Forecasting

I'm grateful to a correspondent for this powerpoint. Click images for full size.

Comments

[dusty]

BBC Lunchtime News today a Tory minister in responding to the fact that an aircraft had flown through the densest part of the volcanic ash 'red' area as predicted by the Met Office and found nothing; I paraphrase:

"......... just because no ash was found does not mean the models are wrong....." I kid you not; I despair.

[ZT]

This model is not dead, it's sleeping.

[sunderland steve]

Its pining for the fjords. :)

[Justice4Rinka]

I find it splendid that climate alarmists think they can forecast the climate 100 years ahead.

An example of a complex system most of us are familiar with is a small child. How would a climate alarmist go about modelling what a small child will be doing in say 5 years' time?

If this seems impossible, in what way is predicting the climate somehow easier?

[Diogenes]

Rules For Forecasting

•"Make lots and lots of predictions; the law of averages says some of them will actually turn out to be right."

•"If you ever happen to be right about something, don't ever let anybody forget it."

•"When forecasting provide either a number or a date— but never both."

_______________________________________________

An excerpt from:
Fiedler's Forecasting Rules

#1 The First Law of Forecasting: Forecasting is very difficult, especially if it's about the future.

#2 For this reason: He who lives by the crystal ball soon learns to eat ground glass.

#3 Similarly: The moment you forecast you know you're going to be wrong, you just don't know when and in what direction...

*****
#8 Correspondingly: If a camel is a horse designed by a committee, then a consensus forecast is a camel's behind.

As listed in The Official Rules by Paul Dickson. New York, NY, 1981.

[Diogenes]

FORECASTING

It was the late fall and the Indians on a remote reservation in South Dakota asked their new chief if the coming winter was going to be cold or mild. Since he was a chief in a modern society, he had never been taught the old secrets. When he looked at the sky, he couldn't tell what the winter was going to be like.

Nevertheless, to be on the safe side, he told his tribe that the winter was indeed going to be cold and that the members of the village should collect firewood to be prepared. But, being a practical leader, after several days, he got an idea. He went to the phone booth, called the National Weather Service and asked, 'Is the coming winter going to be cold?'

'It looks like this winter is going to be quite cold,' the meteorologist at the weather service responded. So the chief went back to his people and told them to collect even more firewood in order to be prepared.

A week later, he called the National Weather Service again. 'Does it still look like it is going to be a very cold winter?' 'Yes,' the man at National Weather Service again replied, 'it's going to be a very cold winter.' The chief again went back to his people and ordered them to collect every scrap of firewood they could find.

Two weeks later, the chief called the National Weather Service again. 'Are you absolutely sure that the winter is going to be very cold?' 'Absolutely,' the man replied. 'It's looking more and more like it is going to be one of the coldest winters we've ever seen.'

'How can you be so sure?' the chief asked. The weatherman replied, 'The Indians are collecting a lot of firewood.'

[David]

Here's something I found, written by Joseph de la Vega in 1688, the oldest book written on the stock exchange (from Wiki), so related to forecasting:

The first rule in speculation is: Never advise anyone to buy or sell shares. Where guessing correctly is a form of witchcraft, counsel cannot be put on airs.

The second rule: Accept both your profits and regrets. It is best to seize what comes to hand when it comes, and not expect that your good fortune and the favorable circumstances will last.

The third rule: Profit in the share market is goblin treasure: at one moment, it is carbuncles, the next it is coal; one moment diamonds, and the next pebbles. Sometimes, they are the tears that Aurora leaves on the sweet morning's grass, at other times, they are just tears.

The fourth rule: He who wishes to become rich from this game must have both money and patience.

Thought it was neat.

[John Shade]

A nice pair of graphics that should encourage humility in all those who choose to make forecasts for complex, poorly understood systems. In particular, it should encourage them to be very careful indeed when talking to the media, just in case the media might take their forecasts more seriously than they really deserve to be treated. Needless to say, the Bank of England forecasters would be able to come up with flannel to blether away their blunders, while producing more forecasts with absurd 'uncertainty limits' for the next wave of meetings and press releases. Ridicule won't stop them. As long as they keep getting paid, they'll do it. The complication here, and more recently with climate, is that the clients for these forecasts have the conceit to think they can control the future - in one case by tweaking interest rates, in the other by begging to be allowed to, and imagining that they can, tweak CO2 levels. Conceited opportunists making good money at the expense of the rest of us? Or am I being too kind to them?

[ManicBeancounter]

The manicbeancounter's rule of economic forecasting.

"The accuracy of a prediction is inversely proportional to its utility"
That is, it is easy to forecast where things are staying the same. But, what everyone really wants to know is when there will be a step change.

Climate models "forecast" a huge step changes, that do not appear to be happening. Economic forecasts are simply extrapolating from empirical regularities, with a bit of guesswork tied in.

[PFM]

If you go on forecasting the same thing time after time, you might eventually be right. I have noticed the Met Office using this technique in their daily weather forecasts.

The Met office claims a certain level of accuracy in its short-term forecasts; but if you forecast the weather tomorrow as being the same as today's weather, you are some 7% more accurate than the Met Office.

[KnR]

First rule of climate science, if the models and reality differ, its reality which is wrong .

[sunderland steve]

Just a thought, if the ash cloud position is based on a model, then it may or may not be where they think its. If its not where they think it is it must be somewhere else. So they could be advising other airports that its safe to fly when in fact its not! As I say just a thought.

[robin pittwood]

A 'model' is our attempt to understand and describe something - it becomes our framw of reference and the way we perceive the world. And as they say in marketing. "Perception is reality". But truly, what has happened, is that we have mixed up perception and reality.

[James P]

I heard Michael O'leary (boss of Ryan Air) making a forthright statement about the volcanic ash. He thought it was perfectly safe to fly from Glasgow, while other planes had been grounded, so he seemed to be putting his money where his mouth was. Not sure if he was allowed to, though.

[James P]

Link here

[diogenes]

james p - it all depends if the customers put their money in the pockets of Ryanair!..But seriously, there are risks with everything...it is how we try to quantify the risks that is important together with the level of kowledge of what the risks are. If you were on an o'leary flight and the engines seized up from the ash, would you sue O'Leary for taking the risk? No...you would be dead. Where is the balance of risk in this case?

[diogenes]

I also love the way that people like Tamino and "hard scientists" ridicule economic forecasting - a very small sub-section of economics. Let's look at climate science: a few physical principles, and hey you guys cannot predict the weather from one hour to the next. Economic forecasts try to deal with the complexities of Governmental, institutional plus millions of consumer interactions (which are not guided by strict rules)...and you "hard" scientists ridicule them. For shame. Would love to see a physicist model an economy. Bratby?

[sandy]

sunderland steve,
what you're not taking into account is that the models predict with an accuracy of 5 decimal places and therefore connot be wrong.

It's only fools who actually go and measure what is there and where exactly /sarc

[Brady]

It seems like Mike O'Leary (Ryan Air) may have a point against the CAA. Once you get under the name calling and the "absolutely safe" fallacy (Is crossing the street absolutely safe?) it appears that the CAA is shifting the goal posts.
http://edition.cnn.com/video/data/2.0/video/business/2011/05/24/qmb.ash.cloud.ryanair.ceo.cnn.html
First the CAA agreed last year that airlines would decide themselves outside the 120km volcano no fly zone. Now airlines must convince the CAA first before they can fly. When the airlines fly unharmed in the "red zone" then the CAA says they were flying too high. Then the CAA says that airlines must also have special equipment aboard but the only planes with this equipment are grounded. As for the CAA relying again on Met Office "projections" well ...
http://wattsupwiththat.com/2010/04/19/ash-cloud-models-overrated-a-word-on-post-normal-science-by-dr-jerome-ravetz/
But maybe I have gotten lost somewhere in this "red zone"? ;-)

[richard verney]

Further to James D's post,
I heard the O'Leary interview and he was very critical of the models. I thought to myself that these are essentially the same models that are used for climate change predictions and if they cannot effectively model the behavoir of the ash cloud over just a day or so, how can they be expected to model changes some 50 years hence in what is simply a caotic system. Of course, the AGW crowd argue that it is easier to predict 50 to 100 years into the future than just a few days into the future. It is difficult to see that that argument can be correct in a caotic system with not all variables known still less understood.

[Bruce of Newcastle]

Wonderful comparison with this successful piece of long term temperature forecasting (ht Joe d'Aleo at ICECAP).

Forecasting can be quite feasible if your assumptions come from reality and not wishful thinking.

And all done on a laptop too. Quite a contrast.

[Philip Foster]

Climate models do not make much use of the laws of physics because we do not know how to apply them to the atmosphere beyond some pretty basic stuff, and even that is an approximation. If they could 'do the physics' then the time taken to 'predict' the next ten years' worth of weather would take the best computers 10 to the power 20 years (roughly the square of the age of the Universe) to complete.... but then just think of the grants!

[diogenes]

ok Bruce...so what are the forces that underly the SOI? It looks to me as if this curve-fitting stuff - we observe a periodicity - is cargo-cult science...can you honestly claim to understand the physical principles underlying this trend?

[Ross]

diogenes -- at least those guys gave accurate predictions , even if they were talking relatively short term.

Green and Armstrong have done alot of work on the the (in) accuracy / methods of IPCC "forecasting"

http://www.ncpa.org/pdfs/st308.pdf

[Brownedoff]

The volcanic ash business:

From Global Warming Policy Foundation

http://www.thegwpf.org/the-climate-record/3005-skeptics-meet-warmists-at-cambridge.html

".............. the Met’s principle research scientist John Mitchell told us: “People underestimate the power of models. Observational evidence is not very useful,” adding, “Our approach is not entirely empirical.”

The Met Office have probably held this viewpoint for decades.

They must have told the CAA this last year when CAA found out that the only aircraft suitablly equipped for test flights was stripped down having a paint job:

Met Office to CAA: "Never mind, not to worry, old chap, observational evidence is not very useful"

This probably accounts for why the CAA never seems to have an aircraft on hand with instruments to make "observations" immediately after an Iceland volcano event. CAA believed the Met Office last year and continue to do so this year.

Until such time as the CAA get an aircraft of their own on 24 hour hot standby, the disruption to air travel after volcano events 'oop north' will continue.

[richard verney]

Further to my post of yesterday regarding the O'Leary interview, I note that the Daily Mail is carring an article on this and his criticisms of the Met Office forecasts. See: http://www.dailymail.co.uk/news/article-1390558/Ash-cloud-2011-Forecasters-half-term-threat-lifted.html

It is not much of a surprise since they struggle to get 24 hour weather predictions correct.

[SimonW]

To my mind the forecasts are the quickest way for the layman to get a grip on the paucity of the AGW case. It is difficult for the public to make a sensible decision on the intricacies of the science but it is easy to tell if someone can make useful predictions, which is after all what the CAGW policy repsonses are all about.. So an "appeal to correctness" rather than the more traditional "appeal to authority" . (This applies to much Green policy of recent decades as well; why do we listen so much to people who are consistently wrong).

Let's have rolling forecasts for regional/global temps, precipitation, hurricanes, ice, sea level, whatever, clearly stated for 1 , 5, 10, 30 years and then updated yearly with results for everyone to see. Sure a big volcano will knock them off but we can accept valid explanations and adjust for them. What's the problem with this? All this after the event tinkering with what was actually forecast (see Hansen 88 debates) is ridiculous and avoidable.

You can be sure if their forecasts were near accurate they would be plastered everywhere as "proof" of their theory.

Here is a dry but informative talk on the limitations of Climate forecasting from last years Heartland conference by J Scott Armstrong PhD

http://www.heartland.org/environmentandclimate-news.org/ClimateConference4

(last one on top row).

[Bruce of Newcastle]

Hey Diogenes, if you made curve fitting unethical you'd do most of the FTSE 'technical analysts' out of a job. While I have a certain concern about said analysts given the GFC, you can't say their services aren't valued given how much the City pays them.

As for the link I gave, try graphing HadCRUT versus previous solar cycle length. This is related to the link I posted. Now explain the correlation please. The same correlation can be seen by graphing the longer CET dataset vs previous solar cycle length.

Furthermore ENSO is cyclical with a close relationship with the 65-ish year wavelength of the PDO and AMO. Unfortunately you need a longer dataset to see the cycles clearly, which is much more apparent in datasets like HadCRUT.

No I can't explain why SOI/ENSO is cyclical, but the data is quite clear that it is.