Stat Model Predicts Flat Temperatures Through 2050

Doug L. Hoffman

The Resilient Earth

Friday, Dec 18th, 2009

While climate skeptics have gleefully pointed to the past decade’s lack of temperature rise as proof that global warming is not happening as predicted, climate change activists have claimed that this is just “cherry picking” the data. They point to their complex and error prone general circulation models that, after significant re-factoring, are now predicting a stretch of stable temperatures followed by a resurgent global warming onslaught. In a recent paper, a new type of model, based on a test for structural breaks in surface temperature time series, is used to investigate two common claims about global warming. This statistical model predicts no temperature rise until 2050 but the more interesting prediction is what happens between 2050 and 2100.

David R.B. Stockwell and Anthony Cox, in a paper submitted to the International Journal of Forecasting entitled “Structural break models of climatic regime-shifts: claims and forecasts,” have applied advanced statistical analysis to both Australian temperature and rainfall trends and global temperature records from the Hadley Center’s HadCRU3GL dataset. The technique they used is called the Chow test, invented by economist Gregory Chow in 1963. The Chow test is a statistical test of whether the coefficients in two linear regressions on different data sets are equal. In econometrics, the Chow test is commonly used in time series analysis to test for the presence of a structural break.

A structural break appears when an unexpected shift in a time series occurs. Such sudden jumps in a series of measurements can lead to huge forecasting errors and unreliability of a model in general. Stockwell and Cox are the first researchers I know of to apply this econometric technique to temperature and rainfall data (a description of computing the Chow test statistic is available here). They explain their approach in the paper’s abstract:

A Chow test for structural breaks in the surface temperature series is used to investigate two common claims about global warming. Quirk (2009) proposed that the increase in Australian temperature from 1910 to the present was largely confined to a regime-shift in the Pacific Decadal Oscillation (PDO) between 1976 and 1979. The test finds a step change in both Australian and global temperature trends in 1978 (HadCRU3GL), and in Australian rainfall in 1982 with flat temperatures before and after. Easterling & Wehner (2009) claimed that singling out the apparent flatness in global temperature since 1997 is ’cherry picking’ to reinforce an arbitrary point of view. On the contrary, we find evidence for a significant change in the temperature series around 1997, corroborated with evidence of a coincident oceanographic regime-shift. We use the trends between these significant change points to generate a forecast of future global temperature under specific assumptions.

Read the rest of the article here.

Advertisements

  Subscribe  
newest oldest most voted
Notify of
phlogiston

This reminds me of Asterix and the Soothsayer – the bogus soothsayer’s favourite quote was “this I had forseen” after everything that happened. The peasant villagers were in awe of him!

Henry chance

The only reason the exaggerating warm mongers predict so far out is because no one has evidence to post against their claims.
Today. Who predicted this northern hemisphere frigid events last summer? If they didn’t they need to be ignored for having defective models.

Henry chance

declare Real Climate a Best Religious Blog Weblog Award Finalist for 2009.
The top four vote recipients, in order, were:
1. One Cosmos 253
2. RealClimate 149
It makes sense they predict warming by reason of faith.

What sense has to base anything on HadCRUT? It is like genetic researcher basing its study on Lysenko pseudoscience.
Back to the drawing board and while at it, check the Arctic temperatures record without UHI: http://climexp.knmi.nl/data/icrutem3_hadsst2_0-360E_66-90N_na.png

P Gosselin

There’s a lot crystal-balling behind all this. +/- 1 degree C either way over 100 years wouldn’t be noticeable if no one talked about it.

Aaah quick ship a copy off to the Australian government – they keep thinking the earth has a terrible fever and they must tax us all to death to stop it!

Jimbo

All these climate predictions / forecasts are utter crap. What matters is what will happen to the climate in the next 100 years and nobody knows. The Met Office (if honest) can tell you how useless their computer forecasts are for the next 12 months.
I will remain a sceptic until they have proof of AGW and not warm weather events, otherwise I will cite the freezing weather in the USA, Europe, China etc., as a sign of the arrival of a Little Ice Age.

People like Sarah Archdale might be here to tell you all about it: http://www.yorkpress.co.uk/news/4835291.Injured_cyclist_found_unconscious_in_snow/
Ecotretas

astonerii

I read this a few weeks back. While it seems to have a good outlook, something that bothers me is the durations of future flat/cooling and warming trends. If you look over the past few trends, they had durations of around 30 years, and these guys are looking at 40 and 50+ year trends.

Eddie

Juraj….i was thinking the same thing. Shouldn’t HadCRUT be tossed out by now in experiments used to predict anything?

AdderW

yet another model that PROVES nothing

Jimbo

On the Thursday, 10 December 2009 the BBC reported:
“The global average temperature could reach a record high in 2010, according to the UK’s Met Office.”
http://news.bbc.co.uk/2/hi/science/nature/8406839.stm
There they go again! Do these climate tricksters have any sense of shame? It’s their continued failed
predictions
that might ultimately falsify AGW.

Henry: “Today. Who predicted this northern hemisphere frigid events last summer?”
I believe Obama’s Science Czar did… back in 1971. He sure knows how to pick ’em!

DirkH

If i understand this correctly they use a technique usually used in technical analysis of stock charts. I won’t say they’re prediction is better or worse than the ones you would get with an unbiased GCM but it’s a fresh start. Gotta look into that Chow test.

AdderW

Henry chance (13:13:05) :
declare Real Climate a Best Religious Blog Weblog Award Finalist for 2009.
The top four vote recipients, in order, were:
1. One Cosmos 253
2. RealClimate 149
It makes sense they predict warming by reason of faith.

That was fun !!

But… but… but… I thought the science was settled! Now I’m confused.

What will the 2 M say to it ;.)

Let’s see if I have this straight: Since the all the previous Climate Change models have now been shown to be fatally flawed we’re using a new model which is based on a test devised by an …
ECONOMIST ?!!
Did everyone else sleep through the past 18 months and miss the great recession?
To paraphrase Gene Siskel, this isn’t scraping the bottom of the barrel, this is digging up the ground underneath the barrel.

So, I could use a random number generator and produce a prediction that is better than the “re-factored” GCM. Actually, I thinks Patrick Michaels showed that this was the case before the re-factoring.
As a note, software developers usually don’t consider it “re-factoring” if the output of code changes. Refactoring generally refers to making code more portable and maintainable.

Royaul52

Slightly OT, but have we reached a tipping point for temp decline acceleration with the increased snow cover? Seems apparent to this untrained eye on Cryosphere that there has been a huge increase in N Hemis. snow cover, including about 1/2 of the British Isles.
Any models have temp predictions that take this into account?
Also, we’re hearing almost nothing current about the arctic ice pack.
Just wondering!

RK

Isn’t this awfully similar to technical analysis in financial trading: statistical models that test well against the past but with uncertain predictive power? As far as I can tell, there is no real hard science (“laws of climate”) within the climate models.
In financial trading, when money managers use the wrong models to deliver suboptimal results, they get fired. Too bad there is no such feedback mechanism for the bad climate models.

Isn’t the stepwise function probably due to poor data?
http://www.theclimateconspiracy.com/?p=180
Anthony, is it at all possible to make some type of chart that shows where all the data comes from, which goes into which data sets, and all of that?

jorgekafkazar

Meh.

And just how good were those econometric models at predicting our present economic melt down?
Look, I am an economist. And there is much about the field that is valid and that I enjoy. But: I make a living from doing what economic theory says it is not possible to do. I “beat the market”. So do a large number of other traders. If computer models of chaotic systems could predict with any skill, you would see a whole bunch of rich economists. Instead what you see is a whole bunch of rich traders (who’s skills are more closely tied to those of gamblers than economists) and a bunch of very poor economists.
FWIW most of the computerized trade models break down for exactly the reason shown in the example above. They are “trend following” models and try to find ever more clever ways to project the trend. They can even work well for long periods of time (for example, I ‘trend followed’ right up to 2008). THEN they break down. Horridly. See 2008-2009 for what that looks like…
Trading is a very different skill set. It is all about following the trend when you have it AND doing very early very fast “inflection detection” when it changes. And chaotic systems DO change, and fast.
IMHO, as a person who trades for a living, 1998 was a “blow off top” in temperatures exactly was we see in stock charts. Now I can’t at all tell you why they both act the same, but I can tell you that I see the same patterns in them. At many tops, you get an unsustainable spike up, then a plummet.
As a trader, catching those inflection points is everything, and that is why 90+% of folks who trade, lose. They win for a while, then miss an inflection, and are wiped out.
That is what I see happening now in the AGW crowd. The inflection happened (IMHO on a 60 year PDO, the 30 year 1/2 cycle has swapped) and they are finding ever more desperate ways to show that their trend following model will start working again Real Soon Now. If they are like most folks, they will continue this behaviour right into the next little ice age.
That they chose to use econometric tools to do this is, to me, confirmation of their error. If you want to be right, use trader tools… I’d use a MACD – moving average convergence divergence. We have about 200 years of decent data, so I’d use a 25 / 12 / 6 years set of simple moving averages for my moving average stack and the usual MACD formula uses a 9 period exponential moving average as the signal line against a 12/26 baseline. Visualizing that (I’ve stared at a million of these, so I can visualize it now even though it will be encrypted Greek to most folks) shows a top inflection. MACD will be screaming down (EMA of 9 years since 2000? Yea, down. While 12/26 period ought to be up tilted rolling into flat). A 25 year SMA will be on top flattening, while a 12 year will be flat and a 6 year will be under it. “SMA stack rolled over”.
And that would be a screaming SELL!! top inflection no doubt about it.
And frankly, I’ll take ‘trader tools’ over “economist tools” any day. One works and makes me money. The other makes pretty pictures and gets you tenure (but don’t ever trade based on it.)
For folks wanting to see what this trader stuff looks like, you can look here:
http://chiefio.wordpress.com/2009/12/22/wsw-monday-december-21-2009/
I’m “due” to make an updated posting of it, but the charts are live charts and you will see what the market looks like today. Down toward the bottom is a link to a description / explanation of the “indicators” that will do a better job than I’ve done in this comment.

joe

Nice model, did they use a ruler and pencil for that graph?
By the way, the Earth will continue to warm after 1000 years pause. There is no way to verify this, but it is what our reliable model(s) predicts.
Nailed it!!!
Global Warming is still ON!

Proves crap GCMs can predict any crap they are programmed to predict — Known as GIGO.
Let’s test these crap GCMs, set the initial date at AD 1 and predict the next two millenium. Publish your results … in a language even a polar brown bear can decipher.
What is the excuse, it’s almost cost free, it’s just computer time.
Engineers call this model verification.

Galen Haugh

Henry chance (13:09:18) :
The only reason the exaggerating warm mongers predict so far out is because no one has evidence to post against their claims.
Today. Who predicted this northern hemisphere frigid events last summer? If they didn’t they need to be ignored for having defective models.
—-
Reply: I dunno, Henry…you have to remember that “weather” isn’t climate until it’s aged at least 100 years, then it suddenly becomes “climate”. This lets all of ’em off the hook, obviously, and puts them in the driver’s seat when it comes to imposing policy, collecting taxes, fudging data, etc. I mean really, who in their right mind could argue with them that far out? It will be our fourth or fifth generation of grandkids that will have to watch this fairy tale work out and by then nobody will remember what the predictions were. But it won’t matter–they’ll be too busy scratching out a living to care.

Paul Vaughan

I’ve just completed a series of wavelet analyses that reinforce earlier analyses I’ve done that support Barkin’s suggestion that the Earth “galloped” in ~1998. Statistical inference is a HOPELESS way to approach the problem (at this stage), since we don’t yet have the conditioning worked out. More effort needs to be invested in identifying the combinations of factors that lead to regime shifts. Mainstream assumptions of randomness will continue to mislead in the meantime. I suggest that people study the EOP (Earth orientation parameter) records very carefully as they are full of clues about terrestrial regime shifts. Because of the polar motion record, for example, I’ve now isolated time-normalized wavelet-power curves for aa index, QBO (quasi-biennial oscillation), & AO (arctic oscillation) that appear nonrandomly related (more detail another day – suffice it for now to say these are NOT “simple linear correlations”).

d

this is just more proof that you can make a model do what ever you want it to do. the in thing now is to hav some stable time then global warming will take over again in some arbitrary span of time. watch out every new model will have this as its new feature.

PaulH

Why don’t they just tell us who will win the 2050 Super Bowl? Their computer-modeled sports prediction will have just about as much value as their weather forecast.

E.M.Smith (13:50:41) :
Sounds like you’re the Engineer Vs the Economists theory group. it seems to me that engineers have the edge Vs theory in that their stuff has to work. Trader tools Vs Economist tools indeed. 🙂

A copy of the paper is available here http://arxiv.org/abs/0907.1650.

Tuck

E.M. Smith: Incredibly well put.

Paul Vaughan

No wool over these wise eyes:
E.M.Smith (13:50:41) “[…] “trend following” models […] can even work well for long period of time […] THEN they break down. Horridly. […] And chaotic systems DO change, and fast.”
What we need is a tool for detecting “the invisible hammer”. At first we may have to settle for identifying when it will pound. Maybe later we’ll be able to work on which way the amplitude will go. Phase first. We need to first get a handle on the phasing of the sign of absolute deviations (from the mean or from zero or from whatever reference is valid, depending on the nature of the phenomena). Warning to researchers: Actual hard work ahead. (Tip: Skip the linear-correlation methods.)

Clive

These future climate predictions are absolute rubbish. We all know that, I just had to get t off my chest. No one knows.
Quoting E.M.Smith, If computer models of chaotic systems could predict with any skill, you would see a whole bunch of rich economists. Instead what you see is a whole bunch of rich traders … and a bunch of very poor economists.
This is both amusing (sorry) and enlightening. Exactly. Thanks.
Clive
From the Frozen North
(Despite what the UK Met office claims about Canada’s recent weather. ☺ )

Peter of Syndey

I take the view making predictions on climate is a total wast of time. Just wait and we’ll know soon enough over the next few years whether the IPCC models have any supporter. Clearly if the IPCC predictions are correct then global mean temperatures must rise dramatically over that period. If they don’t then the models are definitely wrong and will be dumped even by the IPCC if they are to avoid being recognized by everyone as a complete joke.

geo

What, no “return with a vengeance” after 2050? In my eyes, one of the more serious problems with AGW model programmers is that it seems they just can’t believe there won’t be a “monontanous” increase over significant periods. . . so in their eyes, any intermediate cooling *must* be followed by a period of “catch-up” increased warming to return to their model trend lines.
It would be interesting to know what they would show if they had the courage to believe that there would be natural variation intermediate breaks that would be followed by the previous trend of warming rather than a “catch-up” over-trend warming to get back to their overall trend. But then they might have to give up scare-mongering about imminent “tipping-points”, and they certainly don’t seem willing to do that.

rbateman

E.M.Smith (13:50:41) :
Climate has been known to impact economic mood. Sometimes it ominously preceeds it, but only in hindsight.
The Great Depression had it’s Dust Bowl.
The Great Recession has it’s Deep Solar-Polar Bowl.

A model, is a model, is a model. A model can only do what it can do and nothing more. A model that is based on, or where the underlying science is not fully understood, is little better then chance. That said, this does not mean models are of no value. It is just to remind us the difference between mythos and logos. dn

George E. Smith

Well I almost believe the GCMs; well almost right up to the very last real data point (if any) that was entered into the simulation run.
But go one nanosecond beyond that last datum entry, and you are off in the wild blue yonder as far as I am concerned.
I’d rather have you predict in what compass direction (0-360 degrees) the center of mass of the world’s termite population is going to move, as a result of the coming anthropogencimanmade climate change; that is supposed to cause everything to move a quarter of a mile each year. Get back to us when you have that worked out.

They need to apply Black-Scholes to the data to find out what their climate derivatives are really worth.
And yes. That is sarcasm.

Boudu

declare Real Climate a Best Religious Blog Weblog Award Finalist for 2009.
The top four vote recipients, in order, were:
1. One Cosmos 253
2. RealClimate 149
It makes sense they predict warming by reason of faith.

That’s what I love about sceptics. They have a sense of humour 😉

George E. Smith

I’m looking at that wide swath of future dotted lines, and I’m trying to picture which of them proves the logarithmicity of the temperature anomaly/CO2 abundance; that the climate has faithfully follwed in the 600 million odd years since the Cambrian era.

The Great Recession has it’s Deep Solar-Polar Bowl.
What is the point spread?

Boudu

And one of these days I’ll crack those html tags !
[I fixed it. This site uses HTML, not BB code. So use the arrows above the comma and period keys instead of brackets, and just use b for bold. ~dbs]

wsbriggs

Econometrics – voodoo by any other name…
Mandelbrot wrote a book in 2005 about the Misbehavior of Markets, he warned that fluctuations weren’t Gaussian, they were L-stable, fractal distributions with long tails, the long tails are where E.M.Smith is making his money.
Analogously, the Climate fluctuations are L-stable as well, but none of the folks who are “trend followers” are going to make any money at all – bring on the Bears (Polar Brown Bears)!

George E. Smith

I keep forgetting to post this, so here.
http://www.geocraft.com/WVFossils/Carboniferous_climate.html
Now there is the data both CO2 and temperature; Lord Monckton has himself cited this data.
So I don’;t know where the raw numbers are stored, but some of you younger folks with much better eye sight than I have might try extracting some point by point simultaneous values for global temperature, and atmospheric CO2, and then plot us a graph of (T) versus Log(CO2). I’m sure just about any log base will do just fine; but if you are clever and use base (2) , then you will automatically come up with the correct value for Steven Schneider’s “Climate Sensitivity” fundamental physical constant.
Well sorry I don’t have any grant money to support that research; but when Obama finds out about your Herculean efforts, I am sure he will make you the “climate sensitivity” Czar, or Tsar, as you wish.

Roger Knights

E.M.Smith:
IMHO, as a person who trades for a living, 1998 was a “blow off top” in temperatures exactly was we see in stock charts. Now I can’t at all tell you why they both act the same, but I can tell you that I see the same patterns in them. At many tops, you get an unsustainable spike up, then a plummet.
As a trader, catching those inflection points is everything, and that is why 90+% of folks who trade, lose. They win for a while, then miss an inflection, and are wiped out.
That is what I see happening now in the AGW crowd. The inflection happened (IMHO on a 60 year PDO, the 30 year 1/2 cycle has swapped) and they are finding ever more desperate ways to show that their trend following model will start working again Real Soon Now. If they are like most folks, they will continue this behaviour right into the next little ice age.
…………
And frankly, I’ll take ‘trader tools’ over “economist tools” any day. One works and makes me money. The other makes pretty pictures and gets you tenure (but don’t ever trade based on it.)

Three cheers! IOW, to quote a trader’s maxim, “The trend is your friend until the bend in the end.” These warmists can’t conceive of a “bend” beyond the scope of their projections, they have no feel for the dipsy-doodle Nature Of Things, they’ve got no swing — i.e., they forbid nature to have any significant “variations” of its own. They are “straights” or “squares” in the bad sense of the words. They arrogantly impose their logical bean-counting structures (“forcings,” “heat budget”) on reality and think they have all its factors accounted for — and thus they qualify as natural-born losers, like stock-trading, formula-fooled economists.
You can just bet that the Pranksters on Olympus have seen these flatulent flat-landers as ripe for a comeuppance, and have been egging them on into ever-greater hubris, in order to make their nemesis (begun this fall) more fulfilling. “It’s cosmic!” as we used to say, back in the day — i.e., it’s a cosmic joke.
Such a pie-in-the-face is what the fates are poised to deliver to every profession and institution in a top-lofty position: tempt it to go overboard with its fancied omnipotence and/or omniscience, to its ruination (or at least to its spectacular embarrassment). It’s been science’s turn to make a gigantic jackass of itself for quite a while, and the fates have laid out a tempting trap (“save the world and be lauded as saviors thanks to the skills and character possessed by members of your guild”). A trap into which science’s Leading Lights (its gatekeepers and societies and publications and popular popularizers) have boldly put their collective foot.
Snap! Now the thrashing begins. Lay on the popcorn, and damned be him that cries, ‘Enough!’ (After a dozen years of their twisting and turning, maybe we can let them off the hook, if it looks like they’ve learned a lesson.)
Envoi: Currently colleges (in effect) teach that “(those) men were fools,” which implicitly suggests, “but not us,” instead of teaching that “man is foolish,” which is more humbling and truer. But, because this lesson is humbling, it hurts, so people (and especially world-improvers) don’t want to hear it — they “can’t take the truth.” It’s much more fun to laugh or sneer at those OTHER jackasses.
So much the worse for them. “Don’t tempt fate” and “all fame is fleeting” are the lessons a liberal education should somehow get across to students. If OTC it implicitly inculcates arrogance instead, it betrays its students and itself.

George E. Smith

It might be useful to note that the last time that atmospheric CO2 levels were as low as they are today, was about 267 million years ago; go figure.

greg2213 (14:06:14) : edit
E.M.Smith (13:50:41) :
Sounds like you’re the Engineer Vs the Economists theory group. it seems to me that engineers have the edge Vs theory in that their stuff has to work. Trader tools Vs Economist tools indeed. 🙂

Well, I’m an “odd duck” in that I’m in both camps. Most of my professional career was as a “computer geek” building things that had to work (though I traded on the side building my skills). But I love theoretical Econ (just don’t confuse it with reality 😉
My degree is in Econ, but when I went to school there was no Computer Science degree. You could be a math major, or an Engineer, or go into econometrics if you wanted to play with the computer… I have a LOT of Engineering Department units on my transcript as that is where the econometrics guys took computer classes. (Eng. calculus and physics too) Even a couple of college courses in things like semiconductor theory and transistor design. Along the way I also picked up “Biomedical Applications of Computers” – speaking of things that just have to work…
But I loved Econ (got an “easy A” in Econ 1 and loved it while everyone else was moaning… took that as a sign 😉 so I also have “linear programming optimization” and “International Trade & Finance” classes and such.
Though, yeah, I’m much more impressed by the guy who makes hard things work than I am by the theorist with 4 degrees who says he couldn’t.
FWIW the “Consensus” in investment theory is “The Efficient Market Hypothesis” that says you can not beat the market. Despite the long list of folks like Warren Buffet – as a recent example – who regularly do. See Goldman Sachs trading desk for an example of a group of folks consistently beating the market. So which was the correct market price for Goldman? The $200 prior to the crash, the $56 I bought it at during the crash? Or the $120 – $140 I sold it at? Or the price today? The efficient market hypothesis says “All of them!” Or how about Lehman or Bear Stearns – both market darlings up to just a few weeks before they became zero….
So I love Econ, and it has much to offer (for example, understanding fractional reserve banking and foreign trade balances); but breaks it’s teeth on the issue of econometric modeling and trading.
Humor Side Bar:
The story goes that Woz (of Apple) was getting his college degree (at last, having made millions without one) under a pseudonym for privacy; and was assigned to do a floppy drive design. He, of course, turned in a circuit remarkably like the one he used in the Apple and Apple II (that had some ‘different’ behaviours, but also used about 1/2 the parts).
Supposedly, he was admonished by the instructor that he had left out some critical parts (a pull down or buffering resister of some sort, IIRC the story) and had great difficulty not telling the guy it worked fine and had shipped a few million units! (and blowing his ‘cover’) Theory, meet realty 😉 Woz had originally made the design, so the story goes, by starting with a standard complex one then clipping out parts to see what was really needed and what was ‘fluff’, then tuning it.
And guess which of the two made the most money 😉
Always bet on the engineer over the theoretician.
(Though IMHO, having a balance of both works best.)