Click images for more details
A few sites I've stumbled across recently....
I along with others have posted questions on various threads directly to staff of the Met Office. As we all realise it is not always convenient for a reply to be forthcoming before the the question gets lost in the thread or newer posts take the attention away from the one where the question was posted.
Might it be possible to use this thread, with the Bishops consent, to log questions that could be answered at the convenience of those who may be able to respond in their own time. I would hope that those of the Met Office would also find this useful in order to pass on knowledge to us the general public.
I would suggest if a question is directed to a specific individual that others within the MO could just as easily provide an answer as it will most probably be the information required rather than a personal response.
If a specific question has already been answered elsewhere but has been missed by the individual asking it then anybody that could provide a link to the answer, it would be appreciated.
It may also be prudent that any answers that trigger a lengthy discussion be moved to a new thread so as to retain the viability of maintaining this thread.
Obviously if a question is posted in a specific thread it will be prudent to initially wait for a response on that thread before posting here but specific questions that are not pertinent to other threads will also be relevant.
I would hope that those of the Met Office who already respond on a regular basis, our thanks go out to you, might persuade some of their colleagues to also dip their toes into the blogosphere and share their knowledge with us.
An excellent idea, Lord B. I second the motion!
Sure - within reason of course!
May I request that Tamsin Edwards' comments policy (see end of her post) is a good one to follow (although point (e) is probably redundant in this context). This will be critical if Doug and I are to persuade other colleagues to engage!
And please don't overload us... :-)
Let's give it a go on a trial basis and see if it works.
This seems like a good idea so let me re-post an unthreaded comment I made earlier (deliberately provocative at the time in the hope of eliciting a reply) which followed a revelation by BBC bloggist Paul Huson that Met Office global forecasts had been too warm in 11 out of the last 12 yrs.
This is what I said:
The probability of the Met Office global temperature forecast having shown the same directional bias (in either direction) in as many as 11 years out of the last 12 (if we assume that the chances of an error in either direction is equally likely) is about 0.0063 i.e. this is a highly significant result and confirms the existence of bias.This is on a par with the Met Office's claim that the probability of the UK experiencing another winter as cold as those we have been experiencing recently is roughly 1/20 and is independent from one year to the next...They really don't have a clue and this is now being confirmed statistically.
This is on a par with the Met Office's claim that the probability of the UK experiencing another winter as cold as those we have been experiencing recently is roughly 1/20 and is independent from one year to the next...
They really don't have a clue and this is now being confirmed statistically.
Now I have no doubt that Richard or Tamsin will be able to put a better gloss on the value of Met Office forecasting generally but I would like to see a proper defence on the specific charge of bias.
Thanks for that. It's worth posting the whole of Paul's article here for context:
Global temperatures fell quite sharply in January, according to the UAH satellite measure. The anomaly of -0.093C below the 30 year running mean equates to approximately +0.16C above the more standard 1961-1990 time period.As regards 2011 as a whole, according to the Met Office, 2011 was the 12th warmest year in their 150 years of global temperature records with an anomaly of 0.346C. This compares to their 2011 forecast of 0.44C. Although this discrepancy is within the stated margin of error, it is the 11th year out of the last 12 when the Met Office global temperature forecast has been too warm. In all these years, the discrepancy between observed temperatures and the forecast are within the stated margin of error. But all the errors are on the warm side, with none of the forecasts that have been issued in the last 12 years ending up too cold. And, in my opinion, that makes the error significant. Some scientists who I have spoken to suggest that one of problems is the lack of observations in the Arctic, which is known to have warmed faster than other parts of the world. They point out that if proper account was taken of this area of the world, then the overall observed global temperature would be higher, a point acknowledged by the Met Office when I spoke to them earlier this week.In short, it could be that the observations are wrong, with computer predictions right all along. Climate sceptics, however, say that the real reason why the computer predictions are systematically too warm is because they don't properly take into account some of the natural processes that are occurring, such as weak solar activity, which may be holding back global temperatures.But in recent research conducted by the Met Office and Reading University, the possible cooling exerted by a less-active sun was found to have only a small effect on global temperatures. This year, the Met Office is predicting an anomaly which is 0.44C above the long term average.Whatever the reason for the ongoing 'warm bias' in Met Office global temperatures, their forecast for the first half of this decade, published in early 2010, that half the years between 2010 and 2015 would be hotter than the hottest year on record (with an anomaly of 0.52C set in 1998) is already looking in doubt.
The anomaly of -0.093C below the 30 year running mean equates to approximately +0.16C above the more standard 1961-1990 time period.
As regards 2011 as a whole, according to the Met Office, 2011 was the 12th warmest year in their 150 years of global temperature records with an anomaly of 0.346C.
This compares to their 2011 forecast of 0.44C.
Although this discrepancy is within the stated margin of error, it is the 11th year out of the last 12 when the Met Office global temperature forecast has been too warm.
In all these years, the discrepancy between observed temperatures and the forecast are within the stated margin of error.
But all the errors are on the warm side, with none of the forecasts that have been issued in the last 12 years ending up too cold.
And, in my opinion, that makes the error significant.
Some scientists who I have spoken to suggest that one of problems is the lack of observations in the Arctic, which is known to have warmed faster than other parts of the world.
They point out that if proper account was taken of this area of the world, then the overall observed global temperature would be higher, a point acknowledged by the Met Office when I spoke to them earlier this week.
In short, it could be that the observations are wrong, with computer predictions right all along.
Climate sceptics, however, say that the real reason why the computer predictions are systematically too warm is because they don't properly take into account some of the natural processes that are occurring, such as weak solar activity, which may be holding back global temperatures.
But in recent research conducted by the Met Office and Reading University, the possible cooling exerted by a less-active sun was found to have only a small effect on global temperatures.
This year, the Met Office is predicting an anomaly which is 0.44C above the long term average.
Whatever the reason for the ongoing 'warm bias' in Met Office global temperatures, their forecast for the first half of this decade, published in early 2010, that half the years between 2010 and 2015 would be hotter than the hottest year on record (with an anomaly of 0.52C set in 1998) is already looking in doubt.
First let's be clear about the word "bias" - this is meant in its scientific sense (the model systematically going one way more than the other) not in its political sense. I expect you know that, but others may not realise it.
It's completely true that the forecasting for the coming year is still in its infancy, and the forecasts of global mean temperature are essentially an interesting exercise in testing the model "out of sample" rather than being intended to support decision-making - so to you answer your question about "utility", we don't make any particular claims that these should be useful for anything other than seeing how the model is performing.
As for the actual results - I'll have to consult a colleague on that. As Paul pointed out, the lack of observations in the Arctic may be an issue - but I'll get back to you in due course.
Hi again matthu
Actually I found the answer on the Met Office website
It is the Arctic that's the issue. The model forecasts covers the Arctic, but the observations do not, so it's not comparing like with like.
If the forecast is compared with observations for the region where there are adequate observations (60N to 60S) then the match is much better (see second panel in the page linked above).
Richard - thank for the reply.
I think you are trying to say that the reported GAT is lower than it 'really' is because we do not have enough measurements to cover the Arctic (where you assume that the GAT anomaly is higher than normal). But presumably the calculation of GAT has a weighting for the Arctic which takes care of that?
Also, the historical data available to you must have equally sparse data covering the Arctic, so I would have thought that your model (which has presumably been calibrated using the history you have available) should be trying to project the measurements you can expect to make in the future?
Otherwise what you are saying is that you are making forecasts (maybe you would call them projections?) which can never be falsified because you don't attempt to measure reality.
Also, perhaps you would also comment on the other advice I got from the Met Office that the cold winter events we had been experiencing were a 1/20 event and no more or less likely to occur in any future year regardless of what we had experienced recently.
I find this particularly odd.
Are the Met Office revising their stance on this or are they still saying that we have a better than even chance that we will NOT experience another similarly cold winter in the next dozen years (because the chance of not experiencing such a cold winter is 95% and so the chance of not doing so for 12 years on the trot is 0.95 to the power of 12 = 54%) ?
A couple of "presumablies" in there - always dangerous as it risks creating a straw-man.
Again, if you check our website, you'll see that for HadCRUT3 only grid squares where observations were made in a month are used to calculate global average temperature. Other groups do more what you suggest, and do indeed get a warmer GAT due to including the Arctic!
I think I'll have to ask that folks check the Met Office website for answers before posting questions here. I have limited time and it's better spent addressing questions for which the answer is not already out there! I'll probably only be able to deal with 1 question per day at most, sometimes not even that, although I will do my best. (And also I reserve the right not to answer - although in such a case I will try to tell you that I'm not going to answer, and explain why!)
Having said that, thanks for your original question here as I wasn't aware of Paul Hudson's post, which did need addressing. (Shame he didn't read the website either!!)
Also I should point out two other things:
1. While I'll do my best to represent Met Office science correctly, any opinions are my own.
2. Tamsin Edwards does not work for the Met Office so please don't expect her to answer (although she is very welcome to comment if she wants to!!)
On your second question, please can you give more details on where and how you got that response, so I can check the context? Again, I'll respond in due course, which may not be today.
Thank you again for your reply - and you are right: I must try to familiarise myself more with what is available on your web site!
The original response I got from the Met Office was a reply to a personal enquiry but I have located an original source:
12.12 We have also explored whether or not the occurrence of two successive severe winters influences the probability of a third in succession – in other words, is there any evidence of clustering? There is some small influence from year to year but these matters are still very uncertain and it would be safer to assume that there is statistical independence between one winter and the next.12.13 In other words, we are advised to assume that the chance of a severe winter in 2010–11 is no greater (or less) than the current general probability of 1 in 20.
And of course - I do realise that Tamsin does not work for the Met Office, but bearing in mind the name of her new blog (what was it again? SomeModelsAreLessUselessThanOthers.com? oh no, http://allmodelsarewrong.com/ ) I thought she might have a particularly good angle on this one! ;-)
Thanks. I see that report is a couple of years old now.
It's early days in understanding this, but the solar cycle may play a part - see here and the paper that news item refers to here (if you can't get the paper, send me an email (or ask BH to be the middle-man) and I'll pass it on.
To take a risky strategy of cherry-picking a sentence from my colleagues' paper, they say:
The average of recent winters (2008/9, 2009/10 and 2010/11) shows cold conditions over northern Europe and the United States and mild conditions over Canada and the Mediterranean associated with anomalously low and even record low values of the NAO. This period also had easterly anomalies in the lower stratosphere. Given our modelling result, these cold winters were probably exacerbated by the recent prolonged and anomalously low solar minimum
However, as I say, it's early days - lots of interesting research still to do.... :-)
Richard's "arctic excuse" does not make much sense as pointed out by greensand on the Hudson blog. What is an 'observational estimate'?
I'll add that the MetOffice page is badly misleading because it doesnt use an equal area projection in Fig 3, so it greatly exaggerates the problem of low arctic coverage. This trick is also used by Hansen, exaggerating the appearance of arctic warming in all his graphs.
While the false claims on the Met Office web site pointed out previously remain there, I am going to be very suspicious of anything on that site.
Thanks for giving this a try. It's not intended to increase your work load although I do recall you saying once that this sort of interaction was encouraged by the MO and I do apologise to Tamsin as I was under the impression that she was part of the crew.
Anything that helps the passage of knowledge, let me know.
As far as comments policy, I think the Bishop is on the ball in that area which is why most of us visit so regularly. If anything does cause concern to anyone participating I hope that I could offer a temperate opinion if necessary and if posts get as excited as have done on other discussion topics, then I will ask to close the thread.
I and others have asked on various posts about the accuracy of the forecasts presented to us. To try and get to grips with the information that we receive perhaps someone could direct me to any available figures that define the accuracy of the UKV output for precipitation and temperature for next day and multi day forecasts. A search of the UKMO site and the WWW provide frustration,
There must be some form of validation utilised for Ministers to comment on the accuracy of forecasting! Are these figures not available to Joe Public?
One certainly detects a changing slant at the Met Office (MO). Even letting the Bish through the door is a very welcome sign - thank you to those responsible - but will they be listening? I was told a few months ago that it was a disciplinary offence for MO staff to express anything other than a solid AGW view (CO2 is evil that is). Is that true Richard Betts? If true (and I have no idea if it is or not) that is hardly the stance of a scientific organisation. GroupThink is never good in science and imposed GroupThink is much worse. I hope you can tell me it is not true.
Both matthu and Richard Betts quote one of Paul Hudson's blog articles - the 11 out of 12 years etc.
Discussions about measuring against real data (not something that all modellers like) and searching for excuses as to why you got it wrong, are a blind alley - what matters is what advice (forecast) flows from that and we all know that recent years have not been kind to the MO seasonally. It is obvious that there has been a warm bias - whether that is model or AGW belief doesn't actually matter.
As matthu points out - in effect cold events are seen, by the MO, as part of the variability (weather is not climate - well only when it suits), but 4 cold northern hemisphere winters in a row, with all time records being broken in a number of places these last two winters, does not fit the milder winter scenario postulated by the MO and the IPCC 2007 report. Intense cold in parts of South America this last two winters doesn't fit the model warming and please don't try to say it is due to CO2.
Another pertinent article by Paul Hudson was this one back in October last year
His first paragraph is interesting - " For as long as I have been a meteorologist, the mere suggestion that solar activity could influence climate patterns has been greeted with near derision. " -- The influence of Solar variation is surely becoming more evident everyday? The MO has moved from there is no influence to there might be some influence, but it isn't much....................... Just a quick look at the ice core records for the last few thousand years shows a huge blind spot in this regard.
I have said to you Richard in a comment before that models are only useful if they display skill and climate models have displayed very little so far in 25 plus years. I accept your view Richard that it is early days for these models - so why are we being shouted at (not by yourself, for which I thank you) and told the science is settled. The MO has been in the vanguard of the AGW cause and its position has bolstered policy decisions (world-wide) that IMHO will come back to haunt and have already cost lives.
Staring into models while wearing blinkers to data in the real world is no way forward. Is anybody in the MO pointing out that isn't going to plan, or is it an Emperor's garments situation.
AGW scientists (not the rabid activist believers) should engage with scientists on the sceptical side (not the big-oil couldn't give a damn people who I am sure exist) who are increasing in number and eminence all the time - but it has to be genuine and free from the behaviour demonstrated against sceptic scientists for far too long.
Hi Lord Beaverbrook
Very quick holding response (I am overdue on a deadline for returning proofs of a book chapter!) but does this help at all?
I was told a few months ago that it was a disciplinary offence for MO staff to express anything other than a solid AGW view (CO2 is evil that is). Is that true Richard Betts?
Richard if you get free sometime could you please explain why the MO does not update their Decadal Forecasts/Predictions?
Yes, thank you.
Will have a good read through but I didn't realise that Nov 2009 was as absorbing at the MO as for the rest of us by the look of the graph ;-)
At last, thank you. I will need to take some time to study this. Is there any way of converting these indices into percentage accuracy?
At first glance there is nothing about wind speed and direction.
In regard to the temperature forecast accuracy graph I could probably eyeball the figures but is there a data file available to import into excel and ensure accuracy?
Probably best to ask email@example.com about that (it's not my department!) Feel free to mention that we have discussed it here and that I referred you to them.
Hi Green Sand
I'll ask about that, it's a very fair question. The answer is probably that, since it is an emerging research area, the "official" forecasts are only those published in scientific journals (eg: the original Smith et al 2007 paper in Science) but I will my colleagues ask about it.
Thanks for the reply, I look forward to your colleagues comments.
Notify me of follow-up comments via email.