Bishop Hill Met Office to scrap seasonal forecasts
Mar 5, 2010 Climate: MetOffice A couple of days ago, Mrs Hill commented that she could no longer find the Met Office's seasonal forecast. We had a bit of a dig around the Met Office website and there indeed seemed to be no mention of a spring forecast.
In the rush of activity after the hearings on Monday, I neglected to follow this up, but the Met Office has now come clean anyway. The seasonal forecasts are to be scrapped.
The Met Office is to abandon its long-term and seasonal forecasting after criticism for failing to predict extreme weather.
It was berated for not forseeing that the UK would suffer this cold winter or the last three wet summers in its seasonal forecasts. The forecasts, four times a year, will be replaced by monthly predictions.
It's hard to see this as a bad thing, given that the Met Office were becoming a laughing stock as a result of their wildly inaccurate predictions.
But wait - here's a thing. Remember back to those hearings on Monday. What was it that Professor Slingo said about climate models?
At least for the UK the codes that underpin our climate change projections are the same codes that we use to make our daily weather forecasts, so we test those codes twice a day for robustness.
Now, note that she is talking about daily weather forecast not the seasonal ones, so I think there is a reasonable case to make that the forecasts that have been scrapped may be based on a different model to the daily weather forecasts. But isn't this all a bit odd? They can forecast short-term and they can forecast long-term, but in the medium term it's all too hard? It doesn't really make sense, does it? And then again can it really be the case that the models for the short and long-term are different to the ones for the medium term?
I claim no particular expertise here; it just looks rather strange.
Reader Comments (52)
This is inspiring... the obvious needs to be tooned
Good
The problem for the Met Office is that the timescale for medium-term forecasts is short enough for people to remember what they said. For long-term forecasts, I suspect they are banking on people not remembering.
To a baby: short term you'll get bigger.
Long term, you'll die. Medium term: who knows?
There's nothing like a lousy analogy, eh?
Does anyone know whether there's any information available on independent auditing of the MO's daily forecasts?
No, but someone at the BBC was proposing a competition between the MO and Piers Corbyn. That would have been fun. Maybe the MO is running scared. ;-)
You can't believe what Slingo says any more - she has to speak 'on message'. She spent years showing that with current computing power, regional forecasting long term was impossible. That's what the article in 'Nature' also affirmed (Andy Russell's summary stated: "Regional climate prediction: We still don’t have sufficient computing power to run models at high enough resolution to make projections on the scale that would be useful to policy makers. This is clearly required to make big infrastructure decisions.").
Slingo and her peers consistently said that the present state of climate models and computing power made it impossible to be able to give useful predictive information to policymakers, especially at the regional and local level.
And then Slingo joined the Met Office: she soon changed her tune to spout government propaganda rather than reflect the state of the science. She said:
"Through UKCP09 the Met Office has provided the world’s most comprehensive regional climate projections with a unique assessment of the possible changes to our climate through the rest of this century. For the first time businesses and authorities have the tools to help them make risk-based decisions to adapt to the challenges of our changing climate."
After that, it's impossible to take the woman seriously. Fuller details here:
http://buythetruth.wordpress.com/2009/06/24/met-office-fraudcast/
I am really disappointed but not suprised. I loved telling people that if they wanted to look into the quality of predictions made by global circulation models, look at the consistancy of the MET office seasonal predictions. They always predict warmer and recently, they always predict wrong. So now they say they make good prediction out a few days and a few decades but that stuff in the middle is a bit dicey and really, really hard. If they are so good at this business, they ought to know what works and what doesn't so I see a failing in both the climatology and the climatologists. Its interesting that traditional weather forecasters who rely on the set up of the oceans and relate that to historical weather patterns seem to do a much better job at seasonal forecasts.
Short term: Largely, I suspect, a case of looking at the satellite photos to "see what's coming".
Medium term: Computer-assisted guesswork, but suffers from mockery after folks remember they got it wrong again.
Long term: Also computer-assisted guesswork, but has the great advantage that either we're all dead from old age, or nobody can remember what they said... Before they are proven wrong.
@pogo, exactly what I wanted to say...
What is the BBC going to do now for seasonal forecasts?
Use Weather Action?
But that is not based on the CO2 theory.
So, no, they can't do that!
If the models are different, then it says it all. If the science is settled, as the Met claim, then surely there can be no more than one single model. If that single model doesn't make predictions that coincide with observation, ie reality, then it's junk. If they need different models to work on different time ranges, then, a priori, the science isn't settled.
Over in the parallel universe occupied by old school physicists, there's a very successful model for the big things, GR. And there's another successful model for the tiny stuff, QED. When I say success I mean that these models describe with enormous accuracy what happens in reality, and are constantly tested to destruction by experiment and our experience of the modern technological world, much of which is possible precisely because these 2 theories fit reality so well.
But still, I never heard any of those old school chaps claim their science was settled, which explains the work going on for much of the last 100 years to find the better model that can describe both worlds consistently.
AusieDan, it looks like if it were up to Paul Hudson (a lone voice of reason at the Beeb), they'd be talking to AccuWeather. Good Yorkshire lad, calls it like he sees it, and he seems to see it all.
The Bishop said:
"And then again can it really be the case that the models for the short and long-term are different to the ones for the medium term?"
Why not ask them with FOI Act request.
On an different point:
In the Select Committee hearing, reference was often made by Jones and Slingo “...that computer programs where made available …”. Computer programs can be in two forms.
The first, is the computer code, made by the programmer in a computer language, such as Fortran, Pearle, Basic or C++ etc. A program in this form can be understood by persons expert in computer languages.
The second form, is a computer program compiled from an intelligible language (Fortran, Pearle, Basic or C++ ) into “Machine Code”. This second form, is the usual way computer programs are supplied to users and is more or less indecipherable. This form of program can be used, but the underlying code can not be examined in any meaningful way.
It is important that the above distinction is made.
Not exactly a competition but I did see Piers and that Beeb weather man who talks so slowly you forget the start of his sentence back in mid Jan. Piers prediction was spot on, predicting a lull, and then snow coming back on about 24th Feb IIRC. Beeboid obviously had nothing to argue back with as they have nothing but the 5day to work with, but tried to rubbish Piers claim
Surely, this is post-modernism in perfect action -- the more money a government service receives, the less it delivers. (See: NHS, schools, police, emergency services etc etc....)
You can't give up producing forecasts just because you get it wrong! 'Where would that end?'
The MET office daily forecasts are hardly reliable. We were forecast snow last week, never appeared. Comparing the MET office forecasts with uk.weather.com has been a constant source of amusement since the beginning of the year.
Oh well, maybe they do have a point, just give up forecasts altogether and concentrate on the core business - a full time environmental advocacy group like WWF or Greenpeace.
From the MO annual report
On Radio 2 earlier, they were spinning this as the public had said that monthly forecasts would be of more use to them than seasonal forecasts (nothing to do with the Met Office's last four seasonal forecasts being crap, of course).
Personally, I'd just prefer ACCURATE forecasts, whether they're daily, weekly, monthly or seasonal.
The Met are NEVER right with our local forecast.
Just after the item you quoted Stringer points out their inaccuracy (quote from uncorrected evidence)
Slingo:
"At least for the UK the codes that underpin our climate change projections are the same codes that we use to make our daily weather forecasts, so we test those codes twice a day for robustness.
Q210 Graham Stringer: You do not always get it right though, do you?
Professor Slingo: No, but that is not an error in the code; that is to do with the nature of the chaotic system that we are trying to forecast. Let us not confuse those. We test the code twice a day every day."
My conclusion, as a layman, is that if the system is so chaotic that it is not possible "to get it right" even on short term or seasonal forecasts, it is arrogant beyond belief to claim that climate can be forecast 100 years ahead.
Even though that may be, and almost certainly is, the situation they are still obligated to release their raw data, the adjustments made, the hypotheses formed and the computer programmes used to reach their conclusions. My understanding is that they have not yet done so.
Weather is chaotic, climate is not. Didn't you know? Climate 'science' is different.
This consistent argument by the Met is just not credible. They are part of the same chaotic system.
Actually here is a question. What happened to their 'Post Code' based climate change info. You type in the post code and it will give the projections for the next 30 or 40 or 50 years or so?
It always was silly, but isn't it even more so now? Time to revisit it?
A comment I saw on a blog a few weeks ago.
If - in the words of the warmists' favourite mantra - "weather is not climate", why do climate projections make use of data collected by weather stations?
The other day I heard the following comment, attributed to the RAF:
The problem with the Met office forecasts is that they are not even always wrong...
Honestly, I don't like that they've done this. Seasonal forecasting is obviously going to be hit or miss, but there is some skill in doing it. The only way to improve that skill is to make forecasts (with documented reasons why), see what happens, then revise your methods to improve upon those forecasts in the future.
The met office just needs to revise their forecast methods. I have no idea, other than a devout believe that things must warm, for them to have predicted a warm winter. Forecasts for the NAO index were largely negative which typically leads to colder weather in the UK and Eastern US. Just like what we've seen.
Jiminy Cricket
"Actually here is a question. What happened to their 'Post Code' based climate change info. You type in the post code and it will give the projections for the next 30 or 40 or 50 years or so?"
That's UKCP09 - it shows climate projection to the end of the century on a 25km grid.
http://ukclimateprojections.defra.gov.uk/
You need to be a registered user to access some of the services.
Here's my take on that:
http://buythetruth.wordpress.com/2009/06/24/met-office-fraudcast/
The met office themselves say that they use the same model for forecasting the weather that they use for predicting climate change:
"The flagship numerical model developed and used at the Met Office is called the Unified Model (UM), as, unlike most other NWP centres, different configurations of the same model are used for both weather and climate prediction."
http://www.metoffice.gov.uk/science/creating/daysahead/nwp/
Warmists are fond of parroting "the weather is not the climate" when someone questions how successive cold winters can be compatible with statements that the globe is in thermal runaway.
But if the same model is used for predicting both weather and climate, what can be the logical distinction between them?
Maybe it has dawned on the Met Office that, since we can quickly enjoy a good laugh at the inaccuracy of their short term predictions, their statements that their long term climate predictions are reliable also makes a very good joke.
I keep seeing the reference Slingo made to the code being tested twice a day for daily weather forecasts, but what does this have to do with climate change prediction?
We are told, repeatedly, that "weather is not climate", so logically a computer model designed to predict WEATHER cannot be used to predict CLIMATE.
Wow, that's amazing. A rare outbreak of common sense from the Met Office. They have finally acknowledged what many of us have been saying for years, that you can't predict the weather months in advance.
The announcement is on their website at
http://www.metoffice.gov.uk/corporate/pressoffice/2010/pr20100305b.html
"We have therefore decided to stop issuing a UK ‘seasonal forecast’ four times a year. "
It is interesting to see the language used above by the Met Office and their Global Warming apologists (the MSM). About six months ago, the standard Man In The Street warming argument was that "we never see the cold winters that I remember as a child" etc. Now that we've just had one of those "cold winters" "we never see anymore" it has become an "extreme weather event" - no doubt typical of what to expect due to "Climate Change".
I'm surprised the UK press & blogs haven't reported on this:
Professor Julia Slingo OBE, Chief Scientist, Met Office UK at the UK Parliamentary Hearing in answering a question as to why the land based temperature data record has shown higher temperature anomaly than the two satellite records explains that the reason is that satellite data is "an order of magnitude" (10 times) less accurate than the land based thermometer data [you know those thermometers in little latex-painted boxes at the end of aircraft runways and in cities]. She dismisses the urban heat effect by saying "we've looked at it" and that it was insignificant.
I contacted Dr. Roy Spencer about her claim and he said Slingo was wrong on both counts
@PaulM
We can though. They just aren't very accurate. Predictions can be made that are much better than random guess and better than average climatology for the period. Just depends on if you think 10-20% better than the climatological average is worthwhile to you. IMO, the only way to improve it is to keep forecasting and verifying those forecasts. This decison to me reaks of hurting true science so they don't get bad PR that could keep the UK government from implementing their green initiatives.
Note how weather forecasting and climate projections interact in the Met Office. Apparently, according to Julia Slingo, seeing whether the daily weather forecast was OK is a good enough check on the models used for climate projections for 100 years. What does the fact that the Met office can't get seasonal projections right tell us about the model?
She is the Met's Chief Scientist and she says they use the SAME model for weather and climate. So her 'testing the code twice a day' for climate prediction amounts to nothing more than running the Met's supercomputer twice a day to produce the daily weather forecast. If the daily weather forecast is sort of OK, then that means the model is OK for climate projections 100 years ahead. Anyway, she can hang herself in her own words:
"So what these models are is hundreds and thousands of lines of code which capture and represent our best understanding of how the climate system works. So they are not in a sense tuned to give the right answer, what they are representing is how weather, winds blow, rain forms and so forth, absolutely freely based on the fundamental laws of physics.
How do we know that they’re good? Well we continually test them against observations of the current climate in lots and lots of ways. At the Met Office we use the same model to make weather forecasts as we do to make our climate predictions, so every day we are testing the model and saying, ‘how well did we do with the weather forecast?’ We know that on many occasions our weather forecasts are incredibly skilful and that’s increasingly giving us confidence that the science in our models is fit to do this ‘crystal ball gazing’ into the future to say what will happen to our climate as we go really into uncharted territory. Because we are taking this planet to somewhere where it has never been before, or at least for millions of years."
Incredible.
http://www.metoffice.gov.uk/corporate/pressoffice/in-depth/ask/julia-slingo.pdf
Yes - extrapolation is a dangerous thing, as evidenced by the failure of the seasonal forecasts.
How do they measure the accuracy of their daily forecasts, versus reality? Locally? Globally? Just glance outside occasionally? Does anyone know how this can be done with any real meaning?
And if they claim to be measuring this daily accuracy and patting themselves on the backs for doing so well, are they also measuring the validity of their corresponding far future predictions by running their models on known -20 year data, for example, and seeing how well the super model is in relation to today's weather?
Surely, the science is settled single perfect climate model would do exactly that. And then just as surely, it could be used to make the perfect seasonal forecast. And the dailies, and all other forecasts on any time frame.
Scott B says
The only way to improve that skill is to make forecasts (with documented reasons why), see what happens, then revise your methods to improve upon those forecasts in the future.
I agree. But this would involve feeding back into their model to account for all the differences in their daily forecasting. As we all know warm climatologists prefer strongly positive feedbacks. This would inevitably produce runaway temperatures in the modelling supercomputer, blowing it up in the process, resulting in even bigger demands for tax funding for newer, more powerful supercooled supercomputers.
And so on, ad infinitum.
Anyone surprised that the further into the future the models go the less accurate they are?
Make no mistake, this is all about damage limitation.
One of the weakest parts of the Warmist agenda has always been their predictions: sea ice will disappear completely by 2012, London will be underwater in 100 years etc etc
These predictions are made because they are a, TERRIFYING and b, UNPROVABLE. The guilty will be long dead by the time they are proved wrong, but the benefits from making a prediction are gained now.
In the past these dreadful predictions have been made for suitably advanced point in the future. The Met office, unfortunately for them, has become so cocky (and so ingrained with the catastrophic global warming dogma) that they let their guard down. They genuinely didn't think it through that people might actually remember them making their incorrect predictions.
Oh dear.
As the Warmist case unravels it is becoming more and more apparent that their ability to predict is useless (see sea ice). The Met office seasonal predictions have become a laughing joke and are associating their 'gold standard' brand to ridicule. The Met office is such a keystone Warmist institution now it cannot be allowed to fail.
Hence the withdrawal of the seasonal forecasts.
It's quite obvious to me that they want to get rid of verifiably inaccurate forecasting, so that they can say "we can forecast future climate change because we are so good at forecasting the weather up to 1 month ahead" and hope that, over time, people will forget that they are total crap at forecasting 3 months ahead.
They don't want some old duffer telling Andrew Neil that they are very good at short term weather , and long term climate forecasting, but the bit in the middle is too complicated. (When Neil was a rottweiller before he got slapped into a poodle for Monbiot)
Drew : "As we all know warm climatologists prefer strongly positive feedbacks."
Especially Julia Slingo, who, in her Met Office Hadley Centre presentation to the Royal Society of Edinburgh last October, 'The current state of understanding climate change, its impacts and uncertainties', stated
"KEY MESSAGE: There are uncertainties in future projections but our current understanding of the Earth System suggests that most feedbacks will act to amplify global warming."
Check out the Climate-Alarmist-in-Chief: she is now a spokesman for government propaganda:
http://www.royalsoced.org.uk/enquiries/climate_change/talks_slides/j_slingo_slides.pdf
Agree with Stuck Record. Let's just home the next time they prophesy disaster, someone (other than the beebophants) say "but you're no good at long term forecasting, that's why you gave it up isn't it?"
I'd like to think Paxman would say it, but his bread is far too thickly buttered.
Sorry, "hope", not "home" above.
The Met Office seems to be are taking a bit of a gamble here at least with seasonal forecasts they can only be wrong four times a year.
If Professor Slingo did indeed tell the committee that the met Office use 'the same model' for weather as climate forecasts, then she is at variance with the Met Office's own statement (on their web page about seasonal forecasts:
"However, on even longer time scales, such as a century ahead, only physical models are used, as no more skilful statistical approach has been found. "
This (and the preamble which supports it) seems a very clear statement that this is NOT the same model at all. Someone should perhaps point out to Professor Slingo, and to the Hosue of Commons Committee, this discrepancy which seems to me rather an important one.
I liked these extracts from the original draft of the BBC story:
and
But you may not be able to recover those bits from the Google cache.
Foot, meet bullet.
Oh, you already know each other?
Jane Coles: "Staff previously engaged in seasonal forecasting will be reassigned to climate change outreach duties in schools." Ex-WWF eco-imperialist Robert Napier introduced climate change business to the Met Office when he joined them as Chairman. Redeploying weathermen as climate change propagandists is all part of the plan.
Chris Bore: "If Professor Slingo did indeed tell the committee that the met Office use 'the same model' for weather as climate forecasts, then she is at variance with the Met Office's own statement..." She did say that to the committee, and it wasn't a 'mis-speak' on her part because as I pointed out in a previous comment she says the same in a written document as well:
http://www.metoffice.gov.uk/corporate/pressoffice/in-depth/ask/julia-slingo.pdf
"At the Met Office we use the same model to make weather forecasts as we do to make our climate predictions, so every day we are testing the model and saying, ‘how well did we do with the weather forecast?’ We know that on many occasions our weather forecasts are incredibly skilful and that’s increasingly giving us confidence that the science in our models is fit to do this ‘crystal ball gazing’ into the future to say what will happen to our climate as we go really into uncharted territory."
That is unequivocal: the Chief Scientist says that they use the SAME MODEL for weather forecasting and climate projections, and thus the accuracy of the weather forecast is 'verification' of the robustness of the same model for climate projection. No verification or test of robustness could be applied if they were different models. Both in the extract above and to the parliamentary committee she made the point that the climate models are OK because they are tested on a daily basis doing daily weather forecasting (!!). Believe me, Slingo has spent years on the computing aspects of climate prediction, so if she's wrong on this one she shouldn't be in her job: she'd either be ignorant, deluded, or trying to hoodwink parliament and people.
Of course, what this shows without a shadow of a doubt is that since (if we can believe the Met's Chief Scientist) the same model is used, then the climate projections MUST be factored into the weather forecasting because the same model is serving both - the weather is simply the diurnal, seasonal and annual variation on the inexorable climate trend determined by natural variation and anthropogenic forcing. It is no wonder, therefore, since the actual climate has failed to march in step with the climate projection, that the Met's seasonal forecasts are consistently wrong: they MUST by now have a warm bias in them by including the climate trend (which is factored into the model as increasing, but which in reality is flatlining). This is a divergence problem between the presuppositional data fed into the model (based on AGW dogma) and the reality.
Since the seasonal forecasts HAVE consistently turned out to have had a warm bias in them compared to reality (which the Met office says it cannot explain - but I would have thought was pretty obvious following the logic above) then there is strong indication that the underlying bias, which is the supposed anthropogenic signal, is not there, or the model is wrong, or the natural variation has not been accounted for properly (or any combination of the above), any or all of which are fatal to the AGW hypothesis and climate projections.
You don't have to be an expert on climate to make that deduction and reach that conclusion: that drops out from logic based on the premises. Anyone, even a non-scientist, can shoot scientists down in flames when they indulge in fallacious reasoning because logic is common to all, and science doesn't have a monopoly on it.
The Met Office can no longer forecast the weather one season ahead. However their forecasts for a century and half a century ahead is still “robust”.
Their logic – it is easier to forecast a century ahead than just one season, which arrives too quickly and can quickly be verified.
Another example often given, we cannot tell if a particular day will be warm or cold, rainy or dry, but we can forecast the seasons very well, for example winter will be colder than summer.
They have discovered this by means of their complicated GCM’s and supercomputers, without which we would never have guessed.
Planet Earth to Met Office.
We want our money back.
Does anyone know if the "contract" between the Met and its source of funding defines the job it is paid for (is it weather forecasts or AGW advocacy?) and requires it to deliver value?
I'm just wondering if the public have any recourse to a refund given the Met has unilaterally decided to give up the job it's being paid to do for the reason that its "jolly difficult".
If this job was deemed worthy of the levels of funding the Met gets from tax payers, shouldn't the contract now be put out to competitors rather than simply allowing the Met to decide on how it will re-allocate those funds?
I feel certain that if the Met was a US institution, law suits would be drawing up as we write.
Many of the comments here are sadly very ill-informed. I am a meteorologist but not from the Met Office so I have no axe to grind. The short term forecasts are based on complex mathematical models of the atmosphere. The output when expressed in probabilistic format, shows a good level of accuracy out to five days (and often well beyond). Seasonal forecasting is usually based not on an atmospheric model but on ocean circulation and it affected (in the UK) by, for example, the north Atlantic Oscillation (NAO). These forecasts are also essentially probabilistic - the Met Office sadly made a rather stupid 'barbecue' comment that was ill-advised. Seasonal forecasts are better than tossing a coin but, if you had read the Met Office's footnotes (sadly now removed) were never claimed to be as good as short term forecasts. Global climate predictions are just that - Global. It's the local detail that's difficult which is why in the long term, there is more certainty. Textbook school physics also explains why greenhouse gases keep the earth warm (without them the world would average -20°C). That's a fact and it's difficult to see why increased concentrations of CO2 would not lead to a warmer world on average.
PJD
Global forecasts have done almost as badly as the seasonal ones.
Thanks You admin.
Arkadaslık sitesi
Arkadaslık sitesi