's really not that hard to understand.

Most likely he is simply looking for a convenient excuse to dismiss any research that conflicts with his beliefs while accepting any research that confirms his beliefs. No doubt my efforts to pin him down on exactly what his criteria is for accepting/rejecting papers is, but I seriously doubt it will come down to anything other then "A good paper is one he agrees with a bad paper is one he disagrees with"

 

[DogB]

No, no I'm just trying to decipher the meaning of the word 'popular' and 'good' in this context. Is there an empirical measure of popularity or even peer acceptance (other than citation frequency)?

 

[DogB]

BTW, I agree with the premise that citation frequency is a flawed measure of popularity. I'm just trying to decide if there is an alternative.

 

[lomiller]

No, no I'm just trying to decipher the meaning of the word 'popular' and 'good' in this context.

Then you are arguing with the wrong people. We've been arguing on the basis of citations and peer acceptance. Since the paper Coolsceptic is pushing has has neither he tried to write those off as "popularity".

 

[lomiller]

No, no I'm just trying to decipher the meaning of the word 'popular' and 'good' in this context. Is there an empirical measure of popularity or even peer acceptance (other than citation frequency)?

Empirical? Not likely, the process of hashing out scientific discussions is not inherently empirical, it’s about who has the arguments that can win over ones peers. Citation, responses, which journals they appear in, and how high the impact of that journal and of course work based on the papers conclusions all factor into it.

For cutting edge work that is hotly debated you can’t tell who is winning based on citation count. Fortunately in this case there isn’t one of those grey areas. There is, simply, no heated discussions taking place as a result of the paper in question.

 

[TellyKNeasuss]

Yep. And gets it utterly wrong, despite the fact that for him, I went into great detail explaining some of the ideas behind the work. Which underlines to me why explaining these things is a waste of time on JREF.

You clearly don't. He misrepresents Halley's paper. If you had read and understood the paper, you'd know that. But you haven't, and you don't. Which is why explaining these things on JREF is a waste of time. (Can you see a pattern emerging here?)

Yes, how dare I disagree with the world's foremost authority in this area!!! I should be humbled that you were so willing to attempt to explain things to an ignorant wretch like me.

Unfortunately, TellyKNeasuss' attempts fall rather woefully short.

As I've already explained, the principle problem that Cohn, Lins and Koutsoyiannis point out is that the geophysical time series exhibit a particular type of persistence which results in a very high instance of type I errors (i.e., that something is significant or exceptional when in fact it is not). Put simply: in climate science, CIs are underestimated. This was my primary claim above, which you seemed incapable of reading or understanding, yet it is up there clear as day.

Yawn. No one is disagreeing that there is some persistence in climate. We can even come up with mechanisms for it, like the ice-albedo feedback.

What is the first thing Halley does in his paper? He tests to see whether this persistence does indeed seem to exist. His result? Yes, he confirms the presence of this type of persistence.

Just in case you missed this: his first test result confirms the views of Cohn, Lins and Koutsoyiannis.

And just about everyone else's.

then goes on to apply a test to determine whether paleoclimate reconstructions are significantly different from modern instrumental records, even accounting for the above persistence. He finds they are.

However, he also finds that these reconstructions are significantly different from one another. Oops. That is a problematic result.

He notes that the result he finds could be because AGW exists. And, indeed, I made it clear that I don't claim AGW doesn't exist. I just claimed the CIs were woefully underestimated. That is, AGW might exist but it isn't necessarily measurable based on the current data.

However, Halley also points out his result could have been caused by systematic errors in paleoclimate reconstructions, and references another paper that suggest just that possibility. This is a rather more plausible explanation, since it also explains the inconsistency between reconstructions.

Halley then outlines a list of the problems that would need to be resolved before you can make the interpretations that AGW can be confirmed even with the presence of long term persistence.

But even if this test ultimately could confirm AGW exists, it still leaves an enormous number of type I errors in the IPCC reports and the peer reviewed literature by virtue of the type of persistence applied in those analyses. Halley's analysis can never refute that.

All of this is clear as day in the Halley paper, written out in black and white. It isn't well communicated by the abstract (which is why scientists never rely on an abstract). But despite it all being written in black and white, as clear as day, neither you nor TellyKNeasuss could figure it out. And I have to produce reams of text explaining why you're wrong.

If you could actually explain why I was wrong, you wouldn't have to "produce reams of text". Who was it that said "When ideas fail, words come in handy."? No matter how you try to spin it, Halley found that even assuming that the reconstruction with the largest variance was the most correct that there is a very low probability that recent warming fits within natural variability. Nor would it even if he had used a 99% rather than 95% criteria.

Nor can you address the issue with the Cohn and Lins paper, which was that they calculated their "natural" variance from the same time period as the trend they were testing, i.e., they "proved" that the warming trend was natural by assuming that all the variance during that same period was natural. And still they barely missed the 95 percent criteria. (Incidentally, a failure to reach a specific probability value only means a failure to detect what you are testing for, it doesn not prove that the opposite.)

But realising why you're wrong requires you to understand some of the nuances of this debate. I'm not optimistic on this aspect. Nobody else on the AGW threads of JREF has demonstrated the knowledge or ability to follow this. TellyKNeasuss - despite being the best on here by far - couldn't even see that Halley's paper strongly supports my claims.

Support your claims about what? About your strawman that the rest of us don't believe in LTP/feedback?

 

[DogB]

Empirical? Not likely, the process of hashing out scientific discussions is not inherently empirical, it’s about who has the arguments that can win over ones peers. Citation, responses, which journals they appear in, and how high the impact of that journal and of course work based on the papers conclusions all factor into it.

This I agree with, to a point. I rarely (never) investigate the citation record of a paper but I naturally accept that if it's in Nature or Science then it's probably received good quality reviewing.

For cutting edge work that is hotly debated you can’t tell who is winning based on citation count.

100% agree. Who would even try?

Fortunately in this case there isn’t one of those grey areas. There is, simply, no heated discussions taking place as a result of the paper in question.

I’ll take you’re word for it, I haven’t a clue and don’t particularly care.

My point was (and I believe it remains valid) that citation record is somewhat questionable as a check of a papers validity. Given that the other measures you noted are subjective and thus highly vulnerable to personal bias*, then surely the best judge of a paper’s worth is its contents?

*With the possible exception of journal impact factor and even that’s not infallible – Benveniste again.

 

[CoolSceptic]

Just to be clear, these 1900 citations are on a paper that is relevant to this discussion, i.e, critical of AGW and not mentioned in the AR4, and further, the 1900 citations aren't 1900



from climate scientists?

Yes, the paper was one under discussion on a previous thread just before lomiller got desperate and started talking about citations. It is the paper that lays down the mathematical underpinnings and recognises that weather probability distributions have a "long tail" and extreme events have a tendency to cluster (at all scales). This is why the CIs in IPCC AR4 are underestimated.

The paper itself doesn't mention AGW, most likely because it was written in 1968 and AGW wasn't really on the agenda then. Back then, scientists worked on science rather than politics.

 

[CoolSceptic]

Yes, how dare I disagree with the world's foremost authority in this area!!! I should be humbled that you were so willing to attempt to explain things to an ignorant wretch like me.

I don't claim to be an authority, and I gave you credit for going to the literature (which puts you years ahead of anyone else on this forum). It is just a shame you can't follow the detailed points put forward in the papers.

Yawn. No one is disagreeing that there is some persistence in climate. We can even come up with mechanisms for it, like the ice-albedo feedback.

LOL. This statement is priceless. You were doing so well up to here, looking up scientific references, but now you've abandoned science and are just making stuff up. Non-linear interactions in a complex system are the cause, indeed, but to start hand-waving specifically about albedo misses the point by a wide margin.

And just about everyone else's.

If only this were true. Unfortunately, it is not. Correcting for the effects of persistence is virtually unheard of in climate science.

If you could actually explain why I was wrong,

Already done that. I can see I'll have to do it again.

No matter how you try to spin it, Halley found that even the paleo-climate reconstruction with the largest variance still doesn't contain enough variance to falsify the claim that the recent warming exceeds normal climate variability. Nor would it even if he had used a 99% rather than 95% criteria.

Yes, and he also shows that reconstructions have non-overlapping variances, which if you accept the case you put forward, "proves" that one reconstruction has AGW in it and the other doesn't, even though they both cover the last one thousand years. Which is clearly wrong.

Luckily, Halley is honest enough to acknowledge this clearly in the paper and recognises that the data aren't good enough to draw the conclusion you desperately want to draw. Halley recognises the systematic errors in the reconstructions (e.g. von Storch's paper) and outlines what would be needed to improve them.

Good grief, he even notes changing the version of the instrumental data set reduces the magnitude of the effect he is measuring by almost one half! It is blindingly obvious that the data are not good enough for the very powerful test he is applying. It is clear when mere versions of the same data set you are using change the results by a similar amount to the magnitude of the effect you are measuring, that you have a serious problem.

And even after all of this - that isn't the contention I made. I didn't claim that AGW doesn't exist. I claimed that the CIs calculated throughout climate science are flawed and prone to producing false positives because the standard approach in climate science is not to correct for persistence. A point you conveniently ignore.

Nor can you address the issue with the Cohn and Lins paper, which was that they calculated their "natural" variance from the same time period as the trend they were testing, i.e., they "proved" that the warming trend was natural by assuming that all the variance during that same period was natural. And still they barely missed the 95 percent criteria. (Incidentally, a failure to reach a specific probability value only means a failure to detect what you are testing for, it doesn not prove that the opposite.)

Again, a misreading of the paper on your part. If you read Cohn and Lins, they stress that the significance values produced in their paper are not terribly meaningful. The point they are making is that the difference between climate science standard practice p-values and p-values with persistence can vary by many orders of magnitude. It explains this openly and honestly in the paper (just as the problems with Halley's paper are explained openly and honestly, but you seem unable to comprehend).

Once again: in science it isn't enough to read a paper and cherry pick your own conclusions. You need to understand what the paper is saying, with all of the nuances and caveats.

Support your claims about what? About your strawman that the rest of us don't believe in LTP/feedback?

Hilarious that you accuse me of a straw man and then introduce your own.

Firstly, LTP and feedback are different things (particularly the naive linear feedback approximations used in climate science).

Secondly, my point is that it is NOT standard practice to correct for persistence in climate science, therefore in the vast majority of papers the CIs are too narrow, leading to a large number of false positive claims. I have not claimed that AGW doesn't exist, that isn't even on my radar at the moment. You haven't even begun to address my main contention here.