It’s official! This morning’s blood-red sun was caused by Saharan dust, so says Simon King the BBC weather presenter.
With the 30th anniversary of the ‘great’ storm of October 1987 now just a few days away, I thought it was about time that I looked into the rumours that ex-hurricane Floyd had something to do with it. After looking at some surface pressure charts for the Atlantic from the 12th to the 15th , I can now see how this speculation came about, and also better understand that Floyd was the hurricane that Michael Fish was referring to in his now infamous forecast when he said:
“Earlier on today, apparently, a woman rang the BBC and said she heard there was a hurricane on the way… well, if you’re watching, don’t worry, there isn’t!”
In the prologue to the book ‘Storm Force‘ which he co-authored, Michael Fish does explain that he was talking about hurricane Floyd and:
“…my remark had NOTHING to do with the storm”
He goes on to add that the infamous forecast:
“…was NOT made the evening before and NO woman rang the BBC”
Looking at the video on YouTube of the now infamous forecast, it’s not entirely clear to me at what time it was broadcast. It could have been from late the evening before (14th), but it seems more likely it was the broadcast after the midday news on the 15th, because rolling 24 hour news from the BBC started much later than 1987 as far as I remember. Michael Fish reminds us in his prologue is at pains to explain that it was Bill Giles who was the duty forecaster that evening (15th), and it was he who uttered the immortal line:
“…it will be a bit breezy up the channel”
As Michael Fish reminds us:
He kept quiet about this until he had collected his OBE and retired!
The culprit as we know was not any of the messengers, but a combination of poor guidance from Bracknell, poor forecast from what now looks a rather crude and unsophisticated NWP model, which itself was due in no small part to a dearth of ship observations from Biscay during the afternoon. Anyway I digress, as I so often do these days, so back to the main point of this blog.
Is there any truth in that speculation, that the October storm was the remains of ex-hurricane Floyd?
Floyd had just slipped past the southern tip of Florida as a hurricane on the 13th of October 1987, here’s the reanalysis chart for 00 UTC (fig 3).
The chart below (fig 4) shows the last position listed in the Hurdat2 archive for ex-hurricane Floyd, and this is how Wikipedia described the demise of Floyd.
Unexpectedly the storm turned sharply northeastward into the southeastern Gulf of Mexico. Based on reports from the Hurricane Hunters, Floyd briefly attained hurricane status on October 12. Around the same time, the nearby cold front spawned a low pressure area that cut off the hurricane’s inflow. While moving through the Florida Keys, Floyd became the only hurricane to affect the United States that year. However, its convection was rapidly decreasing over the center due to the front, and shortly thereafter Floyd weakened to tropical storm status. The circulation became nearly impossible to track on satellite imagery, although surface observations indicated it passed just south of Miami, Florida. The storm underwent extratropical transition as it weakened over the Bahamas, and Floyd was no longer a tropical cyclone by late on October 18*. The circulation dissipated within the cold front early the next day.
And below (fig 4) is the position of the embryonic low that became the ‘great storm’ at 00 UTC on Thursday the 15th of October. I’ve measured the approximate distance between the last observed position of ex-hurricane Floyd and the first of the October storm, and its a whopping 2,600 nautical miles, and even if it had been doing over a hundred knots as a surface feature, Floyd just couldn’t have made it. Perhaps at the middle and upper levels if there had been a jet that spanned the Atlantic from the Bahamas to western France which was blowing at 240° and 100 knots, tropical air from Floyd could have been pulled eastward to fuel developments in mid Atlantic. I’m no expert, but I suppose it’s possible that tropical air somehow got entrained into a developing extratropical low, Floyd did rather mysteriously turn to the northeast and lose intensity as the Wikipedia article points out, but I would have to download more reanalysis wind data for 500 hPa and above to check if that could have been the cause, there’s nothing quite like a good conspiracy theory is there?
As we now know the rest is history (fig 5) even though the coarse grid of the NCEP reanalysis doesn’t quite get the intensity of the October storm.
*This slip may be intentional or not, but I reckon that 18 should be a 14 according to the NHC record.
When I was a child I use to buy a weekly comic called the Beano out of my pocket-money. One of my favourite cartoons in it was about a character called Roger the Dodger (fig 1). Wikipedia nicely sums up much better that I can what Roger was about if you are not of that age:
His strip consists solely of Roger’s basic remit to avoid doing chores and homework which usually involves him concocting complex and ultimately disastrous plans, the undoing of which results in him being punished (usually by his long-suffering father). To perform these tasks he enlists the help of his many ‘dodge books’.
Weather Presenters Dodge #1
I sometimes wonder if there is a book of weather presenter’s dodges and what some of them are? A recent Twitter posting from the BBC weather team has led me to believe that I’ve found a dodge worthy of inclusion in their book of dodges. In the next few days, its forecast that the two ex-hurricanes Maria and Lee will merge in mid-Atlantic, and from the remnants of the two, a super low will form and threaten the British Isles with what could be storm Brian. As soon as I think of them actually naming this storm Brian I can’t help but smile, because it immediately brings back memories of Brian Conley’s alter-ego stuntman ‘Dangerous Brian’, anyway I digress. So the thing is, the Met Office ensemble models are split as to exactly which track this low will take and how much it will intensify by, and even if Brian will even happen at all. That’s where dodge #1 is so clever because the gist of it is this – tell the viewer that there are two possibilities and let them decide. Brilliant! It gets you off the hook, because you get it right whatever happens, and this is exactly the dodge that Matt Taylor employed yesterday in his weather for the week ahead video (fig 2).
I can’t say what the Met Office forecast for later on Sunday looks like, either from their ensemble or deterministic models, because as you know the Met Office like to keep us in the dark about what our model is saying. The GFS model on the other hand has backed off a forecast with an intense low as deep as 946 hPa for 12 UTC on Monday. It now brings the remains of two hurricanes rather quickly eastward as a discrete low at around latitude 50° north, before it’s absorbed into a trough of the dominant parent low south of Iceland during Sunday morning (fig 3).
I find it almost impossible to access a BBC weather forecast from the previous day on the BBC iPlayer. If I wasn’t such a trusting person, I would say that it’s to prevent people like me reminding them of a poor forecast. It’s similar in a way to the reluctance of the Met Office to archive the warnings that they issue, presumably this is to prevent any possible future litigation against them. It seems that the BBC presenters accepted the minimum temperatures from the model for Sunday night, even after it got things wrong on Saturday night for the very same reason, too much wind and cloud, particularly in eastern regions. I wonder if in the state-of-the-art graphics engine that MeteoGroup are poised to introduce will allow the presenter to edit these values?
The forecast temperature extremes across the UK are quite important to the Met Office, they are used at a selected number of places as a metric to calculate their annual bonus. If I remember correctly, they get full marks in verification if the forecast is within +/- 2°C of the actual value. Presumably, they will now use the temperatures that they publicise in their National video forecast, rather than the one that will eventually be provided by MeteoGroup for the BBC, from whatever model they choose to use.
Its been over a year now since MeteoGroup won the bidding process to take over the running of the Weather forecast on BBC radio and television. In August 2016, Nigel Charters, a project director at the BBC promised that these services would ‘hit your screens, from mobile to television, in Spring next year’ (fig 1). Well, Spring has gone, and so has summer, and now we’re in ‘meteorological’ Autumn, but we are still stuck with the same old BBC graphics, and not the state-of-the-art graphics system as promised by MeteoGroup, and as far as I can see we are still using forecast model data supplied by the Met Office.
So what’s gone wrong?
It’s impossible to say with any certainty, because the BBC and MeteoGroup are both keeping schtum about:-
- What the new graphics will look like.
- When the new graphics will start.
- Who exactly will provide the ‘multiple sources of meteorological data’ (see (2) in fig 1).
If you take a look at the MeteoGroup website they have remained tight-lipped about the whole subject since the news that they had been awarded the contract (fig 2).
MeteoGroup seem keen to get started with their own graphics system, which I understand is not an in-house solution, but one that they buy in from WeatherscapeXT from MetraWeather in New Zealand, although I am not even 100% certain about that. As far as I know MetraWeather provides the current graphics system that the BBC have been using for the past 10 years or more. If that’s so, I can’t see that the slowness in switching is down to the new graphics system itself, but more likely to do with the sourcing of the model data that it uses, and what exactly the visualisation of that data will look like, in particularly the mapping. I hope that we have seen the end of 3D fly throughs, which pardon the pun, never really took off in my opinion.
Forecast model Data
MeteoGroup promise ‘better weather forecasts and solutions’, which to me is the most interesting feature of this whole debacle. If they are not going to use UKMO data, and as far as I know can’t use the ECMWF model commercially, that would just leave the ubiquitous American GFS model. I suppose it’s not outside the realm of possibility that MeteoGroup could decide to buy in some cut-price model data in a deal with the French or Germans. The strengths of the various models was recently put to the test in the Caribbean, forecasting the tracks of hurricane Harvey and Irma, some models perform much better than others. The GFS is primarily built to serve North America, and unlike the UKMO model is not tweaked for an island in the eastern Atlantic. So what happens in the future if the GFS model fails to deepen a low as markedly as the UKMO model does, and on the strength of which, the Met Office issue Amber alerts for storm force winds?
So has anything actually happened?
Well, I suppose that all the weather presenters are now no longer working for the Met Office but the BBC, and some of the old lags have been replaced by women presenters to balance the male-female ratio. I did write back in April that the takeover was imminent, and I suppose that it has happened. But rather awkwardly both for the BBC and MeteoGroup nothing visually has changed, if anything the Met Office have come out of this rather well even if they did lose the contract, their recent national weather video service that’s produced using their new VisualEyes graphics system, which they push out on their website and over social media is rather good.
The graphic that the BBC Weather team used in their tweet this morning (fig 1), is of yesterday evenings forecast minimum temperatures for rural places at dawn. If you compare these spot values with the actual minimum temperatures [18-06] from the SYNOPs (fig 2), you’ll see that they’re pretty wide of the mark for many places, but especially so across England, as overnight temperatures were held up in many places by a combination of too much wind and too much cloud. As an example, minimum temperatures in the Vale of York were forecast to be around 2°C, but were actually around 9°C, the ever reliable cold spot of Exeter Airport was forecast to have been 4°C but was over 7°C, southern Scotland was forecast to have been 1°C, but Eskdalemuir could only manage a minimum of 7.3°C.
Initially I did think this cloud and rain was down to one of those magical troughs that suddenly appear in the analysis, and which run parallel with the isobars, but surprisingly, and rarely, the British isles were free of any frontal structures on the midnight analysis (fig 3). It’s quite obvious though, that this week the Met Office mesoscale model has never quite got to grips with the showery northerly airstream left behind in the wake of storm Aileen, and has performed very poorly, both in the handling of convection, and the subsequent forecasting of overnight minimum temperatures, due to the cloud, showers and wind associated with these features.
I think the person that handles the tweets for BBC Weather ought to be a little more careful, that the graphics used in any tweet are factual, and not as much as 7°C colder than they were in reality. The weather has a way of catching you out if you don’t continuously monitor it, and overly rely on a model that seems to be having one of its heads this week.
You wait all season dreading that a hurricane might come along, and then all of a sudden, three turn up all at once (fig 1). Well that’s how it might appear at first glance on the weather forecast at 11.58 AM on the BBC News Channel. Don’t be alarmed though, as Michael Fish might say, it’s not Irma, followed by Jose and Katia that are forecast to decimate Florida, it’s just some poor graphics that might lead some people to think they’re all queued up. Personally, I think the following graphic from the NHC illustrates how things might pan out effectively enough (fig 2).
As a tribute to all the BBC Weather presenters of the past, I think it would be only right and fitting that because they’re now no longer employed by the Met Office but rather by the BBC, that they should be immortalized in the named storms for the coming season 2017-18.
I notice that the Met Office are a bit slow on publishing the list for this coming season on their website, so I thought that I would give them a list of possible Christian names that they could use from BBC weather presenters past and present.
- Alex Deakin is now at Exeter and doing a grand job.
- Bill Giles or perhaps Bert Foord, Barbara Edwards, Bob Prichard, Bernard Davey or perhaps Bee Tucker.
- Carol Kirkwood everyone’s favourite especially the Daily Express or maybe Chris Fawkes?
- Darren Bett or shall we have Dan Corbett?
- Elizabeth Saary or perhaps Everton Fox?
- Francis Wilson he’s the only F as far as I can see!
- George Cowley the first or what about another old timer Graham Parker?
- Helen Willetts or maybe Heather Reid or even Helen Young?
- Ian McCaskill the inimitable or maybe it should be Isobel Lang – what ever happened to her?
- John Hammond one of the best of the lot, and rumour has it that he is still around. Stiff competition though from the likes of Jim Bacon, Jo Wheeler, Jay Wynne, John Kettley he’s a weatherman, and of course Jack Scott.
- Keith Best (wow, I had to think hard and long about this guy and I can still picture him).
- Louise Lear apparently hails from my home city and I just remembered Liam Dutton now on Channel 4.
- Michael Fish who else? I hope Storm Michael is not too strong, or is it Matt Taylor?
- Nick Miller or Nina Ridge (another meteorological one)?
- Owain Wyn Evans (thanks to quaesoveritas)
- Phil Avery, Peter Gibbs, Peter Cockcroft, Paul Hudson or maybe Penny Tranter?
- Rob McElwee, Richard Edgar or maybe Richard Angwin?
- Stav Danaos. There are a lot of other possibilities here – Suzanne Charlton, Susan Powell, Sarah Keith-Lucas, Simon King, Sarah Blizzard (I’m sure she must have changed her name).
- Tomasz Schafernaker, Trevor Baker (you need to be old to remember him) or even Tori Lacey?
- Wendy Hurrell (I never heard of this lady but I’m so glad I found her).
It maybe that by next weekend they’ll need a new list, because we could have a very early entry for the first named storm of the 2017-18 season.
Today’s showers were a lot sharper than forecast by the Met Office model that was used in the BBC forecast on Spotlight yesterday evening (fig 1), with some white pixels in the weather radar, indicating intensities of >32 mm an hour (fig 2). Thankfully they were moving quickly.
Neither was there any mention in the forecast of any thunderstorms in any of today’s showers over the southwest of England (fig 3).
Just checking to see how they did on last nights national forecast, and again no mention of any thunderstorms (and any flashing lightning graphics), or of the intense showers indicated by the weather radar. All Darren Bett seem to be interested in was saying how cool it will be, he must have been out of the country for the last four weeks or so. All I can think is that the new Met Office model is having problem in convective situations like this, I am assuming of course that the BBC are still using Met Office NWP data at the moment!
Or did they mention this large area of very heavy rain in the afternoon over Wales in earlier forecasts. Is it only me that notices just how poor the weather forecast was?
The Met Office seem to be having problems with both forecasting and observing just how much low cloud there is across the southwest of England again today. Although it’s cloudy across Cornwall at the moment, there is little in the way of low cloud at 09 BST over much of Devon as you can see from the visible satellite image (fig 2), and this at odds with the NWP graphics in the BBC forecast for this time (fig 1). No doubt the SC sheet will roll ENE’ward this morning and things will cloud over here, but in the meantime temperatures are around 24°C and its a lovely sunny morning. A similar clearance of cloud occurred yesterday at lunchtime, when the cloud cleared and the early afternoon was sunny, after another forecast of a mainly cloudy day from the graphics. The model does seem to overgenerate low cloud at times, and it doesn’t seem to make any difference if the base is at 2,000 or 7,000 feet.
The AWS the Met Office use are far from perfect either, both Exeter airport and Dunkeswell were at it again this morning overestimating how much low cloud there was. In my mind, there is absolutely no way there was 7 or 8 oktas of SC at 5,000 feet at either of these sites at 08 UTC this morning (fig 3). I can see why for safety reasons the AWS is programmed to err on the high side at an RAF station, but this is just plain misleading.
No wonder the presenters get the wrong idea about the weather across the region from observations like this. I think it might be a good idea if the Met Office introduced weather or skycams to these AWS as they did when they initially started trialling them, it’s either that or I drive down to the airport and start my observing career over again to see what’s going on.
Of course any outstation forecaster who uses one of these AWS will know full well its shortcomings and limitations when it comes to reporting cloud amounts, thankfully my observing career had come to an end before one was introduced at Kinloss. I should imagine that it can get a bit fraught at a fast jet station when the AWS suddenly announces that there’s 7 oktas of ST at 200 feet when it’s just a patch sat over the LCBR.
You could conclude, that the over forecasting and the over observing of low cloud, are some how linked, but as far as I know low cloud amounts from AWS don’t find their way into any NWP model, or maybe things have changed.