Skip to content

Did The Met Office Forecast The Beast In January?

March 18, 2018

By Paul Homewood

 

The Met Office seem to have been stung by media criticisms of their long range warnings about the Beast from the East at the end of February.

image

Various media reports have been commenting on our longer-range warnings in the run-up to the recent cold snap.

This period of severe weather was very well predicted and the first signs appeared around one month before the start, when we were able to offer broader advice about the likelihood of a cold signal. Our advice to government and the public ramped up in confidence and detail starting from the early signs in late January as events became clearer in our forecasts:

  • 26 January: The first indications of a possible cold spell were given in our one-to-three month outlook for contingency planners. On 26 January we said:
    ‘For February, below-average temperatures are more likely than above-average temperatures. The likelihood of impacts from cold weather during February is greater than normal.’

 

 

[Note: that this public outlook is always updated a week later (2nd February), leaving only the three-month view. The one-month outlook reverts to the 30-day forecast at this time.]

So, in summary, the severe cold snap of late February and early March was very well predicted, even from long-range on this occasion. The Met Office provided clear and regular updates on the increasing levels of risk from late January onwards to ensure everyone was aware of how the weather would impact them and they could be prepared for it.

 

https://blog.metoffice.gov.uk/2018/03/07/a-review-of-our-long-range-outlook-for-the-recent-cold-snap/

The claim that they saw it coming in late January is based on the 3-Month Outlook, published on 26th January, which stated:

For February, below-average temperatures are more likely than above-average temperatures. The likelihood of impacts from cold weather during February is greater than normal

However, quite shamefully the above response does not show the whole of that 3-Month Outlook, which gives a totally different version of events. Fortunately, although the Met Office don’t archive the old reports for public use, I did keep a copy at the time:

 

image

 forecast-temp-fma-v2-1

 

My version no longer includes the reference to February alone, in the first two sentences, but the Met Office explain that this is because it is updated about a week later, which appears reasonable.

But the important thing to note is the penultimate paragraph, enlarged below:

 

image

It cannot be any clearer. Not only did they not forecast the extreme cold spell at the end of the month, or the SSW event that led to it, they actually forecast the opposite, a milder spell up to mid March.

Moreover, as well as failing to forecast the Beast, or even a coldish spell of weather, they also failed to predict the cold and snow we have now had twice this month since the Beast.

 

https://www.metoffice.gov.uk/hadobs/hadcet/graphs/HadCET_act_graphEX.gif

https://www.metoffice.gov.uk/hadobs/hadcet/cet_info_mean.html

 

So what was the cold weather that they forecast in February?

Again, we can take a closer look at that 3-Month Outlook:

 

image

image

 

It was in fact the cold spell in early February, which was already well forecast, and did turn up.

 

Courtesy of the Wayback Machine, we can see what the Met Office’s 30 Day Outlook was saying in Jan 26th:

 

image

https://web.archive.org/web/20180126220822/https://www.metoffice.gov.uk/public/weather/forecast

 

A large degree of uncertainty! Some spells of mild, wet and windy weather. Colder, brighter and drier interludes.

 

The Met Office’s claim that the severe cold snap of late February and early March was very well predicted, even from long-range on this occasion, simply does not stand up to scrutiny.

 

It is true that we began to get some indications from them in early February. For instance they offer this link to the Daily Star on 6th February:

Alex Burkill, forecaster for the Met Office, agrees that wintry conditions are expected to continue into the weekend and perhaps into March.

“There are indications of quite a prolonged cold period,” he said.

“Really much of February and perhaps even into March it is going to stay on the cold side, so temperatures generally below average, with further frosts and also the risk of rain, sleet and snow as well.”

https://www.dailystar.co.uk/news/latest-news/679971/uk-weather-forecast-met-office-bbc-london-snow-cold-snap-freeze-news-latest

 

But by then, this was common knowledge amongst meteorologists. The same Star report quotes James Madden, forecaster for Exacta Weather, as saying the oncoming assault of heavy snow and freezing temperatures could plague Britain for the rest of February.

He said: “A barrage of snow events awaits us throughout the rest of this month with the first notable snow events expected next week.

“There is a risk for a total whiteout across the country as a number of major snow events hit our shores, this risk really kicks in at the start of next week.”

It is ironic that, when their 3-Month Outlooks are wrong, they claim they are not “forecasts” at all but mere ranges of probabilities.

Yet when they actually get one right, they crow about it!

But to claim that they predicted the Beast from the East, when they actually forecast the opposite, and then attempt to cover up the inconvenient paragraph that proves they were wrong, shows a breathtaking level of dishonesty, even by their standards.

Advertisements
23 Comments
  1. malcolm bell permalink
    March 18, 2018 12:19 pm

    — and these same people are destroying our energy security on the basis of their “proven” climate models for the next fifty years.

    They are not even close for three month’s time.

    • Jack Broughton permalink
      March 18, 2018 3:22 pm

      Yes, normally the accuracy of a forecast falls later and is accurate early, but not in the case of the “proven science”. When a variable such as temperature is integrated over 30 years, as in climate science, and the initial conditions are fuzzy, the outcome is even more fuzzy!

      • malcolm bell permalink
        March 20, 2018 10:27 am

        Exactly that Jack.
        I have used fuzzy logic as one part of decision making in AI. One lesson you learn very quickly when you do this is that if you get exactly the same answer twice on subsequent occasions then it is at least wrong once because the start conditions cannot have been the same except by vanishingly small probability.
        Weather modellers don’t seem to understand that, theirbmodels are not intelligent, merely algorithmic – which does demand the same answer wvery time.

  2. Bloke down the pub permalink
    March 18, 2018 12:39 pm

    It appears to be a shotgun attempt at forecasting. If they put out enough forecasts, all slightly different, then sooner or later one of the must be right.

    • TinyCO2 permalink
      March 18, 2018 1:16 pm

      I think that’s the basis of it.

  3. Richard Woollaston permalink
    March 18, 2018 12:41 pm

    Is anyone reminded of Astrology?

  4. NeilC permalink
    March 18, 2018 1:50 pm

    Never mind the 3 moth forecasts. They didn’t forecast of the 6″ of snow we had overnight nor the sub zero temperatures we have had since 1550 yesterday afternoon, 3 DAYS ago.

  5. March 18, 2018 2:02 pm

    I’m still awaiting a response from my MP to the second of the two complaints I made to him about the lies told by Prof Adam Scaife of the Met Office and when Scaife alerted the Cabinet Office about the “incoming weather bomb”. My MP’s first response completely ignored the complaint.

    • Tony McKenna permalink
      March 18, 2018 5:57 pm

      I would like to see a dated order /invoice for the extra oil he ordered.

  6. RAH permalink
    March 18, 2018 2:33 pm

    Well if Joe Bastardi and the gang at weatherbell.com is correct “The Beast From the East” will make an encore appearance pretty soon. What does the MET say about that NOW?

  7. John189 permalink
    March 18, 2018 4:40 pm

    In the Met Office 26 January forecasr highlighted by Paul above is the following statement:

    “For the February-March-April period…the chances of above average temperatures are slightly greater than the chances of below average. (This) is consistent with the observed increase in UK temperatures compared to the longterm average”

    Do we therefore expect an April heatwave of unprecedented longevity to redress the balance?

  8. don penman permalink
    March 18, 2018 8:23 pm

    It has been a cold winter as I have measured it in Lincoln :
    December ext. max temp. 11.3c ext. min temp. -3.2c int. max temp 13.1c int. min.temp. 4.6c
    January ext. max. temp 12.6c ext. min temp -0.4c int. max. temp. 12.3c int. min. temp 6.7c
    February ext. max. temp. 9.1c ext. min. temp. -4.5 int. min. temp.11.7c int. min. temp. 6.0c
    I remember that it was being forecast that it would be cold early in the winter and then warm up later because of the La Nina but this never happened. I think we have been cold because of the upcoming Solar minimum and will get even colder in the next few years. I am not saying that this winter has been unprecedented here in Lincoln but only the sixties were worse even the snow in the eighties did not compare to this winter and I remember both but I may just be getting older.

  9. john cooknell permalink
    March 18, 2018 9:23 pm

    The MET office longer term forecasts cover every eventuality so they cannot be wrong, but they are entirely useless.

    I forecast June in UK will be warmer than January, there you are 100% correct even before it happens! it appears I can tell the future as well!

    • John189 permalink
      March 18, 2018 11:45 pm

      I would go even further John. For June there is an early signal for periods of warm weather especially in the south-east whilst there will also be periods of cooler, cloudier conditions with showers more especially in the north-west. Rainfall is expected to be mainly slightly below the longterm average but will be above average in some places.

      This is the kind of risible drivel that is put out for public consumption.

      • duker permalink
        March 20, 2018 3:06 am

        Its done by their climate models, which really do come up with 101 different results- just like the dalmations.
        20% of the model output is in this range, 20% in that, 50% are same old same old and the rest is so far away from that we darent mention it.
        The meteorologists skill used to be using historical knowledge to tie it all together, thats all gone now its always 20% are above normal, 20% are …..

  10. Paddy permalink
    March 19, 2018 7:28 am

    Why anybody pays any attention to these forecasts when one can prove to oneself that the Met Office cannot forecast five days ahead by taking screen shots of the five day pressure charts.

    • Athelstan permalink
      March 19, 2018 9:21 am

      During, Antcyclonic conditions, the Wet office will prognosticate for a few days in advance and the pressure charts feature well may extend to four or five days ahead.

      If on the other hand computer ‘say um I just can’t say’, a 60 forecast is about the best you can expect and be prepared for it to change – sometimes with great variance.

  11. Athelstan permalink
    March 19, 2018 9:22 am

    60 hrs ± – I meant to convey – sorry Paddy.

  12. Coeur de Lion permalink
    March 19, 2018 10:04 am

    Have you looked at the page on climate change effects on the Met Office website? Pack of lies.

    • Athelstan permalink
      March 19, 2018 10:47 am

      I have done but it was a few years ago, the story telling etched an impression, drawing on that, to muse, it was written by a poorly educated undergrad’ who’s first language wasn’t English.

  13. Vernon E permalink
    March 19, 2018 3:57 pm

    There is a lovely weather lore verse which I can’t quote exactly which says that if the sun shines on Candlemas day (2 February) the winter will return. It did. Must be the old agrarians saying don’t rush to plant too early. Beats the heck out of super-computers.

  14. Joe Bastardi permalink
    March 21, 2018 12:35 pm

    This is a classic. Saying something is likely is not forecasting it, it is playing a probability game that leaves yourself a way out, There is a vast difference also between making an actual forecast and suggesting guidance for colder ( for example, if they said x below normal, we could see exactly what they were forecasting,) Let me explain how this non forecast game is played. Suppose you have 3 options above normal below normal and normal You can say there is a 40% chance of below normal which would be greater than 1 in 3 and then claim if it hits you were right. Actually you are only right on being on the right side of the probability, but in that case since the actual chance of an occurrence is either 0 or 100 at time of verification one can argue you were wrong. I played this game in college forecasting classes. In the case here,. First of all they dont explain what each range is or what they are using, Is it standard Deviation, Is the range in London the same as the range in Liverpool? Of course not, There are wider ranges that qualify in different places ( lets say you live near the coast and not inland,) so right off the bat its a non forecast. In any case if you have only 2 categories and you say its likely to be one rather than another, so what? How much more likely? How great. Is it.1 below normal, 1 below normal, 10 below normal. The fact is that when you actually score a metric ( the actual temperature) as clients I have want, you can see very clearly what is right or wrong, Simply saying that it was more likely to be cold than warm, and then when one of the most severe Februaries since 1962 shows up and then saying you hit it is a stretch, You never made a forecast that can be verified because the actual temperature is the metric. Now if you came out and said temperatures would be a definable metric ( example 3 below normal ) and said snowfall would be x above normal then THAT IS A FORECAST It does not rain or snow probabilities nor do thermometers register probabilities The UKMET and other agencies using probabilities should not claim them to be forecasts. IT IS GUIDANCE! And it may be great guidance. But this is blurring the lines and not real world . The real world is defining measurable metrics and then seeing how close you are. I dare anyone out there to read a probability on a thermometer. Again if you say the guidance hinted at it being colder than normal, I have no problem with that. Thats fine and I think accurate. But saying you made a forecast when no forecast was made since I did not see a forecast for temperatures that severe, just a statement it should be colder than normal PROBABILITIES ARE NOT FORECASTS, THEY ARE MEANS TO AN END. IS IT GOING TO SNOW? HOW FAR WILL TEMPERATURES BE BELOW NORMAL. Those are forecast questions. ( BTW The UKMET is a great office , I just disagree with them on this issue , and other offices that use this,) Put the actual temperature forecast out and see the result. Then you can say what was good, bad and as people that do this for a living know, ugly. Peace out

Comments are closed.

%d bloggers like this: