Buy

Books
Click images for more details

Support

 

Twitter
Recent comments
Recent posts
Currently discussing
Links

A few sites I've stumbled across recently....

Powered by Squarespace
« #Greensgobyair | Main | Vision of awfulness »
Friday
May312013

Met Office responds to Keenan

Doug Keenan's article about statistical significance in the temperature records seems to have had a response from the Met Office.

The text is here and there is a blog post here.

PrintView Printer Friendly Version

Reader Comments (102)

Comment removed following update

May 31, 2013 at 4:57 PM | Registered CommenterGreen Sand

The Met Office should be grateful for Global Warming If it exists or not.Because without it they would have nothing else to research and talk and make pronouncements about and they would be Privatized.The Met Office is no different to any other Weather Forecasting Service in the world.

May 31, 2013 at 5:04 PM | Unregistered Commenterjamspid

To be fair to the MO, the parliamentary questions to which this is a response didn't mention Doug either

May 31, 2013 at 5:13 PM | Unregistered CommenterArthur Dent

The Met Office appears to be claiming that the statistical analysis of actual temperature data should carry little weight in the climate debate. We should have far more regard to climate modelling which Prof Slingo tells us is clearly superior in reflecting the "science" and consensus of AGW. I guess she is trying very hard to say this with a straight face. But at least we have it in black and white from our top climate scientist - climate science is not about observation and testing theories with real data, it is all about tweaking their powerful and oh so accurate models to post-reflect whatever happens. Even economists must be laughing.

May 31, 2013 at 5:25 PM | Unregistered CommenterDerrick Byford

A wide range of observed climate indicators continue to show changes that are consistent with a globally warming world, and our understanding of how the climate system responds to rising greenhouse gas levels.
There's that wonderful phrase "consistent with" again.
And will someone explain the meaning of "a ...range of ...indicators continue to show changes ... and our understanding of how the climate system responds...."
It's not even a non sequitur; it's totally meaningless as a piece of English prose!

May 31, 2013 at 5:26 PM | Registered CommenterMike Jackson

Which now begs the question:- Why did the MO ignore the original Question so many times, before deigning to provide an answer?

May 31, 2013 at 5:27 PM | Unregistered CommenterJoe Public

Gnothos sans statistos. Hmmm. What are the odds?
===============

May 31, 2013 at 5:28 PM | Unregistered Commenterkim

Figure 1 shows no increase in radiative forcing after 1998 - which is consistent with no global warming since 1998. This is good news - the 35 ppmv rise in atmospheric CO2 during this period did not have any effect at all.

We do not need the windmills any more.

May 31, 2013 at 5:30 PM | Unregistered CommenterRoger Longstaff

@ Mike Jackson 5:26 pm

It means: "observations of barometric pressure, sea surface temperature, and sea-ice concentration using a physically-based data assimilation system " are rising.

Do you not follow the latest pronoun cements by PhilJones et al & his crew at UEA-CRU???

May 31, 2013 at 5:34 PM | Unregistered CommenterJoe Public

May 31, 2013 at 5:27 PM | Joe Public

Which now begs the question:- Why did the MO ignore the original Question so many times, before deigning to provide an answer?

Maybe the Met Office provided an answer but it was not passed on.

May 31, 2013 at 5:50 PM | Unregistered CommenterBilly Liar

If the "changes" are consistent with natural variation, why do you need to look at anything else?

There isn't a phenomenon that needs explaining, so what are they explaining with the rest of the science?

I am baffled by this argument.

May 31, 2013 at 6:00 PM | Unregistered CommenterTim Hammond

From the blog post:

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.

We can all start going around in circles again?

If the 'global average net effect of human activities' (something Keenan and Donoughue do not primarily refer to, but are surely interested in) is one of warming, then, surely, 'warming' has occurred.

How do you know that?

Is the above such a difficult question to understand?

May 31, 2013 at 6:00 PM | Registered Commentershub

Billy Liar

"Maybe the Met Office provided an answer but it was not passed on."

Hardly likely (@ 99.99% confidence factor) considering Slingo's pronouncements on the MO website have today's date.

May 31, 2013 at 6:00 PM | Unregistered CommenterJoe Public

The Slingo pdf mentions 10 indicators other than surface temperature (tropospheric temp, specific humidity, ice cover, etc.) that were summarized in a 2010 Met Office paper. The data for these indicators is provided in the following link:

http://www.metoffice.gov.uk/hadobs/indicators/11keyindicators.html

This seems rather useful, except that each dataset is offered for individual download. Since there are about 50 or so datasets, this could be rather painfully slow. Perhaps I have missed some way to get them all in one download. Of course, they all end at or before 2009, so are somewhat outdated now. For example, their glacier mass data has been corrected downwards recently. So if the data were used, they should all be updated. Nonetheless, it's pretty convenient to have all the data in one place.

May 31, 2013 at 6:04 PM | Unregistered CommenterLance Wallace

Joe Public
The answer to your question is "no".
But even if I did that sentence (if it is a sentence which I don't think it is) is meaningless.
Stripping out the adjectives and the adverbs we have ...
"A range of indicators continue to show changes and our understanding of how the climate system responds."
Pare it down to the real bare bones and you have:
"A range of indicators continue to show our understanding of how the climate system responds."
No matter how I tweak it I cannot get this to make sense.

May 31, 2013 at 6:11 PM | Registered CommenterMike Jackson

The AGW hypothesis is predicated on the sole assumption that anthropogenic CO2 increases "radiative forcing", however Prof Slingo's paper shows no increase in forcing while CO2 has risen fro 365 ppm to 400 ppm and temperatures have not increased at all (since 1998). Does this not conclusively falsify the AGW hypothesis, and prove that carbon dioxide does not cause global warming?

May 31, 2013 at 6:19 PM | Unregistered CommenterRoger Longstaff

Well done Keenan and Donoughue. Looks like another leg has now been been officially removed from the already wobbly stool of melodramatic alarm over CO2. Those who have stood on it to preach doom and disaster and persuade the vulnerable, the gullible, and the venal to take harmful-to-society decisions such as the Climate Change Act are looking less impressive to a few more people. Inch by inch, leg by leg, the tipping-point approaches ...

May 31, 2013 at 6:24 PM | Registered CommenterJohn Shade

Roger Longstaff (May 31, 2013 at 5:30 PM)
"Figure 1 shows no increase in radiative forcing after 1998..."
While the MO paper does indeed show a leveling-off of radiative forcing, they base it on GISS estimates (figure, data). The greenhouse gas forcing has increased by about 0.5 Wm-2 since 1998, which doesn't seem to be at issue. However, the total forcing (per GISS) hasn't budged much because of posited increases in (negative) aerosol forcing, which amount to about -2.5 Wm-2. Nic Lewis discussed aerosol forcing earlier at BH; these large aerosol values are not generally accepted, and indeed AR5 WG1 Fig 8.18 (preliminary, see Nic's post) shows an increase in total forcing of ~0.3 Wm-2 over the last decade.

May 31, 2013 at 6:28 PM | Registered CommenterHaroldW

"A wide range of observed climate indicators continue to show changes that are consistent with a globally warming world, and our understanding of how the climate system responds to rising greenhouse gas levels."

MJ, it took me two reads, but it does make sense and is parsable English. Removing a few redundant words, and expanding a bit, I get this:

"... climate indicators show changes that are consistent with a globally warming world, and that are also consistent with our understanding of how the climate system responds to rising greenhouse gas levels."

May 31, 2013 at 6:29 PM | Registered Commentersteve ta

Stevata, that was my parsing too. What sticks out to me is that there may well be "indicators that continue to show changes consistent with a globally warming world" (we'll forgive the tautology); but they must also be consistent with a non-warming world, because that is what the best indicators, thermometers, tell us we have.

May 31, 2013 at 6:40 PM | Unregistered CommenterNeil McEvoy

Figure 1 shows 35 years of warming (1910-1945) with very little increase in radiative forcing (~0.2W/m²) followed by 30 years of cooling (1945-1975) with an increase in radiative forcing of ~0.5W/m².

The 25 years of recent warming is little different from the 1910-1945 warming except that it was accompanied by an increase in radiative forcing of ~1.0W/m² - five times the radiative forcing increase of the earlier warming.

It is evident from Figure 1 that warming does not correlate with radiative forcing from increased greenhouse gases.

QED

May 31, 2013 at 6:41 PM | Unregistered CommenterBilly Liar

Harold, you say "The greenhouse gas forcing has increased by about 0.5 Wm-2 since 1998, which doesn't seem to be at issue", but I do take issue with that! Indeed, either there has been an equal and opposite effect on forcing from aerosols, or the forcing dependency on CO2 (Myhre et al, 1998) is just plain wrong (it has no observational validation).

Occam's Razor makes equal and opposite effects, sustained for 16 years, the less likely explanation and therefore probably wrong.

May 31, 2013 at 6:42 PM | Unregistered CommenterRoger Longstaff

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 | Unregistered CommenterLord Donoughue of Ashton

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.

"very high confidence (using the IPCC's definition) that the global average net effect"

Oh dear me, what is that supposed to mean?

And its: "hundreds of years of scientific advancement" - yup Arrhenius raises his head again - but the point is irrelevant padding...................oops - "supported by the development of high-quality observations and computational modeling" - oh well that's OK then but the Met Office is still saying 'it's mainly our modelling'.

You know, the impenetrable and opaque style of the rationale of this piece - has the footprint of a certain bloke, who comments here on occasion.

May 31, 2013 at 6:58 PM | Unregistered CommenterAthelstan.

Lord Donoughue

"but it suggests MO statements should be viewed with caution."

I think most visitors here are only too aware of that, more's the pity. Like others, I very much wish the Met Office would stick to science and leave the advocacy to the NGO's and other Green activists.

In the fullness of time there will be a public enquiry at enormous expense and only then will the truth come out. It took 20 years for Hillsborough. Probably take another 20 for this scam to be fully exposed.

In any event, this citizen salutes you.

May 31, 2013 at 7:01 PM | Unregistered CommenternTropywins

Roger -
I think there is a third path:
a) CO2 forcing is "plain wrong" (sensitivity zero); this appears to be your position
b) CO2 forcing increased with offsetting aerosol forcing (sensitivity high); this appears to be Hansen's position
c) CO2 forcing increased with little change in aerosols (sensitivity low but non-zero); this appears to be Nic Lewis' position.
In the last case, energy imbalance manifests itself most recently in ocean heat uptake rather than in increased surface temperatures.

May 31, 2013 at 7:05 PM | Registered CommenterHaroldW

Sir, the buck stops at Slingo.

May 31, 2013 at 7:10 PM | Unregistered CommenterMartyn

Harold, I agree with your logical alternatives (a) and (b) - and I clearly support (a) as the most likely. But the "missing heat" option (c) I find to be incredible! How so much energy could transit through upper ocean layers without leaving a trace (see the ARGO data) seems to defy the laws of physics.

May 31, 2013 at 7:14 PM | Unregistered CommenterRoger Longstaff

Keenan’s statistical model is physically wrong.

When you analyze data you choose a model. picking a model that is physically wrong ( for example a random walk for temperature) can get you a better fit, but it’s a mistake.

A good example would be people that look at ice melt in september and fit that data with a linear trend. Well, before we even start we know this model is physically wrong.

How do we know that? well at some time in the future your model will predict negative ice area. So a linear model might be useful for communicating the loss rate, but you know that its physically wrong, so you should not hang anything too heavy on it. That is, if that choice of models leads to stupid conclusions, thats a good hint the model is misleading, regardless of how well it “fits” the data.

Put another way. Keenan chose a model that fit the data better. That model says there is no warming. But looking at the data we know it has warmed. Looking at the Thames we know it isnt frozen. Looking at the sea level we know it has gone up. We know the LIA was cooler. plants know it. animals know it. ice knows it. What this means is that Keenan has chosen the wrong model. There are an infinite number of models that fit the data as well or better than his model. fitting the data “better” is not the acid test of a good model. First and foremost the model has to be physically realistic. Keenan’s is not.

May 31, 2013 at 7:22 PM | Unregistered Commentersteven mosher

I had to read this article several times and like other commenters quickly came to the conclusion that several sentences just made no sense. In fact Sir Humphrey of Yes Minister fame would be hard pressed to inject more obfuscation into any of his ministerial responses.

The comment by Lord Donoughue Of Ashton is certainly valid. I could not help but notice that very fact even on the first reading.

While all this may appear to fall under the normal governmental obfuscation, surely after this it must now be time for Ms Slingo to step aside. The reputation of the Met Office throughout the world has already plummeted under her stewardship.

May 31, 2013 at 7:31 PM | Unregistered CommenterJohnB

As I understand it, nobody is arguing whether there has been warming since the LIA.

It is whether such warming has been natural or not.

In Doug's words

"The Parliamentary Question that started this was put by Lord Donoughue on 8 November 2012. The Question is as follows.

To ask Her Majesty’s Government … whether they consider a rise in global temperature of 0.8 degrees Celsius since 1880 to be significant. [HL3050]

The Answer claimed that “the temperature rise since about 1880 is statistically significant”. This means that the temperature rise could not be reasonably attributed to natural random variation — i.e. global warming is real. "

It seems Slingo has failed totally to address this point.

May 31, 2013 at 7:38 PM | Unregistered CommenterPaul Homewood

They say that they don’t only depend on stats, they depend on ‘a deep understanding of the climate system, and ‘complex models’.

The trouble is that the hypothesis that CO2 drives everything is just a hypothesis, and the models that they use have obviously failed, as can be seen from their outputs. When you ask about these, they justify the CO2 hypothesis and the models by referring to the stats – saying that the models and hypothesis MUST be right, because there is statistically significant warming going on.

This is a common bureaucratic circular argument trick. It needs to be exposed for what it is…

May 31, 2013 at 7:42 PM | Unregistered CommenterDodgy Geezer

Way to obfuscate. Keenan is not promoting a model. He is not even trying to go back to the LIA, but if it has been warming since then, how does that fit with anybody's model? Statistically, we can't even go there, the data is not good enough. Keenan proposed nothing, he merely allows the met office to show how ill-based their position is.

May 31, 2013 at 7:43 PM | Unregistered CommenterRhoda

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

http://wattsupwiththat.files.wordpress.com/2013/05/black-box-analysis-averaged-climate-models1.jpg?w=640

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

The essence is:

Climate models can be regarded a black boxes, Global temperature output is computed by an "exponential moving average" with only 2 parameters in a trivial formula, R value 0.99

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.

May 31, 2013 at 7:54 PM | Unregistered CommenterManfred

Steve, I think Keenan's point was not that his model was right but rather that it was "righter" than the model relied upon by the MET.

As the MET was claiming statistically significant warming based on the output of its less right model Keenan makes his point. The MET claim is a statistical nonsense.

I do agree with you that the model needs to be physically realistic - for example, none of the current General Circulation models predicted the current plateau and are therefore not "physically realistic" and need to be tossed.

Now, whether a system as complex as climate can, in principle, be modelled is an interesting question. And whether we currently have data robust enough (ie. for a sufficiently time period and with a close match to observation) to plug into such a model is problematic.

Climate Science is in its infant phase and that needs to be recognized. Of course, we shouldn't base policy on infant science so the entire policy edifice constructed on non-physical models incorporating iffy data should be dismantled forthwith.

May 31, 2013 at 8:00 PM | Unregistered CommenterJay Currie

Maybe the most amazing thing is that a blog (articles & discussion) led to an official Met Office response.

Could that have happened a number of years ago, maybe it's a combination of the growing importance of blogs and the (slight) opening of the AGW debate?

May 31, 2013 at 8:04 PM | Unregistered Commenterclimatebeagle

The Met Office Blog post said:

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.

I thought the Met Office was supposed to be a British organisation yet the last word of the sentence above is spelt the American way. If their staff don't even know what country they are in no wonder the climate in the past 15 years has not obeyed their models!

May 31, 2013 at 8:04 PM | Unregistered CommenterRoy

Stephen Mosher

the issue is not whether it has warmed but whether the warming is 'significant'. You say the model chosen by Keenan shows no warming. That is not what he claims. He says

"The supplement demonstrates that the likelihood of the driftless model is about 1000 times that of the trending autoregressive model. Thus the model used by HM Government should be rejected, in favor of the driftless model. With the driftless model, however, the rise in temperatures since 1880 is not significant. In other words, the correct Answer to the Question (HL3050) might be No."

The main point Keenan is making is that the MO have chosen a model that shows warming to be 'significant' whereas there are good reasons to choose other models that may not give the same answer. Note he says 'the correct Answer to the Question (HL3050) might be No.'

May 31, 2013 at 8:18 PM | Unregistered CommenternTropywins

I have only skimmed it, but the 16 page pdf file looks reasonably sensible, except that Julia Slingo can't spell "Doug McNeall". Ths summary and blog, however, have all the usual problems.

The problem for the Met Office is that the correct answer to Keenan's question is politically impossible to give. The correct answer to the question "Is the warming since 1850 significant?" is not Keenan's answer ("no") or MyClimateAndMe's answer ("yes"), but the rather unsatisfying "your question is meaningless and unanswerable". Doug McNeall got close to this, but unfortunately chose to disguise it in his extended riff on the meaning of the word "significant". A more accurate approach would be to quote Bill Clinton: "It depends upon what the meaning of the word 'is' is".

May 31, 2013 at 8:27 PM | Registered CommenterJonathan Jones

HaroldW says at May 31, 2013 at 6:28 PM:

"...they base it on GISS estimates"

"Estimates"...kinda says it all...And I estimate it will cost us tax-payers TRILLIONS!

May 31, 2013 at 8:28 PM | Unregistered CommenterSnotrocket

Steven Mosher writes


Put another way. Keenan chose a model that fit the data better. That model says there is no warming. But looking at the data we know it has warmed. Looking at the Thames we know it isnt frozen. Looking at the sea level we know it has gone up. We know the LIA was cooler. plants know it. animals know it. ice knows it. What this means is that Keenan has chosen the wrong model.

This is nonsense: Keenan's model does not say that no warming has occurred: all it says is that the warming which has occurred is just random noise, not driven by anything. Diffusive motion is still motion.

May 31, 2013 at 8:31 PM | Registered CommenterJonathan Jones

The warmists are on the run; that shows in their actions. First the Met Office refused to answer the Parliamentary question repeatedly, then when they finally did answer, they clouded the answer with their explanation. As someone said in this thread : "a blizzard of words". They know they don't have a leg to stand on and they, probably better then we skeptics, know that it is crashing down.

May 31, 2013 at 8:33 PM | Unregistered Commenterstan stendera

steveta/Neil McEvoy
Thanks, guys. I think it's water on the brain; it hasn't stopped raining here for the last three days.
It was that comma what threw me.
But it's still virtually gobbledegook!

May 31, 2013 at 8:35 PM | Registered CommenterMike Jackson

The Slingo paper (did she really write it, do you think?) is actually not bad. It answers the question correctly, explains the statistical issues fairly well, and says that Doug was right - that an AR(1)+drift model is untenable, and that when comparing ARIMA(3,1,0) versus ARIMA(3,1,0)+drift, they're statistically indistinguishable and so there is no evidence for or against an underlying trend.

There are a couple of points that I think are disputable. They claim that they never used an AR(1)+drift model to answer the question - Doug might like to pull out the reference for that and highlight it again. And they say that they instead based their statements on the IPCC detection and attribution methodology. I think it would have been appropriate to point out what the IPCC themselves say about these methods:

"The approaches used in detection and attribution research described above cannot fully account for all uncertainties, and thus ultimately expert judgement is required to give a calibrated assessment of whether a specific cause is responsible for a given climate change."

What this argument comes down to is that the climate models (treated as a statistical model) very rarely simulate rises as steep as the 20th century rise when CO2 is set constant. Of course, we know there are a lot of features of real weather that climate models don't simulate, so it's unclear whether one can conclude anything from this - other than that the weather doesn't work like a climate model with constant CO2. You could argue that climate models are even less realistic than the ARIMA ones.

The paper is a bit slippery about the definition of 'detection', too. They spend some time demonstrating that the temperature has risen over the 20th century, although this point wasn't in dispute. The rise is bigger than the measurement error, so there's definitely been a rise. But the question was whether it was greater than the background variation. They don't claim otherwise, but you might think that given the space they devote to it that it's more relevant. However, that's a triviality.

The blog post, on the other hand, is pretty dreadful.

--

Mosher makes the claim above that "Keenan’s statistical model is physically wrong." It's a true statement, but is missing the point. Being physically wrong doesn't mean it's not a physically appropriate approximation.

When people match data against linear trends or unit root models, they're not generally claiming that such a model applies exactly, for all time. What they're generally saying in the former case is that the underlying trajectory is a curve, which in the neighbourhood of some point can be approximated by its gradient. It takes only the first term in the power series on the grounds that we don't have enough data to estimate any of the higher terms. Similarly, a unit root is assumed when there is a root of the characteristic equation close enough to the unit circle that we can't tell the difference, with the length of data we've got. You act as if it's on the circle, because that makes the statistics more reliable. But we're well aware that it's an approximation.

ARIMA models are particularly attractive for modelling physics because they act as discrete analogues for the solution of differential equations (with random inputs). The physically correct answer is very often a differential equation, and can often be well-approximated over the short term by picking out a handful of dominant terms. ARIMA(3,1,0) means it's approximately the accumulation of a variable governed by a second-order differential equation (i.e. the annual amount heat entering or leaving the system).

The objection applies, with ever greater force, to anyone who has unthinkingly slapped a linear trend on a graph. It seems to me to be objection for objection's sake. But still, this is an important point that it is well-worth understanding.

May 31, 2013 at 8:42 PM | Unregistered CommenterNullius in Verba

@ Jonathan Jones, 8:27 PM

The correct answer to the question "Is the warming since 1850 significant?" is not Keenan's answer ("no")....

That is not my answer. My answer is that the answer given by the Met Office is ill-based. My post described how a certain driftless model is 1000 times more likely than the model used by the Met Office to be the better model: that demonstrates that the model used by the Met Office is failing to explain substantial structural variation in the data, and so it should not be used. I am not proposing using the driftless model, or any other model—as other commenters have correctly noted.

May 31, 2013 at 8:45 PM | Unregistered CommenterDouglas J. Keenan

Jonathan

You say

The correct answer to the question "Is the warming since 1850 significant?" is not Keenan's answer ("no")

Is this really what Doug is saying? My understanding is that he is saying that although it is significant under the Met Office's model, at least one other model, which fits the data better, suggests that it isn't, although that other model has its own issues (physics). My reading is that he thinks the dataset is too short to answer the question.

May 31, 2013 at 8:50 PM | Registered CommenterBishop Hill

Oops, Doug beat me to it...

May 31, 2013 at 8:50 PM | Registered CommenterBishop Hill

Mosher (May 31, 2013 at 7:22 PM),

a random walk or ARIMA can, when sampled over some interval, show an increase. Does this mean that the random walk or ARIMA is not physically realistic because the data show an increase? Of course not. If you are interested in finding the cause of the increase then yes, you might argue on physical grounds that an ARIMA would not be realistic, for example. But you do not give such an argument. In fact, fairly little is know about how natural climate variability should be described.

On the other hand, if you only want to characterize the uncertainty in a model parameter like a linear trend, an ARIMA model might be suitable for describing the residual, and might in fact be more realistic than, say, an AR model, as was concluded from tests on the data.

Having said that, testing of "statistical significance" of a warming trend remains a pretty meaningless exercise (we can see in the data how much it has warmed, so we know). Such tests seem to be motivated by the desire to distinguish natural and anthropogenic warming in the data, and the misconception that a linear trend should be associated with the human influence, and its residual with natural variability.

May 31, 2013 at 8:54 PM | Unregistered CommenterCees de Valk

IOW, the question "has it warmed?" cannot be answered by just looking at the instrumental temperature record.

May 31, 2013 at 8:56 PM | Registered Commentershub

These results have no bearing on our understanding of the climate system or of its response to human influences such as greenhouse gas emissions and so the Met Office does not base its assessment of climate change over the instrumental record on the use of these statistical models.

And what in the blue blazes is that supposed to mean? That the Met Office does not make its forecasts on observations but on statistical models? If that is her admission, then we are doomed.

She (ab)uses English, wrenching it into such convoluted forms – much like balloon modelling, whereby the balloon is twisted until it looks like a dog, even though it will always be a balloon.

May 31, 2013 at 9:09 PM | Unregistered CommenterRadical Rodent

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>