Andy Revkin at the NYT asked me the following yesterday:
I asked him, in essence, if the shape of the 20th-century temperature curve were to shift much as a result of some of the issues that have come up in the disclosed e-mail messages and files, would that erode confidence in the keystone climate question (the high confidence expressed by the Intergovernmental Panel on Climate Change in 2007 that most warming since 1950 is driven by human activities)?
You can see my response at
DotEarth. I am particularly heartened to see the reaction of Raymond Pierrehumbert, of Real Climate and University of Chicago, who I collaborated with on a NRC project and paper some time ago. Here is how he starts his
comment:
I often find myself disagreeing with Roger, but in this instance I find little to fault with his nice summary of the situation.
Please join the discussion over at
DotEarth.
10 comments:
Roger, I liked your even handed and well nuanced response. No one knows how well or badly the processes used by CRU and GISS/NZ/NOAA, etc will hold up to a transparent review.
I suspect that one of the underlying issues behind all the obfuscation was the actual quality of the processing and the raw data. I'm sure you are aware of the issues associated with herding a bunch of undergrad & grad students in doing their research, let alone relying on some of their S/W coding.
Roger:
"If it turns out that the choices made by CRU, GISS, NOAA fall on the “maximize historical trends” end of the scale, that will not help their perceived credibility for obvious reasons."
Recently there have been posts suggesting that the New Zealand, US and Northern European temperature trends put out by CRU and others may indeed "maximize historical trends." At the very least it appears that in each case the anomalies overwhelmingly come from adjustments to the raw data rather than from the raw data itself.
To over-simplify, I think one line--the primary line?--of AGW thinking goes something like this:
1. Present temperatures/climate are "unprecedented".
2. In the models we can only reproduce our unprecedented climate if we take into account a significant CO2 effect.
3. Those same models indicate that things get much worse if CO2 continues to increase.
If the first point is re-assessed and comes to a different conclusion, what is actually left?
I've heard many people say that CRU doesn't matter because there are many independent lines of evidence--but are those other lines of evidence actually independent of the first point?
Roger,
Le Mouel et al. in their 2008 paper "Evidence for a solar signature in 20th century temperature data from the USA and Europe" C.R. Geosciences 340 p421-430 reconstructed Europe and US temeprature evolution using the raw data they had to dig themselves since CRU did (could)not help.
The temperature curves are so significantly different from 2007 IPCC AR4 that, regardless of the rest of the paper's conclusion on solar/climate relations -a work in progress-, that fact alone should have attracted worldwide coverage and another stone in CRUs glasshouse. Yet it did not. Sad indeed.
For the less sophisticated among us, I recommend Richard Lindzen's excellent WSJ article.
The Climate Science Isn't Settled
http://online.wsj.com/article/SB10001424052748703939404574567423917025400.html?mod=googlenews_wsj
You have to laugh at North saying "the model's aren't perfect". They models aren't even adequate! If he needs confirmation of that then how about asking some pertinent questions of the modelers like - "just exactly what parts of the worlds climate do these models manage to reasonably simulate". In his position you'd think he might have asked that question already. He'd have found out that the answer is pretty much nothing of any practical use.
Of course that doesn't stop people using model results for impacts reports - and pointedly ignoring the more logical long-term trend analysis of actual weather events. One of the greater follies in this field being to assume that combining and averaging 20 bad results will somehow magically produce a good result.
A blogger named chiefio has been dissecting the GISTemp data.
http://chiefio.wordpress.com/2009/11/07/gistemp-ghcn-selection-bias-measured-0-6-c/
One thing he noticed is GISTemp drops stations over time and this does have an effect on the trend.
What this suggests is the process for selecting what stations is as important as the algorithms and adjustments themselves.
Roger is absolutely right. The long term impact of climategate will be that the game of hide-the-ball is over. The code, the data, the methods, etc. will all have to be made available for review. And that is all to the good.
The political nature of the argument tends to make everyone focus on the question of fraud, bad intent, ethics, etc. The far more interesting issue over time is that of competence.
As engineers keep pointing out, the code is all crap. Incompetence abounds. GISS, CRU, Mann's studies, Biffra, et al -- we keep seeing questionable work, lousy assumptions, shoddy stats, and monumental hubris in announcing the significance of the results. It's not going to end well.
When the average man on the street understands how bad the thermometer siting is, he's gonna be shocked. When he begins to understand all the machinations and manipulations, he is going to be angry. And when he understands how much the GCMs don't include, he's going to be really pissed.
Want to see jaws drop? Explain how bad the temperature siting is. And how no one in authority even bothered to check. [What kind of scientist fails to check or calibrate his instruments?] Then explain how much of the data is made up to fill gaps. Then explain how much of the warming trend comes from "adjustments". And if they can handle it, run through some of the bizarre adjustments done to the data before it even makes it into the database. Talk about how UHI is handled. Then explain that the adjustment process includes a lot more after that. Even committed AGW believers are shocked.
Bottom line when the dust clears -- they'll want to know who it was that told us the science was settled when it turns out that so much of it was just crap? There will be hell to pay. And reputations will be in tatters.
The explanation of how the USHCN version 2 is created is explained here and the data is available. http://www.ncdc.noaa.gov/oa/climate/research/ushcn/
chefio notes that the number of stations in 2008 is vastly reduced and this matters. This is not the final analysis for GISS of 2008, it takes time to QC and adjust the data. GISS has the intention of continuing to use the same network of stations. The low number is temporary
Over at Real Climate Eric Steig has a few comments on this.
http://www.realclimate.org/index.php/archives/2009/12/who-you-gonna-call/
I responded there as follows:
-------------------------------
Hi Eric-
The question Andy posed to me was not a question about climate science but a question about the social construction of science -- he asked me a "thought experiment" about what would happen if the temperature record were to change, would that "would that erode confidence in the keystone climate question"?
My statement on climate sensitivity is thus not at all different than your own: "radical changes to the long term trend in the surface temperature record would require re-evaluation of our understanding of climate sensitivity" -- which is of course why this topic is a political battle ground.
Your colleague Ray seemed to get this no problem:
http://dotearth.blogs.nytimes.com/2009/12/01/a-climate-science-forecast-in-the-wake-of-climate-files/#comment3
Your comments here and mine there at not at all in contradiction.
-------------------------
A second reply to Steig:
Roger Pielke, Jr. says:
Your comment is awaiting moderation.
5 December 2009 at 4:51 PM
Hi Eric-
Again, Revkin did not ask me if it was plausible that the temperature record would shift much, he simply assumed it in the question and I assumed it in the answer. It would be a little like me asking you, assume a major volcanic eruption occurred in 2010, what would be the climate consequences?
One response might be: “hmmm that is an interesting thought experiment. What might the answer be?”
Another might be “Eric is not a vulcanologist, why ask him this? Besides the chances of a big volcanic eruption are really small.”
Your post is quite a bit like the second response.
Again, I said nothing more than you said yourself: “radical changes to the long term trend in the surface temperature record would require re-evaluation of our understanding of climate sensitivity”.
Are radical changes possible? Don’t ask me, I’m a political scientist;-) But in any case you should be heartened by my answers, which explained that even if this occurred, it does not much change the science or policy equation — from the perspective of the social construction of science — which I take it is just about where you come out asking a different question about plausibility. So it appears we are in violent agreement from very different perspectives.
Post a Comment