Parliamentarians do statistical significance

Wow! Who said parliamentarians don't do science? This from Lords Hansard for 8 November.

Lord Donoughue (Labour)

To ask Her Majesty's Government, further to the Written Answer by Baroness Verma on 30 October (WA 114-5) stating that global temperatures have risen less than 1 degree celsius since 1880, on what basis they assert that there has been a long-term upward trend in average global temperatures.

To ask Her Majesty's Government, further to the Written Answer by Baroness Verma on 30 October (WA 114-5) stating that there has been no significant global warming since around 1998, and deeming that period as a shorter timescale, how many years of non-warming they consider would constitute a long-term trend.

To ask Her Majesty's Government, further to the Written Answer by Baroness Verma on 30 October (WA 114-5), whether they consider a rise in global temperature of 0.8 degrees celsius since 1880 to be significant.

• Hansard source (Citation: HL Deb, 8 November 2012, c224W)

Baroness Verma (Whip, House of Lords; Conservative)

The assessment that there has been a long-term upward trend in global average near-surface temperatures since the late 19th century is based upon three global temperature records, compiled from observations, by groups in the US and UK. The rate of global temperature rise on different timescales is summarised in table 1 below. The underlying trend over the period from 1880 to 2011 is 0.062 celsius per decade, giving a total change of 0.81 celsius. Such a rate of change has been judged by major scientific assessments to be large and rapid when compared with temperature changes on millennial timescales.

Over this period some parts of the world have warmed at a much faster rate. The land surface average temperature has risen by about 1.1°C and Arctic temperatures have increased by almost twice the global average rate. The consequences of this warming are already seen across the globe. For example, northern hemisphere sea-ice and snow cover have decreased markedly, most glaciers have retreated and the risks of certain extreme weather events occurring have increased.

Statistical (linear trend) analysis of the HadCRUT4 global near surface temperature dataset compiled by the Met Office and Climatic Research Unit (table 1) shows that the temperature rise since about 1880 is statistically significant.

Time period	Linear trend (°C/decade)	Absolute change in temperature described by linear trend (°C)
1880-2011	0.062±0.009	0.81±0.13
1900-2011	0.074±0.011	0.82±0.13
1950-2011	0.106±0.025	0.66±0.16
1970-2011	0.166 ± 0.038	0.70 ± 0.16

Table 1. Trends fitted to monthly global temperature anomalies for HadCRUT4, with uncertainties describing 95% confidence interval bounds for the combination of measurement, sampling and bias uncertainty and uncertainty in the linear trend fitted to the data. The statistical model used allows for persistence in departures using an autoregressive process (ie that an individual value is not independent of the previous one).

Statistical analyses and modelling of the global temperature record have shown that, because of natural variability in the climate system, a steady warming should not be expected to follow the relatively smooth rise in greenhouse gas concentrations. Over periods of a decade or more, large variations from the average trend are evident in the temperature record and so there is no hard and fast rule as to what minimum period would be appropriate for determining a long-term trend.

Comments

[Spence_UK]

Well, the response would be drafted by civil servants, with assistance (most likely) from CRU or the met office, so it isn't surprising. Looking at the reply:

The statistical model used allows for persistence in departures using an autoregressive process

But natural climate variability is not an autoregressive process (e.g. see Markonis and Koutsoyiannis, Cohn and Lins, etc, etc)

[geronimo]

I'd go for the met office. It has their fingerprint all over the spin.

" Over periods of a decade or more, large variations from the average trend are evident in the temperature record and so there is no hard and fast rule as to what minimum period would be appropriate for determining a long-term trend."

Two issues here, first of all they haven't noted that that applies to warming as well as cooling, and secondly they don't know what natural forcings have halted the warming. It's pure Met Office spin. When the warming halted they changed their slogan from global warming to " hottest decade on record", presumably to avoid Lord Donoghues question. Now they say they expected the warming to halt all along, and if it's for a century there's still no proof that humans aren't causing global warming. So we're nicely boxed into their green agenda no matter what the science says. They really are fanatical greens masquerading as scientists.

[ssat]

Damned by faint raise. Thank you, Lord Donoughue, for asking.

[ConfusedPhoton]

A cherry picked set of results.

Using a time period going back before the nasty carbon dioxide did its stuff might make the number bigger, but it does not support CAGW. Very noticable that the last 10-15 years of not warming is not mentioned. No mention of the Satellite and Argo observations!

No mention that the temperature no longer correlates with carbon dioxide concentration.

[Pharos]

The noble Lord also said

'My Lords, I support the Bill of the noble Lord, Lord Reay, and I do so as a Labour person. Not everyone on this side takes that view, but I do because it is the poorer parts of our community that are paying the main price for this bizarre programme. As has been pointed out, they are paying often without knowing, because the extra taxes that come through the subsidies are often not revealed. I also resent the fact that it involves a huge transfer of wealth from the less well-off to my good friends who own great estates in Scotland and make millions of pounds out of it. I am very happy for them but unhappy for the poorer parts of our community who have to pay for it. I hope that my party will look more closely at this situation in the future than it has in the past.

I also support the noble Lord, Lord Reay, as an environmentalist. It is bizarre that the environmental warriors support this programme when what it does to the visual environment, as has been pointed out, is quite appalling. I object to the fact that they are described as "wind farms". Farms and the farming community contribute enormously to our visual environment but these objects do quite the opposite-they scar it. We need a new collective name and I think "wind blight" is one that could be used in the future because environmentally they are a menace.

As a one-time economist, I particularly object to the economics of the programme, which are absolutely appalling. I shall not go over it all but the wind, especially, is the most uneconomic part; the cost of it is outrageous relative to its contribution. Its contribution is minute. During the winter, the official figures produced showed that wind contributed 0.5 per cent to our energy, partly because of the feature that during very cold spells-certainly in this country-the wind blows less.

http://www.publications.parliament.uk/pa/ld201011/ldhansrd/text/110610-0001.htm#11061043000112

[John Anderson]

I think the peer is Bernard Donoghue - who was chief policy advisor in No 10 to Harold Wilson and then Jim Callaghan. A very bright guy, not blinkered. (I believe he contributed some of the story lines to Yes Minister)

He will not let go of this - his speech sounds devastating

[Agouts]

Rational questions – asked by a Labour peer: remarkable – inevitably, condescendingly dismissed. But given that we know that the DECC is institutionally wedded to every catastrophic prediction of global warming, what could anyone expect? Our masters – I try not to laugh – sold their sneeringly superior souls a long, long time ago.

[Agouts]

Sorry: I missed an obvious point but one worth making nonetheless. You can be 100% certain that Baroness Verma – is that really her name? – understood precisely nothing of the answer she read.

[HaroldW]

HadCRUT4 trends indeed show long-term temperature trends to the current day to be positive. But look at the past and one gets a strong hint of a ~60-year cycle, suggesting that 2000-2030 will be a period of low -- possibly zero -- warming.

This chart doesn't suggest that there is no warming going on. However, the better figure to plan for might well be the 1950-2011 average, given above by the Met Office as about 1 K per century.

[Richard Drake]

ssat:

Damned by faint raise.

I wish I'd said that. In fact I'm sure I will. :)

[MikeHaseler]

"that the temperature rise since about 1880 is statistically significant" - it is only statistically significant if you have a model for the natural variation and when assessed against that model the change is statistically greater than would be expected by that model.

Now this is utter codswallop.

First, it is known that the variation we have seen in the insturmental record of global temperature varies massively with period roughly as t^1.5. So, anyone who knows what they are talking about would know that the variation when considering changes over a century will be much different when comparing changes over a year or a decade.

So, to know whether a change from 1900 to 2000 is statistically significant, we first have to determine the level of natural climate variation over this period.

However, we only have one century which both contains the "signal" & the "noise". So, there is no way to independently determine noise and therefore it is ignorant or worse fraudulent to suggest that the change is statistically significant.

To use a simple example take the height of the sea. Let us use a simple case where we are trying to ascertain how much sea level rise is due to a low pressure zone. What they are trying to suggest is that by measuring the height of the waves approaching the shore, they can then state with confidence that the rise in the level of the sea must be due to the low pressure zone because it is much greater than the wave-to-wave variation.

Obviously this is BS. Because sea level change also responds to natural variations (tides) which are much greater over the period of hours-days than the short term change due to individual waves. So, the fact the sea rise is greater than would be statistically significant than expected for any short term change measured, does not allow us to say it is statistically significant for long term changes. Long term change has to be compared to long term noise.

Instead, what you need to do is create a model of long term change - obviously tide is regular - imagine if they were erratic. We could then create a model of long term sea level changes that naturally occur and only if the sea level rise seen were greater than that change for LONG PERIODs would it be statistically significant.

In summary. They are either being fraudulent, OR they have a model for natural variation which they are not disclosing.

[Salopian]

Bugger statistical significance. When did they actually start using thermometers capable of measuring to within a hundredth of a degree to collect these data? Without this level of accuracy these 'statistics' are totally meaningless. It's a bit like worrying about millimetre rises in average sea levels when then wave height in the Atlantic is in excess of 2 metres for 70% of the time.

[Roy]

The underlying trend over the period from 1880 to 2011 is 0.062 celsius per decade, giving a total change of 0.81 celsius. Such a rate of change has been judged by major scientific assessments to be large and rapid when compared with temperature changes on millennial timescales.

The reference to "millennial timescales" in Baroness Verma's reply implies that the civil servants had compared the rate of warming since 1880 with the rate in previous warming periods over at least 1,000 years. (Actually the plural "millennial timescales" suggests that they made comparisons over more than one millennium). In that case Baroness Verma should publish the rate of warming figures for previous episodes of warming, and also rates of cooling figures for episodes of cooling.

If those figures show that the recent period of global warming is really unprecedented then the government should be very keen on publishing them.

[Pharos]

'Damned by faint raise'

Stealing the thunder from headline writer of the mighty Sun itself.

[Green Sand]

“The World Meteorological Organization (WMO) requires the calculation of averages for consecutive periods of 30 years, with the latest covering the 1961–1990 period. However, many WMO members, including the UK, update their averages at the completion of each decade. Thirty years was chosen as a period long enough to eliminate year-to-year variations.”

http://www.metoffice.gov.uk/climate/uk/averages/

HadCRUT4 30 year trend peaked in December 2003, at 0.199c/decade since then it has been in a downward trend setting subsequent lower highs, lower lows and at September 2012 it now stands at 0.165c/decade a reduction of 17%.

By the Met Office’s own chosen metrics the rate by which this planet is warming is reducing significantly.

This needs to be explained, just where is the evidence to support

“Such a rate of change has been judged by major scientific assessments to be large and rapid when compared with temperature changes on millennial timescales.”

when the rate of change is slowing and has been since 2003?

Without printing new absolute highs each month the 30 year WMO/Met Office "Gold Standard" rate can only continue to reduce.

[Stuart Huggett]

I find the paragraph beginning 'Over this period some parts of the world have warmed at a much faster rate.' to be extraordinary.If the average increase is so low, and some parts are increasing 'much faster' then there must be yet other parts which are getting much cooler in order for the low average to be maintained. Note that this comment applies equally to the 'low arctic ice' meme and the 'extreme weather' one. Am I missing some fundamental bit of statistics here?

OT, but thanks for Hiding the Decline, Bish - a great read!!!

[Alan Reed]

Yes! Yes!
If I stare at the numbers long enough I stop noticing the natural variations. The 1C for doubling - and the 4 times sensitivity just seem to jump out of the screen.

[Adrian]

Question: Temperature has risen by less 1 degrees C in last 130 years is it significant?

Answer: Yes temperature has risen 0,81 degrees C

i.e. A non answer that is too accurate, lol

[Manfred]

“Such a rate of change has been judged by major scientific assessments to be large and rapid when compared with temperature changes on millennial timescales.”

That seems to be a reference to the unspeakable Hockey Stick literature.

High resolution reconstructions show very different results.

"...In addition, the speleothem record suggests significant multi centennial and millennial variability of N. European winters throughout the past 9,000 years that might have been incompletely recorded in tree-ring archives..."

"...The amplitude of the temperature variations in Spannagel, with an estimated range of 2.7 deg C, largely exceeds the temperature reconstructions based on multi proxy record stacks as well as those from low resolution archives [Mann et al., 1998, 1999; Mann and Jones, 2003]..."

"...This question was recently addressed by Moberg et al. [Moberg et al., 2005], who reconstructed a larger variability of N. Hemisphere temperatures for the past 2,000 years after separating high and low resolution archives. They concluded that the low resolution archives show larger multi-centennial variability than most previously published multi-proxy reconstructions because the low frequency curve agrees better with the MWP temperatures reconstructed from borehole measurements in Greenland and with temperatures obtained with a general circulation model than with reconstructions from tree-ring
stacks. This conclusion is supported by von Storch et al. [2004], who suggested that the centennial variability of N. Hemisphere temperature is probably underestimated by regression-based methods and that past variations may have been at least a factor of two larger than indicated by empirical reconstructions..."

"...The most probable answer is that tree-rings rather record the climate conditions during spring and summer whereas both the HSG and COMNISPA curves mirror winter-like conditions, which are only poorly recorded in tree-rings. The comparison thus infers stronger seasonality. There are several indications for a stronger winter than summer variability during the Holocene.."

"...Thus, in summary, the record from Spannagel suggests significant multi centennial and millennial variability of N. European winter climate throughout the past 9,000 years. Our results further imply that multi proxy reconstructions from Europe, which rely heavily on tree-ring records, registering mainly summer signals, might have missed this winter variability..."

http://www.uibk.ac.at/geologie/pdf/mangini07.pdf

[Uncle Badger]

For anyone wondering who this 'Baroness' Verma person is, Dr North has the answer here

http://eureferendum.com/blogview.aspx?blogno=83324

She appears to be another of those who believe it is acceptable to wield power over us without the troubling detail of having won an election.

[The Iceman Cometh]

The thing that none of them seem to notice is that the CO2 only really started growing post about 1950. The problem arises because the CO2 record of Mauna Loa really only starts in 1963, and appears to climb rapidly. But if you show the zero on the ordinate and extrapolate back, you find it was around 300ppm in 1950 and there is a general, understanding that it was at something like 280ppm in 1750, so it was rather flat between 1750 and 1950. Moreover, if you look at fossil fuel use, it only really climbed after 1950, so it is probably the smokin' gun behind the CO2 increase. One can only conclude that all the warming/cooling that took place between 1880 and 1950 must have been natural, and had nothing to do with us naughty humans at all. Take that out of the equation, and the good Baroness' comments seem something of an exaggeration.

[Douglas J. Keenan]

From: D.J. Keenan
To: Lord Donoughue
Cc: Vicky Pope; Phil Jones; Julia Slingo; Baroness Verma; Andrew Montford
Sent: 12 November 2012

Dear Lord Donoughue,

I write with regard to the Written Answer given to you on 8 November 2012 by Baroness Verma, concerning the subject “Climate Change”. The WA includes a table, with columns for “Linear trend” and “Absolute change in temperature described by linear trend”. There is a serious problem with the numbers in those columns: they are all unfounded.

The Baroness is faithfully reporting numbers that were given to her by the Met Office and the Climatic Research Unit. I have previously contacted both the Chief Scientist at the Met Office, Julia Slingo, and the Head of Research at CRU, Phil Jones. I have asked both for valid justifications for similar numbers. No such justifications have been given. In addition, I have had an e-mail exchange about this with a Senior Scientist at the Met Office, Vicky Pope: a copy of the exchange is attached; again, no justification was given.

Both the Met Office and CRU have repeatedly claimed that global temperatures are significantly increasing, but neither can justify their claim. I have been attempting to press this point for years. For example, in 2011, I published a full-page op-ed piece about this in The Wall Street Journal:
http://www.informath.org/media/a42.htm

The basics of the problem are simple. In statistical analyses, a model is fit to the data, and inferences are drawn from the model (as indicated by the WA); in order for a statistical analysis to be considered valid, the model must be demonstrated to be plausible on both statistical and scientific grounds; yet no one has attempted to demonstrate that the model used for global warming is plausible on either statistical or scientific grounds. In other words, the model was simply adopted by proclamation. The failure to present any evidence or logic to support the selection of the model is a serious violation of basic scientific principles—indeed, it means that what they have done is not science.

Perhaps you could ask another Question, about why they selected the model that they did. I strongly suspect, however, that the Met Office and CRU would respond with circumlocutions.

To my knowledge, the sole attempt to provide a valid statistical analysis is that given by

Koutsoyiannis D. (2011),
“Hurst–Kolmogorov dynamics as a result of extremal entropy production”,
Physica A, 390: 1424–1432.
doi:10.1016/j.physa.2010.12.035

One conclusion of this paper is that global temperatures are not significantly increasing. In other words, according to this paper, there is no global warming. I have found, however, an invalidating flaw in the analysis leading to that conclusion. Hence, I believe that at present no one knows whether global temperatures are significantly increasing.

The underlying problem is due to ignorance about a statistical concept known as a “time series”. A time series is any series of measurements taken at regular time intervals. Examples include the following: prices on the New York Stock Exchange at the close of each business day; the maximum temperature in London each day; the total wheat harvest in Canada each year. Another example is the average global temperature each year. More generally, virtually every climatic data set is a time series.

Evaluating whether an increase in a time series is significant requires specialized statistical techniques. Those techniques are wholly unknown to almost all climatologists. As an illustration, in 2010, I was in a debate with Trevor Davies: Davies is a Pro Vice Chancellor at Phil Jones' university and a former head of CRU. I offered to pay £500/minute to have Jones write an examination in an introductory (undergraduate) course in time series. My offer was not accepted. Indeed, it is obvious to anyone familiar with time series that a large majority of the work published by Jones during his decades-long career is grossly incompetent.

One idea that you might consider is to ask the following three-part Question.

1. How many scientists are employed by the Met Office who spend at least half their time doing work that is closely related to global warming?
2. How many of those scientists have taken an introductory (undergraduate) course in statistical analysis of time series?
3. How many of those scientists have taken a graduate-level course in statistical analysis of time series?

I suspect that the answer to the third question is “zero”. From which you could correctly conclude that most of their statistical analyses of the climate are unfounded. There is a scandal here.

To briefly introduce myself, I used to do mathematical research and financial trading, on Wall Street and in the City of London; since 1995, I have been studying independently. My web site tells more. My background in mathematical finance is relevant, as many financial data sets are also time series. Indeed, the best people in time series tend to be not in academia, but in finance, because remuneration in finance is much higher.

Sincerely,

Douglas J. Keenan
http://www.informath.org

[Jonathan Bagley]

Douglas, here is a link to an article by Richard Davis, a very respected figure in Time Series, discussing the paper by McShane and Wyner, which is the only paper on climate change I can remember being published in a top statistical journal.

[Jonathan Bagley]

Sorry, forgot the link
http://arxiv.org/pdf/1104.4176.pdf