Bishop Hill Models, data and the Arctic
Mar 10, 2012 Climate: Mosher Climate: Surface While I was at the Met Office the other day, I had a very interesting presentation from Jeff Knight, who IIRC runs the season-to-decadal forecasting unit there. Jeff's talk included discussion of a paper that Lucia had looked here. Lucia's comments are, as always, very interesting and address a range of concerns that I'm not going to touch on here. The paper itself is here.
The headline image of Jeff's talk was what the Met Office refers to as "the Celery Stick" graph, which assesses the range of model predictions against data:
;)
ENSO-adjusted global mean temperature changes to 2008 as a function of starting year for HadCRUT3, GISS dataset (Hansen et al. 2001) and the NCDC dataset (Smith et al. 2008) (dots). Mean changes over all similar-length periods in the twenty-first century climate model simulations are shown in black, bracketed by the 70%, 90%, and 95% intervals of the range of trends (gray curves).
As can be seen from the graph, CRUTEM is on the cusp of falling outside the envelope of the projections of the models, although this was surprisingly of little concern to the Met Office. As we saw a few weeks ago, the new version of CRUTEM will be issued soon, and will have an increased warming trend, which is apparently the result of a reassessment of warming in the Arctic. This involves a shift from the "admit our ignorance" approach used now to a new one incorporating a GISS-like extrapolation. The enhanced warming will shift CRUTEM back towards the centre of the envelope of the models.
This bothers me. It seems to me that for assessing model peformance, you should only compare models to data, not models to data-plus-extrapolation. If we don't know anything about the Arctic, then we should presumably test models versus data only for those regions of the planet outside the Arctic.
Bishop Hill
Richard Betts in the comments advises that the new temperature series will not involve extrapolation. I'm scratching my head here, because I'm sure we discussed this while I was down in Exeter, but I'm going to defer to Richard.
Reader Comments (68)
Another divergence problem about to be resolved by a trick. When will someone begin to hold these guys to task for their bodging.
As an engineer, I'd have to say that the only reason I'd have to present data in such an obscure and twisted manner is to hide the fact that my conclusions are bunk.
ENSO adjust global mean temperature changes since a year. 70, 90 and 95% intervals of the range of trends. Yeah right.
All it shows is a commonality of the fiddling and tinkering to get the predetermined 'right' answer.
It's this same 'ole climate technician's gambit. As I told Richard B, they simply do not know the basics of the scientific principle. They have little or no knowledge of the fundamentals of statistics and they do not know that they do not know these things. The sacresinct nature of data is completely lost on these people and, finally, you are allowed to invent data where it doesn't exist to make the climate meet your expectations.
I would tear my hair out if I had any to spare.
Oh and Richard, I'm a fully qualified electrical, electronic, radio and telecomms engineer as well as a ChPhys (form) MSc. I worked for myself and a world renowned private research company. And yes I get very annoyed when the Met off spouts it's crap about the reliability of their model(s).
Where would climate change science be without hindsight?
Well...that even makes sense to me that you should only compare a model to data and NOT to something you made up to justify your religious belief in Mann Made Global Warming (tm). Then again, we are dealing with creationists here NOT rational people.
Mailman
Bish
This bothers me. It seems to me that for assessing model peformance, you should only compare models to data,
I love your phraseology Oh and also your site.
If we don't know anything about the Arctic, then we should presumably test models versus data only for those regions of the planet outside the Arctic.
Absolutely. It's Hansen all over again and Richard says there is no government controlled agenda at the Met Off. My a$$.
I agree with chuckles - what an obscure way to present the results. Did nobody at the presentation question the use of extrapolated "data"? Does everybody at the Met Office accept this perversion of science?
We've had the hockey stick and now the celery stick. What next, the fiddlestick? - oh wait.
I have no idea what that graph shows, a time warp has been applied to the x. A celery stick eh, a few more carefully chosen ingredients and I'd conclude a Waldorf Salad.
BH - Are you sure that the new Crutem record will include extrapolation? I was under the impression that it merely included a number of new Arctic stations (mainly Siberian) to better represent Arctic warming.
Yes, it looks suspicious, but it maybe that suspiciousness lies in our looking and not in the motivation of the data-compilers. Why not ask Richard B for confirmation unless he is deemed suspicious too?
As the saying goes , torturing the data until you get the results you need . And once again an approach normal unacceptable if done by an undergraduate as part of their course, is acceptable as the 'best' scientific standard for the area.
Anteros (Mar 10, 2012 at 9:23 AM)
“I was under the impression that it merely included a number of new Arctic stations (mainly Siberian) to better represent Arctic warming”.
There was a suspicion that in the bad old (colder) Soviet days, the Siberian stations exaggerated the cold to boost their fuel rations. And when Anthony Watts had a shufti on Google Earth,he found a Siberian station in a town where the streets have open air gas-fired central heating. But I’m sure CRU will examine these things dispassionately.
On the graph, is the y axis upside down or the x axis back to front, or is that normal when you’re comparing models to data?
As I have said before they are having to make it up, and make adjustments as they go along. The recent changes by NOAA to GCHNv3 and by Hansen to GISS make these these datasets a joke. e.g. here are a few links:
http://notrickszone.com/2012/03/01/data-tamperin-giss-caught-red-handed-manipulaing-data-to-produce-arctic-climate-history-revision/
http://notrickszone.com/2012/03/07/reykjavik-iceland-deliberate-temperature-manipulation/
GISS temperature adjustments and cherrypicking - http://notalotofpeopleknowthat.wordpress.com/2011/11/19/giss-and-their-african-temperatures/
Greenland - http://www.cru.uea.ac.uk/cru/data/greenland/vintheretal2006.pdf :
“the warmest year in the extended Greenland temperature
record is 1941, while the 1930s and 1940s are the warmest decades.” - see also: http://www.joelschwartz.com/pdfs/Chylek.pdf
http://notalotofpeopleknowthat.wordpress.com/2012/02/28/iceland-adjustments-spread-to-norway-and-russia/
I have also looked at the recent changes to the historical records for a few North Atlantic/Arctic stations and found that the adjustments always seem to make the historic years colder and the recent years warmer. Coincidence? Aye right.
As with all these trend adjustments the past gets colder, the present warmer, the future hotter. The data is simply being adjusted and invented to fit the models. This is how politically correct climate science works and gets done. No other scientific discipline would tolerate such behaviour.
A real scientist would say, "I don't know" when confronted with a lack of robust data, "but I will find out by doing lots of experiments and by taking huge amounts of readings with my instruments". You don't see that behaviour amongst climate scientists, where less is more and nothing is best of all.
http://ocean.dmi.dk/arctic/icecover.uk.php
http://notalotofpeopleknowthat.wordpress.com/2012/01/25/how-giss-has-totally-corrupted-reykjaviks-temperatures/
http://ocean.dmi.dk/arctic/icecover.uk.php
http://notalotofpeopleknowthat.wordpress.com/2012/01/25/how-giss-has-totally-corrupted-reykjaviks-temperatures/
geoffchambers wrote: "There was a suspicion that in the bad old (colder) Soviet days, the Siberian stations exaggerated the cold to boost their fuel rations."
Is this suspicion founded on hard evidence or idle speculation?
I must admit I can't make head or tail of that graph. It looks like the trend is up however. That can't be good.
TPX - I can't provide hard evidence but I have read this also. It is at least possible, it is not as if someone from the Kremlin would leave their relatively warm Moscow office to come and check what the actual temperature in some distant Siberian mining town or research station was.
I have also read that non-Soviet historic Arctic station temperature data are not always reliable, or good quality. Imagine you are by the stove in a relatively warm hut somewhere in Northern Canada, and it is your turn to go out into a well-below-zero blizzard and read the thermometer which is located 50 yards away. After you have faithfully done this a few times, would you or your research chums really care if the reading was -25C or -35C? I would not be very surprised if quite a few historical readings were guessed/approximated/fabricated accordingly.
Nowadays most stations will be automatic sensors, and to keep the cable lengths short will no doubt be located close to relatively warm buildings, increasing the potential for the UHI effect in the data.
Weather not climate but there's not much warmth in Nunavut today.
Also, what is the justification for removing El Nino/La Nina signal (assuming that's what "ENSO-adjusted" means) from the global averages on one hand while including 'extrapolated' Arctic figures on the other hand? How many times must the data be adjusted to fit climate model projections? If I am speaking out of ignorance here, could please a climate scientist enlighten us?
Bishop Hill wrote:
When were the model predictions in this chart made? If it was in 1984 I could see some use in such an exercise. If it was 2008 I can't.
I've seldom seen such a poor, obscure presentation of data as that chart, and it's no better in the context of the original paper that you link to. I've had a brief read through the paper, but find that Lucia's comment at the end of her analysis sums it up well:
"I don’t think they are necessarily worth discussing in any detail. Based on the above, it seems to me that, at best, Knight et all showed that:
If modelers examine a set of 10 runs with different long term trends resulting from a variety of parameterizations and combination of the SRES, all of which seem to have exhibited ‘long term’ warming trends below the average 70 year trend for the multi-run mean of runs forced with the A1B SRES in the AR4, then the current slow down in warming falls in the range of variability of those lower-than-projected-warming runs. "
OK, I've been watching the footy and probably had too many beers to read the graph correctly. But doesn't it suggest the models predict a "temperature" (presumably some sort of non-intensive anomoly measure) of 0.5 deg in 1984 and zero in 2008?
That is a decrease. Aren't we meant to be frying?
That's a deliberately verminous graphic which is designed to confuse - and is a particularly Mannian anticommunication tactic.
We are the priesthood, you ignorant plebs are unworthy of the secrets of the Cause - believe or we will excommunicate you.
Utter drivel.
@Cumbrian Lad
Glad to see Lucia has clarified what they were trying (not) to show!
With a background in modelling non-linear dynamics and CFD in the automotive industry I am shocked at the standards of so called 'validation and correlation' used by the UK Met Office - these people would be a liability anywhere near a safety-critical function. We cannot currently accurately model the airflow around an F1 car let alone the global circulation around our planet.
An analysis of the model outputs I suspect will show they represent few of the modal oscillations of the planets climate and therefore cannot be considered to have an accurate representation of the physics involved and hence the system response to either CO2 or solar forcing.
Also one can only assume the extrapolation algorithms to be used within CRUTEM are going to be the same as those used inside the models to ensure the closest possible fit between outputs!
This would indicate to me that ENSO events are viewed as short term and have no long term influence.
This is an incredible statement. Because the output of the models match what we expect then there is no need to look at anything else.
I see a science-fiction opportunity. A government laboratory has found a way to erase the centre of shame in the brain, and applies it to all key employees. The resulting shameless outputs attract criticism here and there, but their Leader, who is both without shame and without scruples, is there to shrug off the attacks and remind the employees that since they are saving mankind from itself, they can do what they jolly well like. Within 'reason' of course.
Back in the real world, and maybe not off topic, see this report of a lecture by someone who found several opportunities to trigger the centres of shame in a lot of young brains quite recently: http://wattsupwiththat.com/2012/03/10/moncktons-schenectady-showdown/.
Back in the make-believe. Maybe I could use a character like that if I want my fictional tale to have a happy ending? Could he rally the young to revolt and get that laboratory closed down or converted to work that would actually benefit humanity?
lapogus wrote: "I can't provide hard evidence..."
Then quit blowing smoke.
The long sunspot cycle 23 predicts a significant
temperature decrease in cycle 24
Indicating cooling rather than warming is in the pipeline!
GISS Surface Temperature Analysis
From Hansen himself:
There is an expectation that the Sun will return to levels of cycles 22-23 for the peak of SC24. Which is unlikely.
Notice the reliance on ENSO and Sol to retain the underlying warming trend?
TPX
The whole point of "suspicion" is that there isn't evidence. The suspicion that Soviet apparatchiks exaggerated temperatures has been circulating for a long time. Of course, it could just be idle speculation.
I speculate that that is the sort of behaviour that one would expect given our knowledge of how the Soviet Union operated.
What exactly is your point?
To offset the exaggerated temperature rise which is the result of bad heat transfer assumptions, the climate models use twice real optical depth for low level clouds AND incorrect net AIE which is presently near zero but was positive at the end of the last ice age.
In time Betts will have to acknowledge the failure, but it'll be done with subtlety
TPX,
There was a comment from a U.S. Scientist over at WUWT some time back openly admitting during his service, (50s/60's I seem to recall), of exactly the story, of not going outside to bother checking the readings. You stop blowing smoke now, go over to WUWT and look back for the openly honest comment! I seem to think Polar Bears were involved as well so it should fit right in with your suspicions.
SNTF at 10:55. I'm not sure what the emoticon for a sheepish grin is but consider one inserted here! I thought Lucias comment was a lot clearer than the paper was, if that helps :)
My own paraphrase would be: they've selected a set of the lower temperature rise predictions, shuffled them around a bit and added a few sprinkles of magic dust, and found that actual temperatures just about sit within the error bars.
lapogus wrote: "Imagine you are by the stove in a relatively warm hut somewhere in Northern Canada, and it is your turn to go out into a well-below-zero blizzard and read the thermometer which is located 50 yards away. ... I would not be very surprised if quite a few historical readings were guessed/approximated/fabricated accordingly."
Mike Jackson wrote: "I speculate that that is the sort of behaviour that one would expect given our knowledge of how the Soviet Union operated."
Pete H wrote: "There was a comment from a U.S. Scientist over at WUWT some time back openly admitting during his service ... of not going outside to bother checking the readings."
Ideology is irrelevant. That is just plain lazy. We cannot dismiss the entire record by a single example, assuming it's true.
TPX - and we cannot assume the recent and much hyped rise in Arctic temperatures is based on data which is 100% accurate. Especially when scientists with an ideological agenda then compound potential errors through dubious homogenisation and extrapolation techniques.
Not forgetting about blatant but convenient adjustments to the raw station data - http://notrickszone.com/2012/03/01/data-tamperin-giss-caught-red-handed-manipulaing-data-to-produce-arctic-climate-history-revision/
"As can be seen from the graph, CRUTEM is on the cusp of falling outside the envelope of the projections of the models..."
Are we looking at the same graph? I can't make head or tail of the thing.
Who's "dismissing an entire record"?
Geoffchambers' first reference — the one you decided was to be your troll-subject-of-the-day — was about well-known suspicions about Soviet officialdom, which was the one I replied to.
Lapogus' comment made no reference to any particular nationality; he simply cited a possibility using Arctic Canada as an example.
Pete H suggested that there had been failing on the part of the Americans.
It looks to me (having second thoughts) that perhaps we should be dismissing an entire record given that two of the largest nations on earth (three if you add in the whole of Canada) appear to have been producing "off the top of the head" temperature readings for decades.
Ideology is indeed irrelevant (unless of course you're a climate psyentist in which case it's compulsory) and human nature will trump it every time!
How does that play with you, Tippex?
@Cumbrian Lad 12:43
Spot on, if Lucia's clarification of the sophistry was still that opaque (a single sentence covering five lines with and only two commas) and I know how good she is (i.e. very), then I really don't have the life-force to suffer the non-abridged version. :-)
As Shx says, the idea of 'ENSO-adjusted' temperature is a bit of a nonsense. The ENSO index is just a measure of ocean temperature. So they are looking at temperature with some of the temperature subtracted. There are now so many adjustments going on that they have lost touch with reality.
Did they explain why the graph only goes up to 2008?
By the way , the CRUTEM4 paper is now out (since January 2012).
Hemispheric and large-scale land surface air temperature
variations: An extensive revision and an update to 2010
P.D. Jones (1, 3), D.H. Lister (1), T.J. Osborn (1), C. Harpham (1), M.
Salmon (1) and C. P. Morice (2)
Note that (3) is Dept of Meteorology, King Abdulaziz University, Jeddah, Saudi Arabia.
I am sure we would all like to congratulate Professor Jones on this distinguished appointment.
You can read the whole paper here. There is also a pdf version but I can't get that to load.
Remember, CRUTEM is the land-only part. HADCRUT is the combination of land and sea. (Andrew please note!) The HADCRUT4 paper is not out yet but I have a copy of it.
Only in climatology does time run right to left.It is a tough life, obscuring the facts, but someone has to do it - for the good of humanity.
"Near-zero and even negative trends are common for intervals of a decade or less in the simulations, due to the model’s internal climate variability. The simulations rule out (at the 95% level) zero trends for intervals of 15 yr or more, suggesting that an observed absence of warming of this duration is needed to create a discrepancy with the expected present-day warming rate."
This is the critical piece of the MET report that was so damaging to the IPCC models and the reason why the MET was forced to either change the data to reflect more warming than the record currently shows to protect the IPCC models or admit that the IPCC models have been falsified at the 95% level which would result in killing the goose that provides golden eggs. Does anyone expect these "honest" scientists to admit the models have failed?
The next phase of the agenda will be to try to discredit the satellite data because the last 17 years of RSS data also falsifies the IPCC models.
The Arctic region poses a problem, as temperature measurements are scarce in the region. GISS extrapolate over the areas where there are no measurements, which always has given me pause, as most of the extrapolation seems to be from land stations to ocean area. Crutem simply leaves the unmeasured areas out of the average.
However, the global average per se is of no intrinsic interest. It is one metric (arguably an inferior one, but the most common one certainly) by which we attempt to quantify global warming, and with which we assess the accuracy of models' predictions. Given the difficulty of measuring Arctic temperatures, wouldn't it be better to use a metric which is the average surface temperature over the more accessible parts of the globe, say between the Antarctic and Arctic circles? It's certainly just as simple to compute from the models' output as a global average, and more reliably observed. I might add that -66.5 to +66.5 latitude coverage includes about 92% of the global area (and virtually all of the population), so it wouldn't be as though this would be an irrelevant metric.
"If you can't dazzle them with brilliance, confound them with BS!"
Classic example. Chuckles and several others are absolutely right about the Celery Stick and where it should go.
TPX (Mar 10, 2012 at 10:02 AM) asked me whether the suspicion about Siberian temperatures I mentioned was founded on hard evidence or idle speculation.
Idle speculation on my part, based on what I read on blogs.
In the four years I’ve been practicing this form of research, it’s proved infinitely more reliable than reading accounts based on peer-reviewed science.
I find this situation disturbing.
You know the saying - if it looks like a duck and walks like...........etc.?
Well I can remember loads of times when after several parameterisations, only the modeller thought it vaguely resembled a bird and that was at 6 days!
The trend from 1999-2008 was +0.07 deg C per decade but this is attributable to the trend in ENSO. Adjust for that and the trend is 0.00. Does anyone know what this adjusted trend from 1999-2011 is? It doesn't really matter what the models say if the observations stay flat.
Before I scream and tear my remaining stubs of hair out will someone, anyone in any Met off anywhere in the world please release the unadjusted, untampered with straight from the instrument data. Just publish it on the web we will find it. ANYONE RICHARD HELLO !!!
"Celery Stick Graph". I love it, nice phrase.
It just goes to show the old adage "A picture can speak a thousand words" also works reversibly: A good picture can hide a thousand words [or numbers]. When Richard Betts kindly provided a link to the Met Office decadal predictions I bookmarked it. I also saved a copy on the hard disk.
I was looking for some data on atmospheric CO2 concentrations recently, and also wondered why some of them only went up to 2008. Does anybody have a link for up to date figures?
I had a read of Lucia's thread about this paper earlier today, and have been thinking about the comment by tetris, which I will repost here:
Was this pre-climategate admission of the models' inadequacies a one-off?