Click images for more details



Recent posts
Recent comments
Currently discussing

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

Powered by Squarespace
« Cold on Nelson's column | Main | Well sampled science - Josh 209 »

Met Office or bookie's office?

Roger Harrabin takes a wry look at the Met Office's three-month forecast for last spring. These longer-term forecasts are not published, so Harrabin got it under FOI. The results are most amusing:

The Met Office three-monthly outlook at the end of March stated: "The forecast for average UK rainfall slightly favours drier than average conditions for April-May-June, and slightly favours April being the driest of the three months."

A soul-searching Met Office analysis later confessed: "Given that April was the wettest since detailed records began in 1910 and the April-May-June quarter was also the wettest, this advice was not helpful."

The whole piece is worth a look, but I was struck by this bit:

The Met Office explained it this way: "The probabilistic forecast can be considered as somewhat like a form guide for a horse race.

"It provides an insight into which outcomes are most likely, although in some cases there is a broad spread of outcomes, analogous to a race in which there is no strong favourite. Just as any of the horses in the race could win the race, any of the outcomes could occur, but some are more likely than others."

The analogy between weather forecasting and backing horses is amusing, particularly in the light of a Twitter conversation between the FT's Jim Pickard and Bob Ward last week:


HOTTEST MARCH EVER AS MINI-HEATWAVE HITS UK: Is this the worst weather forecast since Michael Fish?

. Yes, that's what happens when journalists confuse Ladbrokes with the Met Office.

Perhaps it's different when the Met Office confuses the Met Office with Ladbrokes.

PrintView Printer Friendly Version

Reader Comments (99)

Paul Butler: "most people fail to understand the notions of risk and possibility". Quite right; I've yet to come across a member of the public who understands what is meant by "risk". I suspect you don't either, otherwise you wouldn't put "risk and possibility" together.

Mar 29, 2013 at 4:02 PM | Registered CommenterPhillip Bratby

If the 'barbecue summer' (or the wet April) had materialised, would Ms Slingo now be wittering about probabilities? Probably not.

Mar 29, 2013 at 4:06 PM | Unregistered CommenterJames P

This is so odd. Betting on the year's weather is not like betting on horses that carry no odds!

Weather is auto-correlated. And you have the super-computers. There are odds. They exist.

The reason why the Met Office is getting it wrong is because it is getting the odds *wrong*. If there is a cooling tendency, and you persist in making warm forecasts, your predictions are wrong because you are betting wrong. Simple.

Mar 29, 2013 at 4:07 PM | Registered Commentershub

The 3-month outlooks are on the Met website here (or at least the current ones are).

I am surprised Harrabin did not know this.

Mar 29, 2013 at 4:15 PM | Unregistered CommenterPaul Homewood

Since we're on weather forecasts, here's something I've wondered for a while.

For how long have the BBC/MetOffice been not showing minimum night time temperatures on their TV weather forecasts?

I noticed a while ago that "minimum" night-time temperatures shown during weather forecasts appeared to be quite a lot higher than we were actually experiencing. This is borne out by the fact that the forecaster will occasionally state something like "Mininmums of around 3 C tonight, but out in the countryside it could get down to -4 or -5 C". The map will show lots of 3s and 4s, but nothing anywhere near the TRUE mininums.

The trick ("trick": a neat way of solving a problem (copyright "The Team")) is that the night time minima that are shown are minimum temperatures for towns and cities. The true, actual, minimum temperatures for the evening - those that occur in the rest of the country (which is to say, the vast majority of it) - are usually not revealed.

This, of course, has the effect that people are given the impression that nights are warmer than they actually are, and were in the past.

I suspect that the maxima that are forecast are also maxima for "towns and cities" and are therefore several degrees higher than the "natural" maxima elsewhere, however there is at least some justification for reporting actual maximum temperatures as maximums.

What there is no justification for, IMHO, is pretending that night times are warmer than they are because of a UHI effect.

I don't know if this is a recent change, but I'm willing to bet it's all part of "creating a story" around climate change.

Mar 29, 2013 at 4:37 PM | Unregistered CommenterAngusPangus

Come on Richard (Betts) support your Boss (Sligo).

Mar 29, 2013 at 5:40 PM | Unregistered CommenterDon Keiller


Please don't bait other commenters.

Mar 29, 2013 at 5:49 PM | Registered CommenterBishop Hill

Much as I have utter disdain for the output of the Met Office , and its pro warming agenda, be wary of placing the "ash cloud" fiasco at their door.

The problem at the time was how the aircraft manuals were written, and how aircraft could be operated LEGALLY AND SAFELY , in that environment . It took me over 12 months to convince a major engine manufacturer to add THREE words to their maintenance manual , which would have allowed the aircraft to fly during the Icelandic event .

Mar 29, 2013 at 6:06 PM | Unregistered CommenterNeil


These longer-term forecasts are not published, so Harrabin got it under FOI.

Actually they are published, see here - but I don't blame you for getting this wrong, as you probably saw Harrabin's incorrect claim that the seasonal forecasts are "withheld from the public". They aren't - they are published online.

I don't know why Harrabin felt he had to use FOI to get last spring's forecast, he could have just looked on Watts Up With That

Scientific criticism is perfectly fine, but incorrectly claiming that information is withheld and unneccessarily making a show of using FOI unnecessarily is another thing altogether.

PS I was going to post this before I saw Don's bait. I'm not going to rise to it - it's the bank holiday weekend and I'm not in the mood for a scrap with Don... :-)

Mar 29, 2013 at 6:22 PM | Registered CommenterRichard Betts

A couple of the comments on this thread have been quoted in a post at WUWT:

Mar 29, 2013 at 6:29 PM | Unregistered CommenterSJF

I pointed out to Roger H that it would have been the perfect April Fool's story to blame the vagaries of the weather on the Mad(den) Julia(n) Oscillation...

Mar 29, 2013 at 6:38 PM | Registered Commenteromnologos

I have noticed this -4/-5 colder in the country thing on East Midlands forecasts for a number of years. My guess is that they will justify it my saying the majority of people live in Derby/Nottingham/Leicester so the forecast is for those people.

Mar 29, 2013 at 6:46 PM | Unregistered CommenterSandyS

The form guide is fact. The predictions by tipsters can be equated with Met Office forecasts.

Mar 29, 2013 at 7:08 PM | Unregistered CommenterBC56

After a few more years of non-warming and probably cooling, together with blackouts and gas rationing from our insane energy policy and the high number of deaths due to the cold, Government ministers are likely to point the finger of blame at the Met Office.

I wonder whether Ms Slingo will give the same reassurances that they are really very accurate because they offer probabilities for the full spectrum of possible weather outcomes and the actual weather result has always had a positive and finite probability, so they got it right.

Mar 29, 2013 at 7:12 PM | Unregistered CommenterSchrodinger's Cat

Richard Betts: "...but I don't blame you [BH] for getting this wrong..."

What???? Richard, how can you say that when Harrabin on the BBC News site says this:

"Last spring's forecast has been obtained by BBC News under Freedom of Information."

I think you owe the Bish an apology.

Mar 29, 2013 at 7:13 PM | Unregistered CommenterSnotrocket

Having been persuaded to go against my religion, by His Lordship, I perused Harrabin.
Could any of the renowned experts on this site tell me what is the Madden Juli(?)an Oscillation.
Please disuade me that it is not another wet excuse from the Met to get their prognostication more wrong in the future.
Perhaps they need to get their M(o)JO working?


Mar 29, 2013 at 7:15 PM | Unregistered Commenterpatrick healy

Actually they are published, see here - but I don't blame you for getting this wrong, as you probably saw Harrabin's incorrect claim that the seasonal forecasts are "withheld from the public". They aren't - they are published online.

Not entirely true:

Following public research, the Met Office no longer issues long-range forecasts for the general public; instead we provide a monthly outlook on our website.

Despite this, you may have seen some reports in the media on Thursday, suggesting the Met Office has produced a forecast for the coming winter.

These media reports have based their interpretation for the coming winter on probability maps on our website. However, they have been selective about the information they have used and you should not take these interpretations as a guide to the coming winter. Instead we would recommend using our monthly outlook and short range forecasts.

I can see how Roger can be getting crossed signals, for I could not find the retraction press release.

Mar 29, 2013 at 7:50 PM | Unregistered Commenterintrepid_wanders

Hi Richard Betts,

with reference to my earlier post, perhaps after the Easter break, would you be able to confirm when the Met Office/BBC started reporting urban minimum temperatures rather than actual minimum temperatures in TV (and presumably radio) weather forecasts? At whose instigation was this and why?

Many thanks in advance.

Mar 29, 2013 at 8:05 PM | Unregistered CommenterAngusPangus


Please read my post carefully. I said it was Harrabin that originally claimed incorrectly that the seasonal forecasts were "withheld". And yes, I know Harrabin's article did say the BBC got last year's spring forecast through FOI - my point was, he didn't have to use FOI, because it was in the public domain the whole time, and linked from Watts Up With That.

Mar 29, 2013 at 9:02 PM | Registered CommenterRichard Betts

Richard Betts is correct - we all knew about the hopeless Met Office forecast. There were dozens of comments on BH about it at the time (several from me), No need to go to WUWT.

If Harrabin has to use FOI to find it what the hell are we paying our BBC taxes for?

Mar 29, 2013 at 9:16 PM | Unregistered CommenterRoger Longstaff

I live in rural Herefordshire. The forecast for this area for this evening is 0 degrees c. My thermometer outside reads -4 degrees c.

Mar 29, 2013 at 9:32 PM | Unregistered CommenterA Lovell

has the met office ever confirmed the text published at WUWT was its?

Mar 29, 2013 at 9:32 PM | Registered Commenteromnologos

Richard, you may not have meant to make it seem like it was the Bish's fault (OK, it was Harrabin's), but Harrabin probably used FOI because the forecasts on the MO site ( ) are so difficult to get to or figure out.

As example, I give you the equivocation of the current three month forecast (my bold):

"Period: April 2013 - June 2013 Issue date: 21 March 2013

This outlook provides an indication of possible temperature and rainfall conditions over the next 3 months. It is part of a suite of forecasts designed for contingency planners.

The outlook should not be used in isolation but should be used with shorter range and more detailed (30-day, 15-day and 1-5 day) forecasts and warnings available to the contingency planning community from the Met Office."

Richard, considering that you are an expert, and that you go on to say that the forecast is based on (inter alia) 'expert judgement', why do you expect lay readers to become experts at interpolating your forecasts. Surely, you should be giving it to them as a finished product?

Mar 29, 2013 at 9:40 PM | Unregistered CommenterSnotrocket

Don't let Richard Betts deflect us from the substance of this post. It is not about FOI it is about the ability of the Met Office to make medium range forecasts. As far as I can see the MO scores nil points on that issue.

If Richard thinks otherwise he will no doubt address that issue.

Mar 29, 2013 at 10:00 PM | Unregistered CommenternTropywins

Just as Feynmann accepted his reponsibility for his theoretical work on the atomic bomb, so Slingo, Betts, et al, must accept responsibility for the consequences of their actions. Or their inaction.

Mar 29, 2013 at 10:45 PM | Unregistered CommenterRoger Longstaff

[snip] Please do not bait other commenters. BH

Mar 29, 2013 at 11:01 PM | Unregistered CommenterDon Keiller

Trouble is, people, you just wait until the MO get a forecast wrong (inevitable since they are probabilistic), then weigh in as if all of their long-range forecasts are wrong!

Quite a good demonstration of how most people fail to understand the notions of risk and possibility
Mar 29, 2013 at 3:29 PM | Unregistered CommenterPaul Butler

On the contrary. My original post was complimentary about the Met Office's performance when I was motorcycling in the UK. I think unexpectedly getting wet 2-3 times in 2 years in an excellent testament to the reliability of the forecasts. My rant was about trivialisation. Whether the Met Office or the BBC is responsible for this trivialisation, I don't know, but it's the Met Office product which is broadcast daily across the UK.

I'm 60 years old, I'm still riding motorbikes after 44 years, and I'm still mostly intact. I'm old, not bold and fully understand the notions of risk and possibility. I also understand the concept of making informed choices and taking personal responsibility. What I don't need is to be patronised by the enlightened ones, from the depths of their inner sanctum.

If the Met Office wants to be taken seriously, it should stick to what it's good at: providing weather forecasts to be presented in a non-trivial manner. As it chooses to meddle in politics, informing policy using projections which demonstrably diverge from reality, then its game on. Those policies have serious negative impacts on those of us outside the public sector/academic bubble. Those in the Met Office who dislike criticism should grow thicker skins or get out.

Mar 30, 2013 at 2:21 AM | Registered CommenterHector Pascal

So far no discussion here about probabilistic forecasting,how it is done and what the implications of its successes and failures means. Just fluff.

Mar 30, 2013 at 2:46 AM | Unregistered CommenterNick

Stephen Richards wrote:
"The BBC changed to their New Zealand bought software a few years ago amid the most complaints ever seen on the BBC."

I worked on the project to modify and deploy WeatherScape at the BBC. The software is perfectly capable of displaying synoptic charts, as well as a number of other traditional met charts. They can be static or animated over a time period. They can be for a limited area in (effectively 2D), or plotted in 3D over all the entire globe.

Whether WeatherScape does so in any particular broadcast is entirely up to the show's producer.

Mar 30, 2013 at 4:11 AM | Unregistered CommenterBruce Hoult

Expect the Met Office to go raving mad, because the gods are obviously out to destroy it. It takes skill or a curse to be so consistently wrong

Mar 30, 2013 at 6:44 AM | Registered Commenteromnologos

to clarify last night's comment : foi might have been needed because it sounds inappropriate for the bbc to write articles on the met office based on unconfirmed text from a blog in America.

Likely the Met Office refused to confirm the text's authenticity so Harrabin had to go the foi route, in another own goal by the MO.

Mar 30, 2013 at 6:49 AM | Registered Commenteromnologos

Ironic that the BBC puts FOI to good use whilst it stretches FOI law to breaking point to avoid FOI scrutiny of itself.

Mar 30, 2013 at 8:21 AM | Registered CommenterMartin A

Richard, until you gave your presentation at Oxford- and let us all remember it;
Then followed them up with " I stand by my comments of 4C or more global warming being possible by the end of this century, with local warming higher in some places, up to 15C in the Arctic as an extreme but plausible case"

Up to that point you had credibility as a scientist. With these comments you lost this, as far as I and quite a few other commentators here, believe.

As far as I am concerned such comments are as out of touch with reality as Lewandowsky. They have no attachment to observational and empirical science whatsoever.

Mar 30, 2013 at 10:54 AM | Unregistered CommenterDon Keiller


That's your scientific opinion and you are entitled to it - that's fine. I don't mind scientific criticism. If you can stick to that, instead of resorting to name-calling, I'd appreciate it.



Mar 30, 2013 at 11:01 AM | Registered CommenterRichard Betts

I think that the problem is, Richard, that as soon as the science gets challenging you feign insult, or pressure of work, and run away. This makes Don, and others, frustrated, and leads to a breakdown in communication.

Returning to the science, I would be grateful if you could answer a very basic question for me. How is it possible that models can lose fidelity (deviate from reality) within a few weeks of simulated elapsed time, then subsequently regain fidelity, enabling them to make meaningful predictions over decadal timescales?

This is a genuine question, and I would be grateful for an answer from you, or anybody else at the Met Office.

Mar 30, 2013 at 11:25 AM | Unregistered CommenterRoger Longstaff

Will also point out that this is a pure blog mechanistic thing - whenever Betts answers a question or two, a whole barrage of questions starts descending and he has no choice but to run. So he ends up picking the ones he can give short answers for. But I'll agree with Roger and say no more.

I followed up on Richard's message yesterday on Twitter about FOI. It seems whenever the BBC is involved in some climate-related FOI business, the information is already in the public domain. It is just wrong to say the MO has been hiding their forecasts when they haven't. At least Harrabin could have said, 'they've stopped publicizing them' or something like that.

Mar 30, 2013 at 11:39 AM | Registered Commentershub

plenty of name calling by the unnamed troll over the months but nobody from the met office stood up against that bully either

Mar 30, 2013 at 11:40 AM | Registered Commenteromnologos

Roger Longstaff +1

Richard Betts - please consider doing a guest post on the top 10 pieces of observational evidence that mankind is driving climate through CO2 emissions. That way you can take a considered approach and define terms etc to support your arguments. If you have some convincing material I think it will go a long way to taking the heat out of the cut and thrust of the comment/response dialogue. Much of what you see is frustration IMO. If you are not up to providing a few paragraphs of solid science then I respectfully suggest that you might be better staying out of the debate.

Mar 30, 2013 at 11:58 AM | Unregistered Commenternot banned yet

If your only support is a ban evading troll, that should give pause for thought.

Mar 30, 2013 at 12:13 PM | Unregistered CommenterNW

On Radio 5live yesterday morning they interviewed the person responsible for the offending forecast, who basically excused the bad forecast by saying it had to be "political": because of the existing drought they wanted to frighten people to make sure people took the right actions (sorry I can't recall exact details). Makes you wonder what other forecasts are similarly political...

Mar 30, 2013 at 2:37 PM | Unregistered CommenterRoger D

When the history of this period is written,the Met. Office's role in the mass hypnosis of the UK population will be compared with the same form of propaganda by the Nazis as they grabbed control of Germany.

Mar 30, 2013 at 3:34 PM | Unregistered CommenterAlecm

Richard Betts,

I will try to cut a deal with you. If you can answer the VERY simple question (copied below), to the satisfaction of the majority of BH readers, I will never question either you, or the Met Office, again.

How is it possible that models can lose fidelity (deviate from reality) within a few weeks of simulated elapsed time, then subsequently regain fidelity, enabling them to make meaningful predictions over decadal timescales?

Do you accept?

Please copy the question to Prof. Slingo.

Mar 30, 2013 at 4:24 PM | Unregistered CommenterRoger Longstaff

Roger Longstaff

I don't "feign" anything. If I say I don't have time, or feel insulted, then that's the truth. And in any case, there is no excuse for behaving in an offensive manner. For what it's worth, I don't approve of calling people "deniers" either. Name-calling, by either side, is completely pointless and unhelpful.

Nevertheless, thanks for the question, which is a very good one:

How is it possible that models can lose fidelity (deviate from reality) within a few weeks of simulated elapsed time, then subsequently regain fidelity, enabling them to make meaningful predictions over decadal timescales?

The point is, there are different things happening on short and long timescales. In the short term, seasonal forecasts are trying to predict natural internal variability, i.e.: variations around the mean state. It is extremely difficult to predict whether the next year will be warmer or colder than average. An for 50 years' time it is impossible to predict whether any particular year will be warmer or colder than the average, so the models are not regaining fidelity in that respect. However, it is possible to estimate how the average itself may change, if the system is subject to constant general warming or cooling influence - so for this, the models are doing a different task.

An analogy may help here. Imagine you're on a beach trying to predict how far up the beach the waves will go. For the next one or two waves, it will be very hard or even impossible to say whether the first wave will go further up the beach than the second - there are too many factors at play. Similarly, it will be impossible to say whether the 100th wave will go further up the beach than the 99th wave. However, if you know that the tide is coming in, you can say with some confidence that the 100th wave will go further up the beach than the 1st wave, as all waves will on average be breaking higher up the beach as the tide comes in.

So there's two things happening there - one is the factors affecting individual waves compared to the average, which is a short-term thing, and then there's the tide affecting the average itself, which is a long-term thing. So the height of individual waves breaking up the beach is analogous to the seasonal forecast (dominated by natural variability) while the change in the long-term average is analogous to greenhouse-forced (or solar-forced) climate change over decades.

Hope this helps. I'll be offline for the Easter holidays - have a good Easter everyone.

Mar 31, 2013 at 12:16 AM | Registered CommenterRichard Betts

Richard, good try, but your analogy is flawed; we can predict the tides quite accurately, and far in to the future, because their minimum and maximum are only dependent on two factors, the position of the Earth with respect to the Sun and Moon. CO2 on the other hand is one of many factors which affect the global climate system, and it is far from being the most significant. After all if CO2 was the most significant driver, global temperatures would have continued to rise for the last 15 years. But they haven't, and we have have been stuck at Mean Seal Level. Just how long do we have to wait before the incoming tide resumes? Another 15 years?

Mar 31, 2013 at 2:04 AM | Registered Commenterlapogus

Thanks for the reply Richard, and thanks lapogus for your comment - which I agree with.

The important point for me is " is possible to estimate how the average itself may change, if the system is subject to constant general warming or cooling influence - so for this, the models are doing a different task". I interpret this as imposing a warming or cooling influence upon a coupled, non-linear, chaotic system which the models can not simulate in a numerical, time-step integration, and in which low-pass filtering and variable "resetting" are necessary in order to retain "stability". In this case all that you are going to see is the constant warming or cooling signal that has been imposed upon what is essentially a random number generator that has been closely throttled to produce average results.

Is this correct? If so, how is this constant influence modelled? In particular, please could you, or someone in the Met Office explain the radiative physics model that is influenced by varying concentrations of carbon dioxide?

Mar 31, 2013 at 11:01 AM | Unregistered CommenterRoger Longstaff

However meaningful Richard B's reply contains several problems, eg:

1. Nobody experiences the "average" sea level. In climate speak, if you need a snowplough 3 years in a decade, you aren't going to use .33333 snowploughs a year for a decade. You are going to use ONE snowplough 3 years in a decade.

2. The changes in the climate relating to the "increasing average" aren't expected for several decades to come (SREX etc)

3. In-between today and the several decades ahead all sorts of things might happen seriously affecting the "increase"

These all combine to make the Met Office advice and importance wrt Climate Change pretty much irrelevant, and the Office itself an object of ridicule.

It's like forcing people inland by several kilometers because the sea level has increased by centimeters, a tide is expected many hours later and the tide's height might be either a few more centimeters, or meters, and nobody actually knows, and it might as well that the tide stops and reverses, and and and.

Actually, it's worse than that. People aren't being forced inland, they are squeezed out of their money and often left to die in the cold on the basis of what might happen to other people in 2040, or not, and as we know it will never happen to anybody because nobody experiences the average, and and and...

Mar 31, 2013 at 11:37 AM | Registered Commenteromnologos

Richard Betts' problem is that his models are predicting 40 foot Severn estuary-sized tides, when in relaity he's sitting on the shore of Lake Windemere and all he's got is a bunch of waves and no tide.

Mar 31, 2013 at 8:24 PM | Unregistered CommenterAngusPangus

So.... it seems the Met Office is now openly calling for "global governance"...

""The massive complexities associated with geoengineering, and the potential for winners and losers, means that some form of global governance is essential," said Jim Haywood at the Met Office's Hadley Centre in Exeter."

but wasn't this the whole purpose of our global warming scare in the first place and why our politicians have invested so much money in it.

And the reasons why our careerist politicians ignore the best interests of UK citizens in seeking to advance their own careers on the global stage.

Mar 31, 2013 at 11:17 PM | Unregistered CommenterMarion

Another thread expires without any answers......

I found a different climate modelling exercise:

Its' conclusions are:

1. Significant temperature drop at about 2016-17
2. Possible unusual cold snap 2021-22
3. Built in cooling trend until at least 2024
4. Temperature Hadsst3 moving average anomaly 2035 - 0.15
5. Temperature Hadsst3 moving average anomaly 2100 – 0.5
6. General Conclusion – by 2100 all the 20th century temperature rise will have been reversed,
7. By 2650 earth could possibly be back to the depths of the little ice age.
8. The effect of increasing CO2 emissions will be minor but beneficial - they may slightly ameliorate the forecast cooling and help maintain crop yields .
9.Warning !! There are some signs in the Livingston and Penn Solar data that a sudden drop to the Maunder Minimum Little Ice Age temperatures could be imminent – with a much more rapid and economically disruptive cooling than that forecast above which may turn out to be a best case scenario.

The conclusions are very different from Met Office predictions. Will they prove to be correct? I have no idea, but at least the methodology is more understandable.

Apr 3, 2013 at 10:18 AM | Unregistered CommenterRoger Longstaff

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):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>