A response on statistical models and global temperature

31 05 2013

Over a period of several months the Met Office has been involved in dialogue and answered a series of questions on the subject of the use of statistical models in relation to the global temperature record.

The Met Office’s Chief Scientist, Julia Slingo, has written a discussion paper on the subject – you can now view the Executive Summary and a link to the full paper in an article on our Research News pages.

Publication of this paper follows a guest article recently published on the Bishop Hill blog site, where one of the people with which the Met Office has been speaking with – Doug Keenan – makes a series of accusations about the Met Office and its science.

Professor Slingo’s paper answers many of the points Mr Keenan makes, and the Met Office has already directly addressed many of the points Mr Keenan raises through considerable previous correspondence we have had with him on this issue. However, here we directly address a few of the key points in Mr Keenan’s article:

1)       Mr Keenan says that there is “no basis” for the claim that the increase in global temperatures since the late 1800s is too large to be reasonably attributed to natural random variation. He goes on to argue that this is because we haven’t used the right statistical model.

However, the claim that the increase in global warming is larger than could be explained by natural variability has a clear and well understood grounding in fundamental physics and chemistry. There is very high confidence (using the IPCC’s definition) that the global average net effect of human activities since 1850 has been one of warming. The basis for this claim is not, and never has been, the sole use of statistical models to emulate a global temperature trend. Instead it is based on hundreds of years of scientific advancement, supported by the development of high-quality observations and computational modeling.

2)       Mr Keenan suggests that Met Office scientists have been ‘trying to cover it [point 1, above] up’.

The Met Office has entered into email discussion at the working scientific level and responded promptly and transparently on all parliamentary matters and questions. We have also responded to numerous emails from Mr Keenan and invited him to come to the Met Office to discuss statistical modeling in climate science. As he points out in his article, so far those invitations have been declined or unanswered. The invitation still stands.

3)       Mr Keenan then goes on to argue that you can only use a statistical model to determine whether the warming we have seen is statistically significant. He argues that the Met Office has used the wrong statistical model and, therefore, our science is flawed.

The study of climate variability and change is broader than the domain of statistics, most notably due to the importance of the underpinning science of the climate system. Our judgment that changes in temperature since 1850 are driven by human activity is based on information not just from the global temperature trend, or statistics, but also our knowledge of the way that the climate system works, how it responds to global fossil fuel emissions and observations of a wide range of other indicators, such as sea ice, glacier mass, sea level rise, etc.

Using statistical tests in the absence of this other information is inappropriate, particularly when it is not possible to know, definitively, which is the most appropriate statistical model to use. In particular, a key test of an appropriate statistical model is that it agrees with everything we know about the system. Neither of the models discussed by Mr Keenan is adequate in this regard. On that basis, this conversation on statistical modelling is of little scientific merit.

4)       Mr Keenan details his argument to say that various different statistical models can emulate the global temperature record better and worse than others.

This is something the Met Office has already spoken about and shown analysis on (such as in an answer to a parliamentary question (PQHL62)). However, this assessment of relative likelihood does not ensure that any of the statistical models are scientifically valid. Because the Met Office does not make an assessment of global warming solely on statistics – let alone the statistical models referred to in Mr Keenan’s article, this exercise is of very little, if any, scientific use.

5)       Mr Keenan also makes repeated accusations that the Met Office did not, or was not willing to respond to Parliamentary Questions.

This is not the case. The Met Office answered every request for input to Parliamentary Questions and answered them in the most scientifically appropriate way to the best of its knowledge. There has never been a refusal to provide information to answer a Parliamentary Question.


Actions

Information

17 responses

31 05 2013
newminster

Is it “scientifically appropriate” not to say whether or not you used a time-series analysis when that was the question that was asked? Eventually — at the sixth attempt — the questioner got an answer that made sense in the context of his question.
Nobody is disputing that there has been an increase in temperature; the question was whether that increase was significant and what method you had used to determine that it was. That was the question and five times those responsible for providing ministers with an answer left them flapping in the wind.
Keenan is only one of many in the blogosphere who know what they’re talking about.The days when any old rubbish is good enough for the plebs are long gone, I’m afraid.

31 05 2013
ntropyalwayswins

Extract from Douglas Keenan’s post at BH

“Lord Donoughue then tabled a Parliamentary Question asking HM Government for their assessment of the likelihood of the trending autoregressive model relative to the driftless model. HM Government did not answer. Lord Donoughue asked a second time. They did not answer. He asked a third time. Again they did not answer. He then asked a fourth time.”

Lord Donoughue then tabled the question 2 more times before it was answered.

Can you explain please why the good Lord had to ask the same question 6 times before he got a response. Thank you.

31 05 2013
Tim Hammond

Huh? You answered on the fifth time – hardly a glowing record. And if the statistics show that it is natural variability, then what are you arguing about? It is ludicrous to claim that you need “other science” to explain something that can be explained by natural variation.

31 05 2013
CJLB (@orach24463_cj)

Global temperatures have been flat since 1998 whereas CO2 has gone up. Ergo CO2 AGW theory has been disproven by emperical evidence. Thus there is no need to deny poor people access to cheap and abundant energy from fossil fuel or allow people to freeze to death due to energy poverty.

31 05 2013
ntropyalwayswins

And Lord Donoughue has just posted this comment at BH

“The Met Office statement, where it relates to my parliamentary Questions is, at best, misleading. It claims that they did not fail to answer the string of Questions and were always transparent. In fact there was just one basic Question, requiring a simple numerical reply. It had to be resubmitted repeatedly , with small verbal refinements, over several months because the MO repeatedly refused to answer it. Instead they gave responses of impressive verbiage, but no straight answer. Had they answered the Question, as the rules require, the Question would not have been resubmitted. It was finally answered after I had raised the issue with the helpful minister, with the Leader of the Lords and with senior clerks of the House.
Others may speculate on why they evaded this Question, but it suggests MO statements should be viewed with caution.
Bernard Donoughue

May 31, 2013 at 6:46 PM | Lord Donoughue of Ashton”

31 05 2013
karabar

In light of the recent weather in the northern hemisphere, the above five point rebuttal is clearly the same product that emerges from the rear end of a bull.

31 05 2013
Linda Serena

Willis Eschenbach recently pubished a post destillating Met Office and IPCC climate models knowledge about global average temperature:

http://wattsupwiththat.com/2013/05/21/model-climate-sensitivity-calculated-directly-from-model-results/

The essence is:

Climate models can be regarded as black boxes, Global temperature output is computed by an “exponential moving average”applied on input forcings with only 2 parameters in a trivial formula, R value 0.99 (see picture above)

https://en.wikipedia.org/wiki/Moving_average

Or in other words, you can compute climate models global temperature without a supercomputer on the back of an envelope.

I would think it is not very likely that climate responds to forcings in such a primitive way. Worse, the input defines the output, about 100% human caused forcings will miss any natural effects or cycles or if they contribute.

31 05 2013
John Havery Samuel

Bishop Hill is wasting hardworking taxpayers’ time and money with this nonsense. Surely the Taxpayers’ Alliance, the TPA, should have a stern word with Mr Keenan. (Of course I’m joking! Anti-science campaigns on many fronts.)

31 05 2013
nuwurld

Dear Met, statistics is mathematics pure and simple. A model either models a ‘statistically similar’ outcome to measured reality or it doesn’t. If your reasoning is one of ‘picking’ from competing models that ‘disagree’ then you have to question the validity of the models.

If you are not going to accept variance and conjecture through statistically approved views then…

What do you say you have?

“but also our knowledge of the way that the climate system works, how it responds to global fossil fuel emissions and observations of a wide range of other indicators, such as sea ice, glacier mass, sea level rise, etc.”

Knowledge of how the climate system works! Like Venus at 1bar pressure? ( exactly like Earth with 1.91x the insolation but 95.6% CO2)

Sea ice, a positive anomaly again. How does that work after what, 130 years of forcing.

Glacier mass and sea level rise. Now come on that’s one and the same. The Maldives are still there as they where 50 years ago. 0.5 m above the oceanic mean. Like so many ‘fragile’ islands not swamped by rising seas.

GRASP, that’s the name of the next generation of satellite info ‘not’ to be dependent upon land based reference and it’s errors. Until then…

Surely, when the ‘science is done’ no one can question that?

Look Met, this could be like the ‘prodigal son’. You can’t all be sacked from returning to reality. The IPCC’s views are going down with global temperatures. Be the first to reject the fiction, en masse.

1 06 2013
vsaluki (@vsaluki)

“Mr Keenan then goes on to argue that you can only use a statistical model to determine whether the warming we have seen is statistically significant.”

The fact that Slingo doesn’t understand that this absolutely must be true is proof that she is unfit to hold her office.

All of the other things that she mentions may be relevant as to why the statistical results turn out to be what they turn out to be. But they are not cumlative reasons along with the statistical results, they are only reasons for those statistical results. If they are good enough and complete enough as reasons then that will be ultimately shown by the statistics.

1 06 2013
Peter Maddock

If you say something is statistically significant be prepared to explain on what basis the claim is made. Six parliamentary questions later is an embarrassment to the Met office and makes your last paragraph a lie.

1 06 2013
roncram

Unfortunately, this blog post is full of circular reasoning and unsupported claims.

In #1, you write: “The basis for this claim is not, and never has been, the sole use of statistical models to emulate a global temperature trend. Instead it is based on hundreds of years of scientific advancement, supported by the development of high-quality observations and computational modeling.”

You appear to say the statistical models do not support the claim the temperature trend is too large to be natural but the claim is supported by “high-quality observations and computational modeling.” The issue here is that high-quality observations are at odds with computational modeling. There has been a consistent and significant rise in atmospheric CO2 since 1998 but no clear corresponding rise in temperature.

In #3 you write: “Our judgment that changes in temperature since 1850 are driven by human activity is based on information not just from the global temperature trend, or statistics, but also our knowledge of the way that the climate system works, how it responds to global fossil fuel emissions and observations of a wide range of other indicators, such as sea ice, glacier mass, sea level rise, etc.”

This is circular reasoning. You are saying “We know the physics of atmospheric CO2 are causing global warming because we know the physics of atmospheric CO2 cause warming.” I’m sorry, but that is circular reasoning. It is not science. You have a hypothesis, and it is a reasonable hypothesis – but you cannot allow the hypothesis to dictate what the data is telling you. Statistical analysis of the data is essential to determine is the hypothesis is true or if something else may be in play such as a natural negative feedback to the climate system.

In #4 you write: “Because the Met Office does not make an assessment of global warming solely on statistics – let alone the statistical models referred to in Mr Keenan’s article, this exercise is of very little, if any, scientific use.”

This is clearly wrong. Mathematics is the language of science. If you cannot make your argument based on statistics, then you cannot make your argument.

1 06 2013
Bobby Graham

The paper by Julia Slingo misses the point of the Parliamentary Questions. The point is not about the state of the IPCC science. It is to determine if there is “independent” statistical verification that global warming we have seen since the industrial revolution is natural or caused by CO2. Basically, they are trying to determine if there is an independent statistic that would help Parliament decide if they should or should not base decisions on the Met Offices assessments.

The paper is also scientifically inadequate. Example: “Physical understanding of the climate tells us that in the absence of external forcings, global temperatures should vary around a climatological average temperature. Temperatures may exhibit long-term persistence in these excursions away from the climatological average as a result of internal variability in the system.”

The current temperature is well within the range of temperatures that have occurred in the Holocene. Given the difficulties in determining the actual temperature thousands of years ago to good precision, it is difficult to determine what the average temperature was over the last 8,000 years. Some papers indicate the current temperature is below the average of the last 8,000 years of the Holocene and others show we are above the average. It is fairly certain from the literature that the current temperature is within the range of temperatures that have been common during several periods within the Holocene. The implication that Julia Slingo is trying to make is that the current temperature is unusual as it is not varying around a “climatological average” and therefore, the basis for the Parliamentary Questions is erroneous. But this is incorrect. Julia Slingo is basically looking at a few hundred years to determine the climatological average rather than the entire Holocene. She gives no scientific basis as to why this is correct.

Julia Slingo goes on to say that, “the Met Office does not use one of these statistical models to assess global temperature change in relation to natural variability.” In other words, she essentially states that the question is moot because she believes that the IPCC models are correct and that the current temperature is not varying around a “climatological average”. Based on this, they go on to not answer the question posed by Parliamentary Questions.

6 06 2013
jdey123

The claim by the MetOffice and other environmental extremists is that basic physics says that CO2 in the atmosphere will warm by the planet. Basic physics doesn’t hide for the 16+ years, that the “missing heat” has failed to show up in the global surface temperature record. Antarctic sea ice extent continues to increase. Arctic sea ice extent is currently in the normal range. May 2013 was only the 10th smallest arctic sea ice extent in the record which is only 34 years in totality. May 1989 had less sea ice in the arctic than May 2013.

The fact is that no climate model has accurately predicted observed temperatures, so climate model runs should be ignored. That leaves only the evidence around sea ice extent, sea levels etc.

CO2 levels have continued to steadily increase and there are no natural factors which are cooling the planet. There is no La Nina. There has been no major volcanic eruption and solar cycle 24 is at it’s maximum. The global temperature should be at record levels now, according to the AGW hypothesis. To insist that the hypothesis is not bust and that we have to wait for 30 years is not science, it’s scientific fraud.

15 06 2013
nebakhet

The cause of the warming has never been based on statistics. I have no idea what these climate deniers are trying to do other than create a false fuss.

The cause of the warming has always been determined by attribution studies that include actual physics. Which they must.

18 06 2013
karabar

Exactly what ‘warming’ are you imagining? With record lows being tested all over the globe, temperature declines since 2002, nothing significant in 17 years 4 months, do these ‘attribution studies’ and ‘actual physics’ of yours refer to some other planet?

11 07 2013
OgreMkV

Citation needed.

Weather is not climate. Even a year’s worth of weather is not the same as climate. Just because it snowed in 2010 in Austin, TX does not mean that the global average temperature is dropping.

Here: http://www.skepticink.com/smilodonsretreat/2013/07/10/global-warming-by-the-decade/

Actual temperatures, not anomalies based on 3 year moving averages. It’s quite simple.

Oh, and to the person who said that fossil fuels are cheap… in case you weren’t aware, unsubsidized wind power is cheaper than all unsubsidized fossil fuels except for natural gas.

And here’s a discussion that you might find interesting: http://www.skepticink.com/smilodonsretreat/2012/10/30/on-fossil-fuels/

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s




Follow

Get every new post delivered to your Inbox.

Join 6,158 other followers

%d bloggers like this: