Slashdot Mirror
Scientific Journal Nature Finds Nothing Notable In CRU Leak
eldavojohn writes with an update to the CRU email leak story we've been following for the past two weeks. The peer-reviewed scientific journal Nature has published an article saying the emails do not demonstrate any sort of "scientific conspiracy," and that the journal doesn't intend to investigate earlier papers from CRU researchers without "substantive reasons for concern." The article notes, "Whatever the e-mail authors may have said to one another in (supposed) privacy, however, what matters is how they acted. And the fact is that, in the end, neither they nor the IPCC suppressed anything: when the assessment report was published in 2007 it referenced and discussed both papers." Reader lacaprup points out related news that a global warming skeptic plans to sue NASA under the Freedom of Information Act for failing to deliver climate data and correspondence of their own, which he thinks will be "highly damaging." Meanwhile, a United Nations panel will be conducting its own investigation of the CRU emails.
... will be bamboozled. There is nothing new in that. It lies behind all political folly.
The data that was adjusted was paleoclimate data, and what it was being adjusted to was temperature data (i.e., the more reliable modern temperature data). As far as I can tell, they neither could nor did adjust the measured temperature data.
The OP did not quote the really important part of the Nature piece :
Nothing in the e-mails undermines the scientific case that global warming is real — or that human activities are almost certainly the cause. That case is supported by multiple, robust lines of evidence, including several that are completely independent of the climate reconstructions debated in the e-mails.
The evidence for this is literally all around us. Throw all of the CRU data out if you want. It won't change a thing.
The source code is another thing that has been taken completely out of context. For example, the "Very ARTIFICIAL correction" in was computed, but the line where it was applied was commented out. This is clearly a case of someone playing around with the output to explore various effects and scenarios. They did the experiment, then commented it out to removed that effect from the production output. Does anyone really believe that a real attempt at fraud would be blatantly labeled in the code?
Just because you are paranoid does not mean that no-one is out to get you.
FFS, not this shit again. Frankly, given the number of times this has been pointed out, I can only assume that people still mentioning it are wilfully ignoring anything that contradicts them.
However, one last time. It's commonplace to have multiple versions of analysis code with variations including "artificial" changes. I've done it lots of times, mostly for testing purposes - it's quite useful [1] to know how the output of your analysis depends on variations in the input. Only a small fraction of the code in existence was actually used to process the data "for real". Unless you have some evidence that a) arbitrary modifications were made to published data without explanation in the publication and b) any such modifications were not scientifically justified, please stop with this ridiculous and discredited point.
[1] By which I mean "essential for any non-trivial analysis"
If you look at the actual code instead of some blog you'll see the reference to the adjusted value is commented out and never used in the plotting call.
; ;filter_cru,5.,/nan,tsin=yyy+yearlyadj,tslow=tslow ;oplot,timey,tslow,thick=5,color=20
; Apply a VERY ARTIFICAL correction for decline!!
;
yrloc=[1400,findgen(19)*5.+1904]
valadj=[0.,0.,0.,0.,0.,-0.1,-0.25,-0.3,0.,-0.1,0.3,0.8,1.2,1.7,2.5,2.6,2.6,$
2.6,2.6,2.6]*0.75 ; fudge factor
if n_elements(yrloc) ne n_elements(valadj) then message,'Oooops!'
;
yearlyadj=interpol(valadj,yrloc,timey)
;
;
filter_cru,5.,/nan,tsin=yyy,tslow=tslow
oplot,timey,tslow,thick=5,color=21
;
oplot,!x.crange,[0.,0.],linestyle=1
;
plot,[0,1],/nodata,xstyle=4,ystyle=4
Without revision control one can't say for certain but there's no evidence any adjusted data made it into a paper. There's only evidence that a single piece of code from the thousands of modeling sims had at one time an adjustment that was commented out.
I will, however, admit that the researchers should have noted the issues with the tree-ring data in question.
Good thing they did, then. Only ten years ago, mind you.
Seriously, this whole "climategate" debacle tends to run like this:
1- Deniers exhume some e-mail / piece of code which they don't understand, but assume is definite proof of evil scheming on the part of the great academic conspiracy ("Trick!" "Hide the decline!" OMGconspiracy send teh copz!!) .
2- Scientists post explanation, showing the deniers' allegations to be baseless (The "hidden" decline in tree ring growth was published a decade ago - see Nature link above; in this very publication, it was shown to diverge from the actual instrumental record after 1960; so for the post-1960 period we basically replace tree rings with the actual instrumental data, because we trust thermometers more than tree rings when the two fail to agree; we cited the relevant articles in the caption for the graph just to be sure).
3- Deniers completely ignore scientists' explanation, and keep fantasising about their glorious victory over evil scheming scientists. See GP for an illustration.
Rinse. Repeat.
To GP and all the folks who keep harping about this "VERY ARTIFICIAL" correction code: the code in question is a one-time code for temporarily re-calibrating the tree ring data. The reason, and the coefficients, are ultimately derived from the Nature article I linked to above. For an interesting hypothesis concerning the source of this code, see comment #147 and linked manuscript on this thread.
Which version was used to produce the plot that is in the Nature paper? It is easy enough to tell by looking at the Nature paper. Hint, the plot in the paper does not have this correction applied.
Just because you are paranoid does not mean that no-one is out to get you.