From the “(pick one: 90% 95% 97%) certainty department, comes this oopsie:
Via Bishop Hill:
=============================================================
Doug Keenan has just written to Julia Slingo about a problem with the Fifth Assessment Report (see here for context).
Dear Julia,
The IPCC’s AR5 WGI Summary for Policymakers includes the following statement.
The globally averaged combined land and ocean surface temperature data as calculated by a linear trend, show a warming of 0.85 [0.65 to 1.06] °C, over the period 1880–2012….
(The numbers in brackets indicate 90%-confidence intervals.) The statement is near the beginning of the first section after the Introduction; as such, it is especially prominent.
The confidence intervals are derived from a statistical model that comprises a straight line with AR(1) noise. As per your paper “Statistical models and the global temperature record” (May 2013), that statistical model is insupportable, and the confidence intervals should be much wider—perhaps even wide enough to include 0°C.
It would seem to be an important part of the duty of the Chief Scientist of the Met Office to publicly inform UK policymakers that the statement is untenable and the truth is less alarming. I ask if you will be fulfilling that duty, and if not, why not.
Sincerely, Doug
============================================================
To me, this is just more indication that the 95% number claimed by IPCC wasn’t derived mathematically, but was a consensus of opinion like was done last time.
Your article asks “Were those numbers calculated, or just pulled out of some orifice?” They were not calculated, at least if the same procedure from the fourth assessment report was used. In that prior climate assessment, buried in a footnote in the Summary for Policymakers, the IPCC admitted that the reported 90% confidence interval was simply based on “expert judgment” i.e. conjecture. This, of course begs the question as to how any human being can have “expertise” in attributing temperature trends to human causes when there is no scientific instrument or procedure capable of verifying the expert attributions.
So it was either that, or it is a product of sleep deprivation, as the IPCC vice chair illustrated today:
There’s nothing like sleep deprived group think under deadline pressure to instill confidence, right?
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
The liars picked 95% because it was higher than the last AR, IMO. As some wit commented earlier, it will probably be Cook’s legendary 97% in the next AR, if there be one.
Same same this time around: 95% probability does not equal confidence interval. Unscientific jiggery-pokery!
“Probabilistic estimates of quantified measures of uncertainty in a finding are based on statistical analysis of observations or model results, or both, and expert judgment2.”
2 In this Summary for Policymakers, the following terms have been used to indicate the assessed likelihood of an outcome or a result: virtually certain 99–100% probability, very likely 90–100%, likely 66–100%, about as likely as not 33–66%, unlikely 0–33%, very unlikely 0–10%, exceptionally unlikely 0–1%. Additional terms (extremely likely: 95–100%, more likely than not >50–100%, and extremely unlikely 0–5%) may also be used when appropriate. Assessed likelihood is typeset in italics, e.g., very likely (see Chapter 1 and Box TS.1 for more details).
From what I have seen of the raw data, the ‘confidence limits’ are untenable – the 95% CL on the ‘regression’, for example, should bound roughly 95% of the data. Anyone who has read chapter 2 of a statistics book discussing linear regression would know that the CL on a regression line comprises a pair of hyperbolas (one above the regression line and one below) with their vertices pointing at the overall data mean (the midpoint of the regression line). These hyperbolas are asymptotic to the lines passing through the data mean with different slopes, as illustrated here: http://www.statsoft.com/textbook/multiple-regression/
representing the regression slope plus or minus the CLs on the slope itself.
I suspect what they are reporting as a CL is on the variance of the slopes among the *models*, which is totally meaningless, as has already been pointed out elsewhere.
When the IPCC pulls numbers out of their collective behinds this is what happens. I am praying their credibility will be shot sooner rather than later.
Billion more dollars needed to fix report.
Well we cannot expect the Summary for Policy Makers to reflect the science can we? After all, the science hasn’t even been published yet! Worse, policy makers had a heavy hand in the wording of the summary of the science which hasn’t yet been published!
A better name might by “Summary of the Policy Makers, By the Policy Makers, For the Policy Makers”.
From the University of Colorado Boulder, Headline: “Shrinking atmosphere linked to lower solar radiation”. In summary, the upper atmosphere has shrunk 30% and cooled by 74 degrees since 1998.
“It is now clear that the record low temperature and density were primarily caused by unusually low levels of solar radiation at the extreme-ultraviolet level,” Solomon said.
C02 had an impact of less than 5%.
Hmmmmmmmmm……
Forgot to post the URL for the shrinking atmosphere article…
http://artsandsciences.colorado.edu/magazine/2010/08/shrinking-atmosphere-linked-to-low-solar-radiation/
When I was going to school, a mark between 85% and 100% was considered an A grade. I do not know what I would have done if it was further divided up into “likely,” “very likely,” and “most likely.” If it had, I would have been so confused.
The Belgian news is still on the CAGW line. In the Netherlands there was at last more nuance. Marcel Crok pointed out in a very gentle way that the models don’t fit and the warming could be far less than assumed, whereupon an AGW supporter made the remark that the warming disappeared in the oceans. They don’t realise that models failing to predict the actual conditions, predicting for the year 2100 is scientific blasphemia.
Manny M says:
September 27, 2013 at 11:37 am
From the University of Colorado Boulder, Headline: “Shrinking atmosphere linked to lower solar radiation”. In summary, the upper atmosphere has shrunk 30% and cooled by 74 degrees since 1998.
“It is now clear that the record low temperature and density were primarily caused by unusually low levels of solar radiation at the extreme-ultraviolet level,” Solomon said.
C02 had an impact of less than 5%.
Hmmmmmmmmm……
Manny M says:
September 27, 2013 at 11:38 am
Forgot to post the URL for the shrinking atmosphere article…
http://artsandsciences.colorado.edu/magazine/2010/08/shrinking-atmosphere-linked-to-low-solar-radiation/
Stephen Wilde has been saying this for some time.
My guess is Julia Slingo will not respond or acknowledge any problem, that is her role!
The following is a comment from the InterAcademy Council review of the IPCC process and procedures in 2010:
“The IPCC uncertainty guidance urges authors to provide a traceable account of how authors determined what ratings to use to describe the level of scientific understanding (Table 3.1) and the likelihood that a particular outcome will occur (Table 3.3). However, it is unclear whose judgments are reflected in the ratings that appear in the Fourth Assessment Report or how the judgments were determined. How exactly a consensus was reached regarding subjective probability distributions needs to be documented.”
I couldn’t find any such documentation in the SPM. Perhaps it’s in the AR5 WG1 report coming out soon? Or perhaps it doesn’t exist. Hmmm….
Over a year ago we had the SREX IPCC report that said.
Recently we had the draft Summary for Policymakers.
Did the instruments develop over the last week? Wasn’t it scary enough? Bring in the government representatives and what do you get, a consensus among civil servants. The scientists are left scratching their heads, but he who pays the piper……..
since the mid-20th century, and likely that human influence has more than doubled the probability of occurrence of heat waves in some locations (see Table SPM.1). {10.6}
http://www.climatechange2013.org/images/uploads/WGIAR5-SPM_Approved27Sep2013.pdf
Sorry,
Messed up the html. The last paragraph should also be indented and is part of the quote.
Since a linear trend model plus error has least squares estimators that are asymptotically nomal, one standard error of the trend margin is about .2 /1.645 degrees C, or on order of magnitude 0.15. So 0.85, the quoted trend estimate, has a Z-score of about 5+, which is way more significant than 95, 99%, 99.99%, etc…. Seems to me the authors are downplaying the evidence. While I have not read the report, I wonder what is Doug arguing? That the AR(1) model is inappropriate? Sure. But statistics has advanced way beyond this the last few decades. Elaborating, put in a fractionally differenced long-memory component in the model if you want. You’re not going to knock that z-score below significant. Please reinforce your argument.
Hi. Thanks for the post. RE:
Worth noting: the Solomon of this 2010 (?) paper is “Stanley”, not Susan, who used to work at NOAA, and won the “Nobble Prize” along with other members of the Algore fan club.
Weather forecasts use the same approach. If three weathermen say it will rain and one disagrees, there’s a 75% chance of showers.
“While I have not read the report, I wonder what is Doug arguing? That the AR(1) model is inappropriate?”
There’s a bit of history behind this. The IPCC in their last report used 90% confidence intervals based on REML linear regression to state bounds on the amount of warming. Doug had a very long argument with the UK Met Office when they used this model to claim the warming was “significant”, via a number of Questions in Parliament, that eventually ended with the Met Office conceding that AR(1) was unphysical and far less likely than some other noise models they could have used. The chief scientist there, Julia Slingo, said that AR(1) was unrealistic and tried to claim they hadn’t used it in making their assessment, instead like the IPCC using a wide range of evidence.
So evidently, when the new report came out Doug immediately checked what model they were using. Turns out they’re still using AR(1), the model Julia Slingo said was rubbish. The confidence intervals are wrong, because the error model used to generate them is wrong – on the authority of the Met Office chief scientist.
The question is, will she say so?
“Elaborating, put in a fractionally differenced long-memory component in the model if you want. You’re not going to knock that z-score below significant. Please reinforce your argument.”
Actually, yes you can. That was the basis of Doug’s earlier argument – that a trendless ARIMA(3,1,0) model fits the data a thousand times better than AR(1). There are links to the context at Bishop Hill.
—
In case anyone else wants to check, the following R script ought to replicate (roughly) the IPCC’s calculation. I used GISTEMP here, although the IPCC didn’t say I’m assuming they used the combination of several global temperature series, or possibly different versions. But the closeness of the result indicates that this is indeed what they’ve done.
# ###################
# Replicate IPCC’s confidence interval for warming 1880-2012
library(nlme) # nlme contains gls
# Read in GISTEMP data obtained from
# http://data.giss.nasa.gov/gistemp/tabledata_v3/GLB.Ts+dSST.txt
# Downloaded 27 Sep 2013
gistemp<-ts(c(-22,-13,-16,-19,-27,-25,-24,-31,-19,
-10,-33,-27,-31,-36,-32,-25,-18,-18,-31,-20,-14,-21,
-30,-36,-44,-29,-26,-42,-43,-46,-45,-44,-41,-39,-23,
-16,-36,-44,-31,-29,-27,-21,-29,-26,-24,-22,-9,
-18,-16,-31,-11,-7,-10,-25,-9,-15,-10,3,6,1,6,8,5,6,
14,1,-8,-4,-10,-11,-19,-6,2,9,-11,-12,-18,4,4,3,
-4,5,4,7,-20,-10,-4,-1,-5,6,4,-7,2,16,-7,-1,-12,15,
6,12,23,28,9,27,12,8,15,29,35,24,39,38,19,21,28,
43,33,45,61,40,40,53,61,60,51,65,59,63,49,59,66,
55,58)/100,start=1880)
# Do the regression using AR(1) model, restricted maximum
# likelihood, and show the coefficients of the best fit
glsREML<-gls(gistemp ~ time(gistemp), cor=corARMA(p=1,q=0), method="REML"); coefficients(glsREML)
# Calculate 90% confidence interval on the slope
confint(glsREML,level=0.9)
# Calculate 90% confidence interval on the increase from 1880 to the end of 2012
confint(glsREML,level=0.9)[c(2,4)]*(2013-1880)
# Plot the data and slope on a chart
plot(gistemp)
abline(glsREML)
IPCC news flash ,the global temperature trend going forward is going to be DOWN, not up.
Natural causes can explain the temperature rise from 1880-1998 from high solar activity, to a warm PDO post 1980, up through 1998, featuring more El Nino activity.
Davidmhoffer:
Dont you mean “Summary of the Money Makers, By the Money Makers, For the Money Makers”?
Nullius in Verba says: September 27, 2013 at 1:38 pm
“Turns out they’re still using AR(1), the model Julia Slingo said was rubbish.”
Would you care to quote Julia Slingo saying AR(1) was rubbish?
This seems to be just another episode in Doug Keenan berating people for not using his pet AR(3,1,0) model, which gets a better fit at the expense of extra parameters and physical impossibility. But the IPCC is not claiming that linear+AR(1) is the best model of temperature. They are simply using it as the basis for calculating temperature change over the period.
According to the MO, AR(3,1,0) would give a temperature change of 0.73°C, which is well within the IPCC stated range. So I can’t see how this can be construed as an error.
Simon, except they are not “making money”… they are spending our money.
How about “Summary of the Tax Spenders, By the Tax Spenders, For the Tax Spenders”.
There, much better.
Summary for ScareMongers — 100 Ways to Spread Scare Stories and Make a Million!
Manny M says:
September 27, 2013 at 11:37 am
Would like to hear Dr. Leif Svalgaard’s take on this.