Thursday
Jan102013
by 
Bishop Hill
Bishop Hill Spot the difference
Jan 10, 2013 Climate: MetOffice Climate: Models I'm still suffering. Even whisky isn't working. It must be serious.
In the meantime, Paul Homewood has found something interesting about the Met Office's forecasts.
Reader Comments (216)
Richard, thank you for your comments here. In one comment you state:
"These decadal forecasts use "initialised forecasting" techniques, ie: the models are started at the observed state for the current time "
Can you say:
(a) which year's data and date is used for initialisation?
(b) Is any later data used in this process? ie re-intialising during the run/constraining with later data?
(c) Is this a single model run is or is the model run many times and this particular one selected? If so, how many times is the model run in order to select this case?
Regarding a further comment you made:
"Yes, "retrospective forecast" means hindcast here."
My understanding is that, because of processes such as diffusion it is not possible to run the models backwards (ie with negative time step). Therefore:
(d) what is meant by this phrase? Does this describe a parameter matching to the data (in Reservoir Modelling this type of approach is referred to as "history matching")?
Would be very interested to hear your responses.
I still think that the MO tried a cheap PR stunt by issuing this when they did and with the level of confusion discussed above, given that it would be too complex for the MSM. This is is not honest and transparent science from a (supposedly) trusted tax payer funded public science organisation.
Shame on all its managers. This is a disgrace.
Richard, if the white line is a hindcast, why is it described as a "previous predction"?
For the umpteenth time, the cynical attitudes expressed here are a direct result of this type of behaviour from the met office.
"Even whisky isn't working."
Double the dose and go to bed for a week. Irish whiskey might help - take a double dose of that too.
I suspect Richard was not best pleased by the way the metoffice mismanaged the whole shambles, not least since he had to intervene in the (by mistake on purpose?) misleading Guardian/BBC reports as well.
I must say that I think there was a bit more than computer run times/costs behind the limiting of the model projection to 5 years. Not least the more dramatic difference that would result from the previous projection, but mostly the political hot potato.
Bishop you may try the Hat Cure. Bequeathed to me by my grandfather.
Make some hot spiced wine.
White wine, about 3/4 liter.
Sugar
Cinnamon
Cloves
Heat wine slowly, do not boil. Break a cinnamon stick, toss it in the wine. Add half a dozen cloves. Add sugar slowly till it tastes pleasant. Go to bed and put a hat on one foot. Drink hot spiced wine until you see two hats. Go to sleep. Next morning you will be fine.
This is a clumsy attempt by the Met Office to reposition themselves. The lack of warming is a big problem for them. The politicians who passed the Climate Change Act, an Act designed to decimate UK industry, cast pensioners into fuel poverty and litter the countryside with useless bird chomping windmills, will be quick to blame the Met Office for the overconfidence it placed in its worthless computer models when they finally realise that they have been fooled. The sooner the Met Office give up the propaganda and start acting like scientists, the better for everyone.
Paul Matthews:
But with Delingpole in the Daily Mail it isn't just us. Betts I hope has a good future but changes are surely afoot. It's just hard to tell which head of the hydra will be cut off (or, shall we say, revised downwards) first.
Just made my first visit to Paul Homewood's website.
I noticed the sub-header on his homepage:
A discussion of that statement would be a significant post in its own right.
Applied, it would devastate the IPCC: 1) assessment processes wrt lack of any meaningful openness and transparency, 2) biasing of processes for personnel selection, 3) cherry picking of research uncritical of alarming AGW by CO2 from fossil fuels.
Well done on your website Paul Homewood.
John
But, isn't it quite normal in climatology to re-predict the past in order to conform to the basic theory of climatology? (cf. the disappearing MWP).
By the way, is this model, with its improved prediction of the past/hindcast (whatever we're calling the white line these days), is that the model that the Met Office is testing every day in their weather forecasts (as Julia Slingo told parliament)? Or is that a different, even more elaborate, past/future prediction program?
In 2007 how many CPUs did the MetOffice have? In 2012 how many CPUs does the MetOffice have? How much more computationally expensive is the new model?
Joe Bastardi posts graph of temp anomaly vs CO2 since 1990:
intriguing graph
Am I reading this right?
How is the near-term natural variability not relevant to global warming? Imagine you are in the year 1988. Hansen walks up to the US Senate and tells the world that it would warm. The next 10 years of warming are taken unequivocally to be due to anthropogenic CO2. These ten years could very well have been due to 'natural variability' but that explanation was never allowed, was it? And now it is?
And it is not just natural variability at decadal scales. The current centennial warming trend began in the 18th Century, and has continued for most of the time since then /the truth about Alaskan glaciers - the long slow thaw...
"Heads should roll at the Met Office." --Schrodinger's Cat
Please, let's all refrain from such imagery, lest we be thought Jacobins.
Much gratitude to Richard Betts, who restores my confidence that members of that organization do, indeed, have heads.
Green Sands, maybe Richard is feeding your input back to the met office, but it is then treated like most IPCC review comments - "no change needed" :-)
Richard Betts at 11:19 PM said, when talking about the new model,"it's not yet been shown to be useful to anyone".
Can someone (anyone?) please tell me in what way the old model was "useful to anyone"? (apart from getting research grants and green subsidies for windmills etc) Oh, hang on, ok, yeah, forget it, I've answered my own question!
Bish
It must be serious.
It' s Man Flu, Bish, and this is serious indeed. I know - I am recovering, so there is hope.
I'm getting preety cynical now, met office replaced whole decadal forecast page, with a new one, same url.
Public have no way of knowing or guessing, that the one they see now, is fifferent to the previous ones.
And I'm with Paul Matthews on hindecasts as forecasts/predictions
Richard Betts, I have a further question for you:
(e) If a model is run from the same initial conditions do you always get the same result/output or do you get different results?
I'm more grateful than usual to AlecM (7:14 AM) for his input into The chocolate teapot.
I think we have reached a point in the whole debate where the Science and the politics have reached a cross-roads, the Science no matter how far the numbers are tortured to fit the theory can no longer support the CAGW hypothesis in the face of new information from other areas (solar, cosmic rays etc) nature is no longer playing along throwing the whole alarmist CAGW climate science industry into disarray.
The political, Media, business and social culture that has grown around the CAGW cause has now reached a stage where it needs to either force the Climate Science it has fostered and grown to produce and promote the cause to continue the farce or it throws the Science under the bus where it is in an effort to maintain the CAGW meme as it currently stands long enough for it to become a faith based belief system and the whole CAGW becomes a new creationist type belief debated about as usefully as creationists debate evolution.
That is my take also. I can almost forgive their over-egged models but not the mendacity. Taxpayer-funded organisations should not behave like this.
Close it down.
Hi folks
Just quickly dropping to say that I acknowledge your comments and will try to read them properly later - I'm in meetings all day today and (as mentioned before) also under deadline for IPCC so have very little time!
If this thread continues to fill with more questions for me then I'll find it difficult to answer them all, but I will do my best to answer ones that have come in so far :-)
Thanks for all the comments and feedback though, it is very useful to be able to see what people think. Thanks for all your interest.
Cheers
Richard
The next couple of years will prove interesting as the HadGEM3 model still applies the current IPCC climate sensitivity even though the forecast output is now lower.
If that sensitivity is too high then even this forecast is too high, it is still warm biased!
I might also add to those thanking Richard for his perseverance, not only with the media but with us as well.
Has it ever been anything but a faith based belief system masquerading as "science"?
It is here to stay.
Key point Lord B. Once the IPCC has fully imbibed the latest findings of Nic Lewis then what kind of downward adjustment from the Met are the good readers of the Daily Mail - and every other part of the MSM, by then - going to have to chew on?
The fact that recalibration leads to such different post-dictions suggests that the model itself is not at all robust. I should not want to use it for prediction.
The basis of CAGW is identical to all the major faith-based religions. Humans are evil and must change their ways to satisfy the wrath of the gods.
As for the Met Office, as pleasant as it is to have Richard Betts explaining what has gone on inside the tent, outside there has been spin and obfuscation for years selling a political agenda. If the models had shown what Richard tells us they show now 15 years ago the output would have been binned and the scientists ordered to go back and find out what they'd done wrong. Now, after nearly two decades of hiatus in temperature rises with a further 8% of CO2pumped into the atmosphere the new outputs are a godsend, giving the activists that run the Met Office the opportunity to postpone thermaggedon for a few years in the hope that temperatures start rising.
I must say that I think there was a bit more than computer run times/costs behind the limiting of the model projection to 5 years. Not least the more dramatic difference that would result from the previous projection, but mostly the political hot potato.
Jan 11, 2013 at 12:03 AM | Registered CommenterPharos
Pharos,
I'm a cynical old *astard and agree that the potential for a 10 year version of the model to continue it's sharp downward trend (see the graph) is enormous. Either it was cut short out of fear, or edited back to 5 years once the results didn't go 'as required' ?
I think Richard Betts has fully explained the misunderstanding regarding the white lines - it appears to be simply a misreading of a couple of words, no malice or jiggery-pokery, but perhaps sloppy language, from the Met Office. Thanks are due to Richard for taking the time to clear it up.
One of the Met Office quotes (can't locate it at present) said that they were currently trying to understand why the newer forecast was significantly lower than the old one.
I assume this means that using HADGEM3 model, with all the latest information plugged in (why would you not run it with the latest data?) it clearly shows a better hindcast, and thus is probably a better model. However this does not mean that the reduced forecast was expected. Given the complexity of the model, I'd guess this isn't unusual.
Richard, is it the case that the exact reason that the better model produced a lower forecast is still not fully understood? And was it limited to 5 years instead of the previous 10 due to the computational run requirements, or did it start to deviate too much to be believed?
remember the previous graph and forecast, on the exact same url. was used in publications, the press and publically.. it also reinforced Popes predictions of next 5 years half will be hottest, predictions.. these have now been vanished.
http://www.bbc.co.uk/blogs/paulhudson/2013/01/met-office-scale-back-global-w.shtml
Previous graph used in literature like this (same url as new one):
http://www.metoffice.gov.uk/media/pdf/6/9/PresentationMOSAC_16.7_Scaife.pdf
so policy makers, when looking at the decadence forecast webpage, will have no idea that the old one existed (no links, no comment)
so if it is suggested that the new one is just experimental (yet same url) I become cynical, then they will be misled that the Met office previously predicted current temps, and a lower projection...
despite as the pdf shows a year previous the predictions and outputs were much higher.. (also not knowing that what is described as previous predictions, actually means 'hindcasts' !! )
BBC's Paul Hudson (ex Met Office, relevant degrees, an extract)
http://www.bbc.co.uk/blogs/paulhudson/2013/01/met-office-scale-back-global-w.shtml
"One of their most high profile forecasts came in late 2009, coinciding with the Copenhagen climate conference.
It stated that half the years between 2010 and 2015 would be hotter than the hottest year on record, which I wrote about on my blog.
This already appears wide of the mark.
The latest projection seems to address this error with a prediction to 2017 in which temperatures rise 20% less than previously estimated.
In November 2009 I wrote about this levelling off in global temperatures, using research available at the time on the Met Office website.
In it, the Met Office explained that the levelling off of global temperatures that we were experiencing can be expected at time periods of a decade or less, because of the computer models internal climate variability.
But intriguingly, the research ruled out zero trends for time periods of 15 years or more.
The new projection, if correct, would mean there will have been little additional warming for two decades despite rising greenhouse gases.
It's bound to raise questions about the robustness and reliability of computer simulations that governments around the world are using in order to determine policies aimed at combating global warming."
-------------------------------------
Bonus points for the bright-red confidence intervals. No climate graph is complete without plenty of bright scorching-hot red.
The white line showed how badly their model was already failing on a "hindcast" basis before they introduced their new shiny model.
Any reputable modeller would have been pointing out this divergence some time ago and pndering on what that tells them about their model (i.e. why is it wrong??).
In essence, the Met Office have been sweeping under the rug their modeling failure for a while until they changed the model.
Then they swept under the rug the news that the new model - which corrected the problem they didn't want to talk about - leads to far less alarming conclusions about the future.
The errors are part of modelling. Entirely acceptable and understanble.
The way they have comminicated - or more accurately not communicated - these errors and issues is completely disreputable in my view.
I think the irresponsibility of climate science workers is highlighted here by shub ( 1:18 AM ):
It was irresponsible of Hansen to make such testimony and irresponsible of his backer Wirth to have opened so many windows the night before that the air conditioning could not cope during the hearing [http://www.pbs.org/wgbh/pages/frontline/hotpolitics/interviews/wirth.html].
It was irresponsible of the Met Office and others not to point out, loudly and clearly, that the subsequent warming decade was not necessarily due to the additional CO2. Lindzen, on the other hand, did show responsibility back then and since, and his sensible perspective is standing the test of time. For example, from his 1989 presentation at MIT:
[http://www.fortfreedom.org/s46.htm]
It was irresponsible of John Houghton to make so much of the MBH Hockey-stick plot within the IPCC, and irresponsible of the Met Office and others not to have thoroughly audited such a study purporting to erase the MWP before giving it such credence. Thank goodness there were responsible adults in the form of McIntyre and McKitrick to do that job later. Thank goodness also that we have an outstanding piece of writing on this part of the history of science. It is called ‘The Hockey Stick Illusion’ [http://www.amazon.co.uk/Illusion-Climategate-Corruption-Science-Independent/dp/1906768358]
It has been irresponsible for the Met Office to act as if computer models are capable of being run to give climate forecasts fit for practical guidance. Buried within even the IPCC documents are statements that they are not fit for prediction.
'So states the IPCC’s Working Group I: The Scientific Basis, Third Assessment Report (TAR), Chapter 14 (final para., 14.2.2.2), p774.' [http://junkscience.com/2012/08/24/does-the-ipcc-really-believe-anyone-can-predict-the-future/]
The pantomime-like record of Met Office attempts at seasonal forecasts would be laughable were it not for the harm they must have caused. It is generally better, after all, to have no forecast than to have a wrong forecast that discourages preparations against a wider range of plausible conditions. Stocking less road grit, and less de-icer at airports, are but two examples of responses to a forecast of a milder winter to come.
The biggest fail of all has of course been associated with the flatling of nominal global mean temperatures for the past 16 years or so - a spectacle that might well be puzzling the concerned citizen who has been soaked in exponentially rising temperature projections, and convinced that CO2 is the culprit. Mind you, I am not aware of the mass media going big on this flatlining, so perhaps my imagined citizen is not yet puzzled. But one day, that puzzlement will arrive and ought to be followed by considerable anger. But will a raging citizenry be any match for the unbounded sophistry of computer climate modelers?"Our computer models can do flatlining - we've always known that. You can expect that from time to time. So we're not wrong. We can't be wrong. The science is settled. One day the warmth will come. Just you wait and see."
And, of course, one day it will get warmer year after year as measured by global means. Cue a renewed crescendo of alarm about over-heating. Then it will get cooler again, year after year. Cue a new offensive spun around climate disruption making things colder. And, from time to time, those mean temperatures will not do much, year after year. Cue the hunt for and press releases about extreme weather events. Or maybe not. Maybe one day, we’ll just see that stuff on sandwich boards on the High Street alerting us to the end of the world.
@ Richard Betts regarding this comment of yours:
Quite so. it appears that the recent historical data falls clearly and consistently below the confidence bands of the HadCM3 derived hindcasts.
Why is this the first time we are discussing this point - when conveniently you can say "we've solved the problem"?
Where was the press release last year? The fact that HadCM3 was producing simulation that were too pessimistic was BIG news. Much more important than news of a new model. In fact even more important than the news that HadGEM3 is producing more benign simulations.
People - taxpayers - have a right to feel very aggrieved about this.
Bish, have you tried a hot whisky? 50/50 Whiskey and hot water. Spoon of sugar or honey and some cloves. Works for me in extremis.
Pointman
And a slice of lemon.
Re whisk(e)y
I was given a bottle of Bushmills at Christmas and I have to say it has rather transformed my opinion of whiskey.
I'm fond of Canadian Club with ginger ale as a smoothly pleasant long drink, but it's so smooth as to be quite nondescript and one could equally well be drinking Pimms. Jack Daniels and indeed all American ersatz whiskies I find just revolting. Someone gave me a bottle of Jameson for my 21st and I poured 3/4 of it down the sink, to which it seemed better suited. I've sampled a variety of Scotch whiskies at a Burns Night or two and found them interestingly variable; until sampled alongside one another, I hadn't appreciated how a whisky could be smoky, or peaty, or fruity etc in flavour.
Bushmills, I have to say, I found extraordinary from the moment I opened the bottle. It just exudes class. The closest I've found to it in impact is Mount Gay Silver Rum, which bizarrely it quite resembles in nose and depth if not actual taste. In both cases one can sit and happily sniff the fumes for ages, they are almost as satisfying as the drink itself - a bit like one can with well-roasted coffee.
Mount Gay makes all other white rum smell and taste like paint stripper. Bushmills does the same for every other high street whisky. Have any cognoscenti any other whisky recommendations? I have long wanted to get into whisky properly as a retirement hobby but as that's still a long way off I'm quite minded to crack on now. I find the greater the quality the smaller the quantity that satisfies.
Logically, as the climate in parts of Japan is similar to ours, Japanese whisky ought to be quite good. Is it?
I have a number of comments held in moderation at the Met Office news (some since yesterday)
One that I have just added(pending moderation)
http://metofficenews.wordpress.com/2013/01/10/addressing-the-daily-mail-and-james-delingpoles-crazy-climate-change-obsession-article/#comment-893
"The new graph, implies the Met Office accurately forecasted the eruption of a volcano Mt Pinutobo (before it happened) and the also forecast resulting temp change very accurately (compared to prev graph)
quick somebody tell the volcano experts the Met Office can predict major volcanic eruptions!!
Why are the outputs of a new model producing hindcasts, described by the Met Office as ‘previous predictions’ (the white lines)
old graph:
http://notalotofpeopleknowthat.files.wordpress.com/2013/01/image19.png
new graph:
http://notalotofpeopleknowthat.files.wordpress.com/2013/01/image18.png
-------------
I've had one comment at Met Office News blog held for 15 hours now, others coments, also with urls, have appeared: ( I reproduced here)
http://www.newstatesman.com/sci-tech/2013/01/was-there-anything-james-delingpoles-daily-mail-piece-which-was-true#comment-762868781
The MET office, what a joke.
How do you confuse a MET office employee? Ask them what the weather will be next week. (Which would be funny until you realise this is their main job).
By the way they've said it will be snow a\ll over the country for the next two weeks, I'm sure you've seen the reports in the paper based daily drivel.
Pointman, Geckko, Justice4Rinka and many others: I worry for our host that, due both to the severity of his fever and his commitment to the experimental method, giving all of the excellent suggestions on this thread a try may lead to something more than George Steiner's recommended 'two hat' outcome. Eighteen hats and a startling vision of a complete solution to the Navier-Stokes equations seems more likely - but will the man be able to recount the tale many days hence, when he has finally returned to us. :)
Martin A wrote:
I agree. The Church is still here five hundred and 150 years on from having had its cosmology and palaeontology comprehensively demolished.
To a catastrophile, the fact that there's no evidence of warming is no reason to think there's no warming; 20 years of data are enough to prove warming will continue but 20 years of flatlining are not enough to prove it won't; and so on.
Environmentalism is simply malevolent lunacy.
"It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead."
Barry,
I don't think the MO claim to predict volcanoes. I think that the new "retrospective forecasts" not only use the new model, but the most up to date data, so they search for a "best fit", consistent with their physics, that matches all historical data. I think that is what they mean by hindcasting.
The only way to judge the forecasting capability of MO models is to compare their predictions at the time (with the best model and up to date data at that time) with previous and later predictions. If I am correct in this then the worst that the MO can be accused of is confusing terminology.
Of course, none of this shows any predictive capability. RB now says the models are "experimental". It is a shame we bet the farm on experimental models, when the earlier ones have now been proven to be incorrect.
Four Roses
Despite Professor Betts’ best efforts, it’s still not clear what’s going on.
Roger - i was being sarcastic.. of course the met office can't predict volcanoes, the original now replaced decadal forecast graph shows they cant (big gap oberserved temps, vs projections because of the volcano
what they can do when they run a new model retrospectively, is of course include that information..
the sarcasm came about, because they label the new graph, the white line model output, with the volcanic factor retrospectively included, now matching observed temps, is described as 'PREVIOUS PREDICTIONS'
Thanks Barry - I now understand your point. The term "previous predictions" here is not misleading, it is factually incorrect and must be corrected! RB please note.
Well lots of good stuff on the models and who said what to whom and when. A nice fireside chat with Richard Betts. Thank you Richard for at least talking to us realists (sorry sceptics) - it is very much appreciated. We are not all nutters, as I am sure you have found out, and there is a fair bit of knowledge and expertise, both of modelling and climate/meteorology amongst the assembled commenters.
BUT the fact remains that the climate models have shown very poor skill at predicting future temperature trends for most of the last 25 years. OK early on as the temp continued to rise, but divergent from reality from the mid 90's on, looking ahead.
Global temperature is now below even the standstill predictions (no further CO2 emitted) of 20 - 25 years ago. I think we are all agreed that there has been no warming for 15 years, despite MSM steer droppings pronouncements fed by those who wish to obfuscate and drive an agenda that has little to do with science.
In that 15 years it is estimated that 25% of the CO2 ever emitted by man has been released.
There never has been any real evidence that CO2 is a major driver of the Earth's climate, and scientists such as Lindzen have been proven right so far.
If the latest MO model is now correct in the short-term that will be 20 years of no warming during accelerating CO2 emissions - are we to believe that any natural variability and "cooling effect" will have exactly matched the CO2 "warming effect" in that two decades? Seems a stretch.
What annoys sceptics most (IMHO) is the continued insistence, including advice to government, of a warming that has not happened. The policies of government(s) developed on the back of this have lead to the spending of billions we do not have to solve a problem that (a) may not exist (b) anyway, we probably can not affect.
Deaths in the Third World, rising fuel poverty in the first world (not only the UK), increasingly uncompetitive industries in the first world (Europe especially) at a time when we are struggling to make our way. The lunacy of bio-mass burning and bio-fuels, not to mention windfarms that only the salesmen thinks will fit the bill. Committed to a Climate-Change bill that will run into £100's bn. These are all predicated on forecasts that the Met Office have said are beyond question - the science is settled.
Surely it is time Richard Betts that somebody from the Met Office had an HONEST talk with the politicians, whose understanding of science is as good as mine is of brain surgery.
I may be wrong (my wife says I often am) but if reality and proper science, that takes note of the real data, doesn't return to Climate Science, it will destroy the reputation of the Met Office and the Royal Society. It might even get a lot more serious than that.