By Steve Goddard
Yesterday, the Guardian reported :
Meteorologists have developed remarkably effective techniques for predicting global climate changes caused by greenhouse gases. One paper, by Stott and Myles Allen of Oxford University, predicted in 1999, using temperature data from 1946 to 1996, that by 2010 global temperatures would rise by 0.8C from their second world war level. This is precisely what has happened.
Huh?
The temperature rise since WWII reported by CRU is 0.4C (not 0.8C) and it occurred prior to the date of the study. Climate models use thousands of empirically derived back-fit parameters. Given that fact, the only thing remarkable is that their prediction was so far off the mark. Their forecast is the equivalent of me predicting that Chelsea wins 12-0 yesterday. Off by a factor of two, and after the fact.
I recently attended a meeting of weather modelers, who told me that their models are effective for about 72 hours, not 60 years. GCMs use the same underlying models as weather modelers, plus more parameters which may vary over time.
h/t to reader M White
John Finn says:
August 16, 2010 at 4:06 am
I don’t think NOAA or UAH deliberately fabricate data, but I still haven’t seen an explanation for why the July 2010 UAH anomaly is only ~0.05 deg higher than the July 2009 anomaly while the raw AQUA Ch5 temperatures are ~0.2 deg higher.
I wish I could help you. Better brains than I have been looking the issues with AQUA Ch5.
http://magicjava.blogspot.com/search/label/Aqua%20Satellite
You could try Dr Roy Spencer.
http://www.drroyspencer.com/2010/08/july-2010-uah-global-temperature-update-0-49-deg-c/
I don’t have the raw AQUA CH5 data. As I understand things it needs processing and as I graph lots of data I can’t really spend much time looking at the intricacies of each data set.
stevengoddard says:
August 15, 2010 at 8:15 pm
Ian L. McQueen
What is obvious is that WWII (1939 – 1945) was about zero anomaly and current is about 0.4.
http://www.cru.uea.ac.uk/cru/info/warming/
How do these useless discussions get started?
Usually Goddard jumps the gun and makes a mistake in a post and rather than admit he’s wrong desperately tries to defend his mistake against all comers. Notable previous examples include ‘CO2 snow in the Antarctic’ and his numerous Barrow posts this summer.
stevengoddard says:
August 16, 2010 at 6:02 am
More clearly :
“moving backwards from the present, there is no date prior to 1915 which saw a 0.8C rise.”
Apart from ~1904-14, ~1890-5, and 1855-65 judging by your graphic.
More stupid reporting:
http://ht.ly/18AmDV
If it never crossed his mind that there was a problem, why would he ask for a ruling?
Glenn says:
August 16, 2010 at 1:15 am
Is that based on what Steven reported?
Of course, he is usually quite trustworthy.
Alan the Brit says:
August 16, 2010 at 4:05 am
I’d be more inclined to pay attention to a 0.75°C or 0.85°C approach than an exact one! I’m from the “that’s near enough” school of engineering!
It is standard practice in most science that if you state a number like 0.8, it could mean anything between 0.750000… and 0.849999… If one really had it down to a hundredth of a degree, it would be stated 0.80.
OMGZ world is ending quickly convert to communislam to save yourselves from…from…
The 0.8 figure is correct. Look, CRU data shows 0.44, and when you round it off 4+4 =8. There… Simple! [I got my math degree from U of Phoenix–cost me $30 by mail]
“I recently attended a meeting of weather modelers, who told me that their models are effective for about 72 hours, not 60 years. GCMs use the same underlying models as weather modelers, plus more parameters which may vary over time.”
There is a difference between weather and climate. Being unable to predict the weather 72 hours in advance doesn’t mean you can’t predict the climate 60 years in advance, becuase GCMs do not attempt to predict the weather, instead they perform multiple simulations of the weather and take averages to cancel out the stochastic variability (“weather noise”) and leave the response to changes in forcing (the “forced climate change”).
For an example of a chaotic system, whos exact path is unpredictable, but for which its long term statistical behaviour is predictable, see here
Comments such as the above from the OP merely indicate a lack of understanding of the basics of climate modelling.
Dikran Marsupial
You are missing the point. It is the same models, and errors compound.
Do you think that two week forecasts are more accurate than four day forecasts?
Are five year forecasts more accurate than four day forecasts?
There is a huge gap in logic which climate modelers choose to ignore.
Dikran Marsupial says:
August 16, 2010 at 9:59 am
“[…] instead they perform multiple simulations of the weather and take averages […]”
Oh shove it. Simulating a chaotic system over a century and pretending that a little averaging buys you, what, precision? There’s not even an estimate about the distribution of results, do you assume they’re distributed equally? How will you then estimate the increase in precision your primitive averaging buys you? How many runs do you do? Say, 1000? Simulating a system with billions of degrees of freedom for the initial state?
That’s like buying 10 tickets for the lottery and saying, now we got it covered; we’ll win the jackpot now for sure.
Phil,
You seem to have missed the part :
“moving backwards from the present”
Do you see any dates between the present and 1915 which were 0.8C?
Steve>
You wrote ‘prior to 1915’ when you meant ‘after 1915’, I think. Must have been all those oil dollars making your keyboard slippery 🙂
On the subject of the Liverpool own-goal, I think there’s an interesting parallel between the Arsenal fans convinced Almunia had a worse game than Reina – because they’ve been screaming for Almunia’s head – and the climate faithful. You can see what you want to see, and then you don’t have to admit you were wrong.
Richard Lawson says:
August 15, 2010 at 2:26 pm
“………. Last week we had highly stimulating ‘Is it OK to go wild camping’ – I have yet to understand what the difference is between wild camping and using a campsite.
”
Richard isn’t that the one where you fail to ask permission and end up being chased by a ‘wild’ landowner/farmer/gamekeeper or other such aggrieved party.
Leif Svalgaard says:
August 16, 2010 at 8:19 am
Glenn says:
August 16, 2010 at 1:15 am
I’ll ask you again, how did they arrive at the prediction with 1946-1996 data
“I have no idea, but I assume that Goddard reported that correctly [I could be wrong], and predicting future climate is what climate scientists do, no?”
So you don’t have any idea, but you know “precisely”, quote: “Which it did precisely, so good prediction”. Is that based on what Steven reported?
Is that based on what Steven reported?
“Of course, he is usually quite trustworthy.”
Say Leif, if 2010 turns out to be say +0.762C or +0.786 over the 1946 anomaly, will you claim the prediction of 0.8C came true for 2010?
Apparently the concept of moving backwards in time is too complicated for some.
Leif Svalgaard says:
August 16, 2010 at 8:19 am
“It is standard practice in most science that if you state a number like 0.8, it could mean anything between 0.750000… and 0.849999… If one really had it down to a hundredth of a degree, it would be stated 0.80.”
“Most”? What science does not follow this “standard practice”?
I think Leif is playing with Steve. But, he is correct as far as it goes. All it demonstrates is one can cherry pick and get whatever result you want. Steve did basically the same thing when he chose his WWII values. Someone else might choose the mid-point. It’s all a game.
Now, let’s say this La Niña is a strong one and next year the anomaly drops to .2. Does that now make the prediction wrong? How could it be if it is already right? Specific choices of beginning and ending dates can lead to strikingly different results.
In this case the prediction was nothing more than lucky to have picked a El Niño year as the end date. What if they had picked 2008?
“I recently attended a meeting of weather modelers, who told me that their models are effective for about 72 hours, not 60 years. GCMs use the same underlying models as weather modelers, plus more parameters which may vary over time.”
I am sure you know perfectly well that climate and weather are different phenomena. So, what’s the value in this statement? Who exactly were these modellers, and did you ask them at any point about climate models, or only about weather models?
The fact that climate and weather models often share code is entirely irrelevant. They are tools to investigate different phenomena.
[actually…. as an anonymous commenter, your irrelevance takes precedence ~mod]
“thousands of empirically derived back-fit parameters”
Can you list some small proportion of them? Say, 50 or so?
Richard M
I did not “choose the WWII values.” They are taken from HadCrut.
This conversation has been frustrating, to say the least.
stevengoddard wrote
“You are missing the point. ”
No Steve, you are missing the point that climate modellers don’t make projections of future climate by predicting the weather, so it doesn’t matter if the models can’t predict the weather 72 days in advance or 72 seconds. It just isn’t how climate prediction is performed, it is how weather prediction is performed.
“It is the same models, and errors compound.”
No, the point of using Monte Carlo simulation is that “errors” cancel. Monte Carlo simulations are a very well understood field of statistics. Again, you are demonstrating that you don’t understand the basics of climate prediction – in this case why an ensemble of model runs are used (they are used in weather forecasting as well for much the same reason).
“Do you think that two week forecasts are more accurate than four day forecasts?”
No, but as I have pointed out, climate prediction is not performed by predicting the weather, so it isn’t relevant.
“Are five year forecasts more accurate than four day forecasts?”
Of weather, of course not, but then again climate prediction is not performed by predicting the weather (have I got that point across yet?).
Did you look at the link I gave, which explains the difference between weather and climate, and explains why not being able to predict the weather a long way in advance doesn’t mean you can’t predict the climate. It even explains why ensembles are used.
Dikran Marsupial
Your comments indicate to me a lack of familiarity with climate models.
Monte Carlo techniques are used in climate models for small tasks, like approximating cloud coverage in a grid cell. But the models themselves are necessarily deterministic because of positive feedback effects. For example, if a model mispredicts changes in albedo during year one, the error will get worse in year two.
There is no mechanism for them to correct, because they assume that climate is dominated by positive feedbacks. They only get worse.
stevengoddard says:
August 16, 2010 at 11:36 am
Phil,
You seem to have missed the part :
“moving backwards from the present”
Do you see any dates between the present and 1915 which were 0.8C?
No I thought you were writing in English not Gobbledegoop, you should express yourself more clearly.
Had you actually said ‘there is no date between the present and 1915’ there would have been no confusion, with your clumsy construction I had to assume that you meant something else. Even then 1919 would probably be closer to the mark.
The mistake the Grauniad made was not to say ‘post WW II’ which would have been consistent with the work they were quoting. Quite a minor nit to base a post on.
StevenGoddard writes:
“Monte Carlo techniques are used in climate models for small tasks, like approximating cloud coverage in a grid cell.”
ROTFLMAO! The whole ensemble is one big Monte-Carlo simulation, and I think you will find that cloud cover is generally parameterized in the GCMs, not simulated.
“Your comments indicate to me a lack of familiarity with climate models.”
Oh the irony! ;o)
Glenn says:
August 16, 2010 at 1:26 pm
Say Leif, if 2010 turns out to be say +0.762C or +0.786 over the 1946 anomaly, will you claim the prediction of 0.8C came true for 2010?
Absolutely, wouldn’t you? wouldn’t any reasonable person?
Glenn says:
August 16, 2010 at 1:33 pm
“Most”? What science does not follow this “standard practice”?
Here is one example:
stevengoddard says:
August 15, 2010 at 4:30 pm
“Average HadCrut anomaly during that period was -0.010125.”
The ‘accuracy’ of this number is not commensurate with the number of decimals shown. A more ‘standard’ way of expressing that number might have been -0.010, but perhaps this was not meant to be ‘science’.
Dikran Marsupial
GCMs us the same core code as weather models. They have to. They are modeling the same processes.
The code you imagine would be completely useless. Weather/climate consists of massive amounts of detail. Again, it is quite clear that you are not familiar with how GCMs work.