Super mild

2016-11-14_140456

1300 UTC 14 November 2016

Super mild this afternoon at Hawarden in Flintshire, with the 1300 UTC air temperature already at 15.8°C (60.4°F). I should imagine that they’ll be the mildest place today in the UK, all thanks to the mountains of Snowdonia generating a foehn wind in the moderate to fresh south-westerly  flow.

air-temperature-c-1300-utc-on-monday-14-november-2016-in-wmo-block-03-highest-air-temperature-c

Look back at October 2016

monthly-weather-report-oct-2016

Data & Images courtesy of the Met Office

What is it about halloween?

2016-10-31_185023

Last year you’ll remember that the temperature at Trawsgoed in Ceredigion, Wales on the 1st of November 2015 – all Hallows’ day – reached 22.4°C and set a new record for the warmest November day in the UK. Similarly 2016 seems to be following suit, because the maximum temperature today on the 31st of October 2016 – all Hallows’ eve – and spookily again at Trawsgoed reached 22.2°C.  Although this wasn’t an October record it got me to thinking when was the warmest 31st of October up until now? To be honest I don’t have the exact answer, but with the help of the Central England Temperature [CET] series I thought that I could make a very good educated guess. I already have the code (now there’s a surprise) to list the CET for any particular day of the year. You select a particular day and it lists the warmest or coldest maximum, mean or minimum temperatures for that particular day back to 1878, or for means back to 1772. And here’s the output from that program for the 31st of October:-

warmest-central-england-temperatures-for-the-31-october

Warmest 31 October in Central England (1878-2015)

So it looks like the Halloween of 2014 with a maximum CET of 20.6°C looks a pretty safe bet. The maximum anomaly for that Halloween was a massive +8.3°C, the second warmest on record occurred on the 31st of October 2009 with a maximum of 16.6°, a full 4°C lower. So what did the chart for Halloween 2014 look like?

synops-1800-utc-on-fri-31-oct-2014

1800 UTC on 31 October 2014

As you can see a broad SW’ly warm sector with fresh southerly winds.

Courtesy of the Met Office

Courtesy of the Met Office

Here are the plotted 0600 to 1800 UTC maximum temperatures for Halloween 2014.

maximum-temperature-c-1800-utc-on-fri-31-oct-2014

Maximum Temperature [06-18] on 31 October 2014

And as you can see our old favourite hot-spot of Gravesend was the warmest place again with a maximum of 23.6°C. I’m sure I heard Louise Lear (5.57PM) say that the 22.2°C maximum today at Trawsgoed made it the “warmest Halloween on record”, I also heard her at 6.27PM when she had changed that ever so slightly to the “warmest Halloween in West Wales” (which is probably correct), I think she might like to take a look at Halloween 2014.

maximum-temperature-c-0600-1800-utc-on-friday-31-october-2014-in-wmo-block-03-highest-maximum-temperature-c

Maximum Temperature [06-18] on 31 October 2014

And just for the record these are the warmest places today across the UK for you to compare.

maximum-temperature-c-0600-1800-utc-on-monday-31-october-2016-in-wmo-block-03-highest-maximum-temperature-c

Correlating Summer temperature, rainfall and sunshine in the UK

A while I go I looked into a summer index for the UK, this gave me the idea of producing a couple of charts that correlated mean maximum temperature for the summer [JJA] against rainfall and temperature.  And as you can see from the two charts below it gives a very quick and easy way to spot the warm and sunny years against the cold and wet. Instead of using my old favourite of Daily Central England Temperatures [CET] and England and Wales Rainfall [EWP], I’ve used instead the historic regional data series from the Met Office that extends back to 1910 for temperature and rainfall, and from 1929 for sunshine. The regional monthly and seasonal values are of course calculated from gridded values and that’s the reason they are so short. The application can display charts for any region but for now the entire UK is what is of interest to me. It maybe that a linear trend line is not the best way of testing any correlation, and there must be a way of doing a three-way correlation, but for now this will have to do! Click the image to enlarge it as usual.

UKMO Climate series UK Summer Sunshine, Rainfall & Temperatures

The wettest year on record

I’ve just read a Guardian article that surprised me because it stated that:-

Meteorological records show that six of the seven wettest years since records began occurred from 2000 onwards.

Courtesy of The Guardian (30 July 2016)

Courtesy of The Guardian (30 July 2016)

I did comment that I thought this was “not true and misleading”, but I have a terrible memory so I had to dig out the facts, and so to do this I first looked at the Met Office’s abridged 1910 data series. I say abridged because as most people who are in the know know, detailed rainfall record are available for most of the UK from at least 1859 and some records way back into the 16th century. The table below is for annual totals from 1910 for England and Wales, and as you can see in that only four out of the top seven occurred since 2000.

Climate Data 1910 - 2016 England & Wales

So undeterred, I looked again and looked at the UK annual estimates since 1910, and there they were, six out of seven wettest years since 2000. What’s also true is that six out of the last thirteen wettest years also occurred since 2000 but that’s certainly not as punchy.

Climate Data 1910 - 2016 UK

Climate Data 1910 – 2016 UK

Undeterred again, I thought I would see what the England and Wales Precipitation [EWP] series (which started 250 years rather than 106 years ago) revealed about the wettest years, and this was even less punchy again with only two of the last seven wettest years occurring since 2000, but a very creditable five out of the top twenty wettest years since 2000. The wettest year 2012 in the shorter series drops to third wettest in the longer series. But I think you get my drift – the shorter series is very useful for headlines, the longer series is well more accurate?

Tabulated England Wales Precipitation - Annually's 1766 - 2015

Tabulated England Wales  Annual Precipitation 1766 – 2016

And so is the writer of the article Robin McKie, Science editor at the Guardian right to use the shorter series? Yes I suppose so, but the longer series does put it into a better perspective. Although having said that the simple linear trend on the longer EWP series does show that annual rainfall in England and Wales has increased from 894.3 mm in 1766 to 941.9  mm in 2015, an increase of +5.3%.  Robert McKie is of course free to use my trend, but the six out of the seven wettest years sounds so much more alarming than a boring +5.3%. I won’t be issuing a retraction to my comment to the Guardian, but I’m sure they’ll understand.

England Wales Precipitation - Annually's 1766 - 2015

England Wales Annual Precipitation 1766 – 2015

A Summer Index: 1929 – 2016

UK Summer Index 1929 - 2016

UK Summer Index 1929 – 2016

I remember reading the article “A simple Summer Index with an illustration for Summer 1971” in the Weather Magazine when it was published in April 1972. The author R.Murray  of the Met Office (which was then in Bracknell), had come up with an algorithm to calculate a simple Summer Index of how good the summer [JJA] had actually been. You can read the first page of the article here. I’ve used the freely available gridded data to calculate the Summer Index using the mean temperature, rainfall and sunshine values which I download from the Met Office and which extends back to 1910, this divides the country into regional areas as well as national ones. Unfortunately the sunshine data doesn’t start till 1929, but at least I can quickly jump from region to region using this application, to get a feel for how the summer was in other parts of the UK. It was a little fiddly to set up and calculate the quintiles for the mean temperature, and the terciles for the total monthly sunshine and rainfall. Using Murray’s algorithm the absolute best summer can score +48 and the absolute worst -48. The result I get for the UK is -8 for 1971, the year that Murray uses as an illustration in his article, but it’s -13 if I choose ‘England’, which is exactly what Murray calculated! I did think of using Central England Temperature values but there was little point without the monthly sunshine totals which as I say only start in 1929, so my Summer Index can’t reach back as far as 1881 as Murray’s did, but of course he was working for the big MO and didn’t have to scratch around for his data. The beauty of the algorithm is that it can work for any region or site, but obviously it’s good to have as much data as possible.

Best Summer

Best Summer Index in UK 1929 - 2016

Best Summer Index in UK 1929 – 2016

So, no surprises about 1976 having the highest Summer index in the UK since 1929, followed by 1995, 1975 & 1933. In fact 1976 scored a maximum of 48 because the three temperature quintiles were at a maximum of five, as were the rainfall and sunshine terciles at one and three respectively. Interestingly, 1976 isn’t the highest Summer index in every region, in the North of Scotland for example, 1955 had a much higher Summer Index (42).

Best Summer Index in Northern Scotland 1929 - 2016

Best Summer Index in Northern Scotland 1929 – 2016

Worst Summer

Worst Summer Index in UK 1929 - 2016

Worst Summer Index in UK 1929 – 2016

As you can see from the above table the worst Summer Index in the UK since 1929, was that of summer 1954 with an index of -48, the absolute minimum, it’s a shame about that because it was in that summer that I first came into this world. I was going to graph my results but decided against it even though Murray’s results did suggest a quasi-biennial oscillation in summer weather. Good summers in odd years, poor summers in even years, but with the benefit of hindsight and the summer of 1976 I thought I wouldn’t bother!

My Ideal Summer

If I were to devise an index then I would use daily values and then mean the total of these daily values to calculate a summer index. A perfect day for me would have a maximum temperature of 21°C and no higher. It would of course need to be dry and reasonably sunny but occasional puffs of cumulus or cirrus wouldn’t reduce the index for that day. More importantly visibility would have to be exceptional, the dew point (and resultant humidity) would have to be low, and there would have to be at least a light or moderate breeze during the afternoon. Of course this wouldn’t fit Murray’s original algorithm that was only dependent on monthly values, but that’s’ me being picky as usual!

June 2016: Rainfall

I thought that because I’d already written about temperature and sunshine across the British Isles yesterday, I should finish it off with a quick look at the rainfall totals in June. So here they are and guess what Capel Curig with 184.8 mm (7.28″) is the wettest place from the list of all the available SYNOPs!

Wettest Places Total Accumulated Precipitation 1 June 2016 - 1 July 2016 WMO Block 03

Wettest Places Total Accumulated Precipitation 1 June 2016 – 1 July 2016 WMO Block 03

I’m still trying to get my head round the 32.1 mm from Herstmonceux on the south coast, I seem to have 98% of the obs but the value looks low. Perhaps the gauge has a spiders nest in it like mine did early last month. The 153.3 mm from Kenly (to the south of London) made them #5 in the list from all the thundery rain they received.

Accumulated Precipitation for 0600 UTC 1 June

Accumulated Precipitation for June 2016

At the other end of the list Kirkwall finished with the lowest total of the month with just 11.3 mm of rain (0.44″). This is another one that’s hard to believe, yes the first 10 day of the month were fairly anticyclonic, but as the westerlies returned mid month, the Northern Isles were again open to all the various lows that tracked across the country.  Is this another spider? Or is it bugs in my program? That is a strong possibility, as you the reader are my testing team, and I’m notoriously bad at testing the applications that I write.  I am heartened to see that Wick only received 45.8 mm (89% received) in June , so hopefully some kind person at the Met Office might let me know how close I got! Who am I kidding, there is very little chance of that happening with just 13 subscribers, but you can live in Hope, well you could if you can afford the house prices.

Driest Places Total Accumulated Precipitation 1 June 2016 - 1 July 2016 WMO Block 03

Driest Places Total Accumulated Precipitation 1 June 2016 – 1 July 2016 WMO Block 03

Accumulated Precipitation for 0600 UTC 1 June

Accumulated Precipitation for June 2016

June 2016: Sunshine in the UK

I’ve just done a quick bit of coding to allow me to produce graphics for the sunshine totals for June from across the UK. The daily sunshine totals I parse from each day’s 0600 UTC SYNOP reports which I download from the OGIMET site. Top of the shop across the British Isles in June was Valley on Anglesey in Wales, with 218.8 hours of sunshine (43.2% of possible), closely followed by a whole host of other stations from western parts.

Daily sunshine 1 June - 1 July 2016

Daily sunshine 1 June – 1 July 2016

Daily Bright Sunshine 1 June 2016 - 1 July 2016

Daily Bright Sunshine 1 June 2016 – 1 July 2016

Across Europe Valley’s 218.8 hours couldn’t quite match the 404.1 hours (90.4% of possible) from Toledo in Spain. Oh to have a field full of solar panels in southern Spain.

Daily Bright Sunshine 1 June 2016 - 1 July 2016

Analysing the sound of thunderstorms

Analysing the sound of thunderstorms (courtesy of the Guardian)

I, as an observer, am slightly offended by this article that I found in the Guardian this morning. Not just personally but for the thousands of dedicated amateur observers across the country who also can tell the difference between a firework and a rumble of thunder!

I was a weather observer for over twenty years with the Met Office and NO observer worth his salt would report a thunderstorm because he heard a firework going off either on November 5th or New Years.

Most weather stations in the Met Office network since the last war are based on military or civilian airfields, and because of the risk of lightning strikes when refueling aircraft they are very sensitive to lightning and thunderstorms within a 10 nautical mile radius of the airfield and issue a lightning risk 1 or 2 warning if conditions like this occur to halt refuelling. The chance of a professional observer at a site like this reporting a firework as a thunderstorm are near zero. Granted, from a coast guard or auxiliary observer the chances may be higher, but still low, because these observation were QC checked by the professional meteorologists who collected them, but obviously in recent years even this has been automated.

The Met office have been “listening to lightning” for many years, and not just in the UK but from across the world using Arrival Time Difference [ATD] lightning (sferic) detection, and have a very clear idea of the number of thunder days there are during the year and don’t need poor observations that were obviously used to generate these stats used in this research mentioned in the Weather magazine. The only problem with ATD data is that the Met Office would rather sit on it that disseminate it to the public of this country who pay for it. Thank goodness for the Blitzortung organisation for building a network of lightning detectors for the world and making it available for all people to freely use (shame on you Met Office).

I suggest that M.J Owens either (a) redo’s all the research to exclude all but military or civilian airfield observing sites, observatories such as Lerwick and Eskdalemuir, and radiosonde stations (obviously this would be a very small group of stations now in 2016!) , or (b) uses all the ATD sferic data that the Met Office holds for the last 40 or 50 years, to assess whether the frequency of thunder in the UK has changed in recent years? I would like to bet that any changes that the findings show are down to the introduction of Automatic Weather Stations [AWS] in the last twenty years, which as far as I know in the UK cannot detect thunder as ‘present weather’, although I am aware that the more sophisticated AWS from the likes of Vaisala can.

By sheer coincidence – and as the few loyal followers of this blog will know – I wrote an article about this very subject days of thunder only yesterday!

Days of thunder

Accumulated Precipitation for 0600 UTC 1 January

Days of Thunder 1 Jan –  26th June 1016

It’s very difficult, I would say impossible to find the latest detailed statistics on thunder across the UK, let alone across Europe and other parts of the world. The obvious answer would be to generate monthly frequency maps from the output of the Blitzortung lightning system, but I am not a member, and even if I were it may still not be possible to get my hands on the data to do this with. Despite all this, I have come up with a simple and very effective method of compiling ‘days of thunder’ from SYNOP observations. It depends on the present weather code [WW] and the past weather codes [six hourly ww1 & ww2 code from the main synoptic hours] and the fact that thunderstorms have such high priority there are reported above any other present or past weather codes. So it’s simply a case of writing some software to count any thunderstorms that are reported in any of the SYNOP observations. There is one problem, and it’s a very big problem nowadays, and that is automatic observations which make up as many as 90% of SYNOP observations in countries such as the UK, don’t report thunderstorms as far as I can see (although there may be exceptions – it’s a big world). That’s why when you look at the maps that I’ve produced you will see a lot of spurious zero values plotted. It’s easy to work out an automatic observation in the UK & Ireland, but not so in some other countries such as France, in time I may be to winkle out all the automatic like this, but for the moment they are included.

Anyway the top map is of days of thunder for the 1st of January to the 26th of June, and as you can Erzincan in northeast Turkey is top of the European list with 40 days of thunder, in comparison Brize Norton in Oxfordshire with 9 days is top of the available manual stations in the UK. A quick look at America and the Caribbean reveals that David in Panama is top of the list there with 58 days of thunder so far this year. Tampa has only 1 day because the observations are missing for a good deal of the time. Hopefully with a little more polish, the output from this application might be a little less ambiguous the next time you see it, for now its work in progress!

2016 certainly seems to have been a very thundery year so far, but I would have to do an awful lot more data processing to calculate a thirty year mean for a great many stations before I could say it is. At one time the Met Office published an hourly text bulletin of sferics across Europe [SFUK26], but over 10 years ago they stopped issuing it which was a big pity. The more frequent and more detailed high-resolution bulletin which replaced it [SFUK27] may still be being produced but was never made public on the internet, which is another great pity. Thank goodness for the Blitzortung organisation who had the foresight to see the importance of monitoring lightning from thunderstorms and making it freely available to all, something I thought that national weather services were supposed to do.

Days of Thunder for 1 January - 27 June 2016

Days of Thunder for 1 January – 26 June 2016