HadUKP is a series of datasets of UK regional precipitation, which incorporates the long-running England & Wales Precipitation (EWP) series beginning in 1766, the longest instrumental series of this kind in the world. Maintained by the Hadley Centre at the Met Office.
If I’ve got my programming right, the latest 12 month precipitation accumulations at the end of November for the UK, are still on the low side for Central, Southwestern and Southeastern parts of England (fig 1). The daily precipitation data is from the UKP gridded set which I download from the Met Office. The lowest anomalies at the moment are across the Southeast with 85.9% of the 1981-2010 long-term average (fig 2). After a wet spell in mid-summer accumulations were creeping back close to average, but it looks like a dryish October and November has put them back into a deficit.
The index of reservoir totals for England and Wales looks close to normal (fig 3), despite a number of reservoirs in the south being on the low side.
Rather surprisingly, Northern Scotland currently has the lowest annual rainfall anomalies for the whole of the UK, with an annual running accumulation to the end of September of just 90.9% of the 1981-2010 long-term average. The spell started with a very dry October and November in 2016, and the gap has only started to narrow in the last couple of months. It’s rather strange to think that somewhere can be seen as drier than average, when it’s seen rainfall totals of 1487.1 mm (58.5 inches) in the last year, but that’s the wonder of statistics for you.
The latest 365 day precipitation accumulations for England and Wales for June, are still returning negative anomalies despite recent some heavy rainfall events during the last month places. The England and Wales accumulation for the 30th of June was 797.9 mm, or 85.7% of the long-term average, the driest since Spring 2012.
It’s as if someone turned a tap off at the start of October last year in Northern Ireland. The graph above (fig 1), is a rolling 365 day moving accumulation, the yellow highlighted line is a 365 day moving average of that moving total (if that makes sense), to smooth out the accumulations. This sudden drop is mirrored in most, if not all regions, across the UK despite the rainfall in May, but is more marked in Northern Ireland. The steep decline in annual accumulations was brought about by a series of drier months from July 2016 onwards, that followed the previously wet Autumn and Winter (2015-16), but what triggers these sudden dry spells to suddenly stop and start is another question. The last one in Northern Ireland, occurred in 2004, it will be interesting to see just how quickly it bounces back. It happens the other way round of course, when a wet spell suddenly starts like it did at the end of 2015, but it never seemed as marked as when a dry spell starts.
Here’s another way of looking at the same daily values (fig 2), and as you can see, the accumulations deficit went negative at the start of last years anticyclonic October.
And here’s one that I prepared earlier for the Southeast of England, the deficit is still evident at the end of May, but has probably narrowed in the last week.
I reckon that there has been an almost 60% decline in annual snowfall since 1931. This won’t surprise a lot of people, because snow has become something of a scarce commodity in recent Winters, especially the further south that you are. Before I go any further the science behind this article is a bit thin, it’s based on a mix of daily Central England Temperatures [CET] and daily UKP rainfall (central region), but what the hell, you’ve got to start from somewhere, and I don’t think the Met office would have provided me with the required climate data to do this for free.
The biggest fudge factor, and don’t forget that even the most sophisticated and complicated NWP software employ some kind of fudge, is the algorithm that takes a daily maximum and minimum temperature, and decides if there’s any precipitation reported for that day, what likely probability is that it would fall as snow and accumulate. This is obviously easy if the maximum temperature for the day in question was below freezing, but not so easy say if the maximum is +5°C and the minimum is -1°C, so all you can do is give it your best guess. I suppose you could also look up the LWT for that day, and maybe get some kind of idea of what kind of air mass the country is under, but I didn’t go that far, and kept it as simple as possible.
Getting back to the chart (fig 1), you will see that my trusty algorithm has identified the snowiest winter in the last 86 years as being that of 1946-47, so that’s a good start. The second snowiest season it reckons was 1978-79, and having experienced of winters since the early 1960’s in various parts of the country I wouldn’t disagree with that. I guess that the accumulated snowfall for that season, ignoring melting of course, was close to a 100 cm, with 1946-47 producing an accumulation of just over 140 cm. I entitled the first article that I wrote about using daily CET and UKP to estimate a snow depth; ‘Central England Snowfall’; in fact it’s probably more appropriate to imagine the value as an index rather than a specific depth of snow. There have been some years with no snowfall when the algorithm couldn’t detect any snow, these were for the years 1988-89, 2013-14 and 2016-17, perhaps the code I calculate the daily probabilities requires a bit of tweaking, but then again, the last snow that I can remember settling down here in our part of Devon, was way back in 2010.
I did the original work for the application, which I call Central England Snowfall, about ten years ago now, but after deciding to give the code a bit of a spring clean today, I thought that I’d write an article around it. It’s a bit of fun because the science that I use could be regarded as suspect, but as far as I know, there are no long-term graphs available on the internet for annual snowfall totals for any climate station in the UK. If you do find one, or perhaps know of a graph than spans 60 years or more please let me know, because I would be very interested to see it, and would certainly include it in this article. I do have some evidence though, in the shape of daily climate records courtesy of Alistair McClean, Curator of Natural Science at the Weston Park Museum in Sheffield, for the period 1950-2010, which luckily includes daily values of fresh snow and snow depth. As you can see in this shorter period (fig 2), annual totals on the linear trend have also declined, by almost 40% in the last sixty years. The graphs are quite similar, even the estimate of snowfall that I make for the winter of 1978-79 are reasonably close (actual 116 cm estimated 98 cm). Having experienced that winter firsthand in the higher suburbs of that city though, I would say that the reported values maybe rather on the low side.
It’s a great shame that the Met Office stopped producing the Snow Survey of Great Britain in 1991, if that information could be digitised and collated it would make a wonderful climate resource for snowfall in the UK. Finally, I would just like to give a special mention to the website of Dr Richard Wild and is thesis ‘Spatial and Temporal Analysis of Heavy Snowfalls across Great Britain between the years 1861-1999‘, a fascinating read for the snow lovers amongst us.
The 365 day running totals of daily precipitation totals are below average in eight out of the nine regions (fig 1) across the UK. In England and Wales the accumulation over the last 365 days is 94.7% of the long-term average (fig 2). Not overly low of course, but many reservoirs in the south and west of the UK are well below 100%, where at this time of the year they usually are.
But it’s in Northern Ireland were anomalies are at their lowest at only 81.5% of the long-term average (fig 3), and probably the lowest they’ve been in the Province since 2004 (fig 3).
Here’s a closer look at the last year in more detail (fig 4). So rainfall totals dropped from September in what was a very anticyclonic Autumn and anticyclonic Winter.
NIWater don’t seem too worried about the reservoir levels (currently at 87%) there though (fig 5).
Let me know if I’ve screwed up with any of my UKP rainfall stats that I download from the Met Office website.
In a recent article about the Summer Index in Central England, summer 1954 came in at the bottom of the table with the worst possible scores for temperature, rainfall and sunshine since at least 1929. In researching for this article by looking back in the online archives of the Royal Meteorological Society, I did find that 1931, 1922 (mental note to find out why the summer of 1922 was so cold) and 1912 all rivalled 1954 as the worst summer on record, but I have a special affection for 1954 because it was in the summer of that year I was born. Here are the headlines for the months of the extended summer of 1954 that I’ve grabbed from the Monthly Weather Report for each month that the Met Office make available online (they are Crown copyright – so I hope they don’t mind).
May 1954 Mainly dull and wet, with frequent thunderstorms; large variations of temperature.
June 1954 Mainly dull and cool; periods of rain, heavy at times.
July 1954 Notably cool and dull; wet in some areas.
August 1954 Cool and dull, mainly wet in England, Wales and southern Scotland.
September 1954 Cool and unsettled; wet in most areas; sunny on the whole.
To begin with I thought that I would just look back at the circulation patterns of the summer using the reanalysis MSLP data from NOAA, so the next three charts are the mean pressure for each meteorological month of the summer, followed by three more anomaly charts for June, July and August. As you can see from the first three charts the summer was dominated by a west or northwesterly flow, which during July was quite strong.
The monthly anomaly charts for each month all show an anomalous low to the northeast or east of the British Isles, with mean pressure between -5 and -9 hPa lower than the monthly long-term average.
And here are the 1200 UTC surface temperature anomalies also from reanalysis data which show how cool, if not cold it was, not just across the British Isles but also the near continent through each of the summer months.
Here for good measure are the mean temperature anomalies for the entire summer courtesy of the Met Office.
The British Isles were in quite deep (for the time of year) cold air for long periods during the summer of 1954. I count 49 of the 92 days of that summer when sub 552 dm partial thicknesses (the green blobs) covered all or some part of the British Isles.
Of all the information and graphics that I have packed into this article, none of them conveys as much as the next graphic just how exceptional the Summer of 1954 was. It’s a chart of daily Central England Temperatures [CET] and anomalies for the entire Summer. The fifth chart from the top displays cold or warm spells that have lasted for four days or longer and were +/- 2°C above the long-term average for that day, and as you can see in the entire summer there were only three warm spells, and all of them spells of high night-time minima rather than day time maxima.
Here is a list of the coldest summers in the CET monthly series since 1659. Although 1954 was cold, coming in at the 17th coldest with a mean anomaly of -1.19°C, there have been colder summers including that of 1922 with an anomaly of -1.72°C.
And finally a look at the daily rainfall totals for England and Wales from the UKP data series that are maintained and made available by the wonderful Met Office yet again. It was a wet summer and according to the figures it was over 40% wetter than average. It looked very wet from the 5th of June for at least ten days or so, with wet spells again in late July and again through the first three weeks of August.
In writing this article I’ve finally come to realise that I have developed an amazing set of applications to display climate data, but charts, tables and maps aren’t the be all and end all of what makes an interesting article about past weather events. To glue all those disparate images together you need meaningful textual information about whatever event your article is about, and if that event happened over sixty years ago like this one did, when you were either perhaps too young or not even born, then you can’t always write about it with the benefit of first hand experience.
The drought of 1976 has been making headlines recently because it occurred 40 years ago this summer. Yep, not much of an excuse but what the heck. One of the best ways at looking at the event is by means of the daily UK gridded precipitation series for all the national and regional areas of the UK, and that’s the data I’ve used in the next few charts I’ve produced. I didn’t look for the classic 14 days without rainfall that is the definition of a drought because the values are derived from gridded values dry days are not common, and believe it or not the longest dry spell I found in 1976 was just 11 days. The first chart is for the England and Wales series which started in 1931, eventually the Met Office will extend these data sets back to 1910 or even earlier, I live in hope!
This chart is displaying a 365 day running total as you can see, the black horizontal dashed line is the long-term average (1931-2016) which for England and Wales is around 930 mm. You can plainly see 1976, the only time that the running 365 day total had dropped below the 600 mm mark for any time in the entire series, with only 1934 coming close to matching it. Here’s the chart in a bit more detail.
In a recent look back at the drought in a BBC Countryfile special about weather events, John Hammond correctly pointed out that the drought had its roots in the previous year, June the 5th 1975 to be precise. Some of my readers may remember that snow stopped play at Buxton on the 2nd of June 1975, another very unusual meteorological event (were they connected?). Anyway back to the drought and the above chart, as you can see 365 day moving total erratically declined from being at 1,065.7 mm or +14.4% above average, to just 552 mm or -40.8% of average on August 24th 1976, the zenith or is that the nadir of the drought.
Here are the values from all the other UK national and regional areas for part of the summer of 1976.
In Scotland as the chart below shows the drought is not as clearly delineated and far less severe as it was further south and peaked in early October rather than in late August, with anomalies of -22% of average. Having said that, there were several periods in the previous twenty years where average rainfall in Scotland had fallen to similar or even lower levels, 1959 and its great late summer for example peaked at -35.5% in mid October, late 1964, 1969 and 1973 were also drier.
How low can you go?
If you examine the monthly England Wales series which extends further back than the daily series and way back to 1766, there looks like there is only one other period that was sub 600 mm and could challenge the drought of 1976 and that was the year 1785. The 12 month accumulated rainfall in July of that year was just 522.1 mm (-42.9% of average), the daily rainfall could have been even lower, but the 12 month accumulation easily beat the 365 day accumulation of 552 mm in 1976. So it’s close (probably the worst drought in over 190 years), but no cigar for the drought of 1976 as far as I can see.
I was just looking at the latest Lamb Weather Types [LWT] from the Climate Research Unit [CRU] and ran a comparison on the beginning of June in previous years. Interestingly on the 14 days between the 30th of May and 12th of June, 2009 tops the list of best analogs to 2016, with a match of 88.7% when you compare the indices from the objective LWT data for both years. I will stress at this point that my method is just a simple comparison using these indices, and drawing any conclusion on what the rest of the summer might bring from a short 14 day analog would be very foolish. Having said that so few people read my ramblings, and since I’m paying for the privilege of hosting the xmetman blog, I thought I would do it anyway, so if you’re reading this remember you heard it here first! Of course if you’re reading this in the future and it turned out that 2016 was a great summer, just put it down to those pesky analogs.
If you scan down the closest matches in the list many of them from recent years. Here’s a look at the daily charts for 2009 and 2016 just to see how alike they are.
Both were initially anticyclonic, with high pressure to the north of the country, before pressure fell and low pressure systems spread in from the southwest. This is how the summer of 2009 turned out as far as Central England Temperatures [CET] were concerned, a warm spell at the end of June and start of July, the rest of July was cool and August was only just slightly above average.
And below is a chart of how the England Wales daily precipitation looked through the summer of 2016. July did turn out a very cool and wet month indeed. Looking back in hindsight, the rainfall accumulations were only 120% of average for the whole summer, but I suppose that you have to set that against what a disappointment it was after all the euphoric hype and raised expectations that the seasonal forecast by the Met Office received back in the April.
I don’t recall if the north Atlantic sea surface temperatures were as cool as they are this year so I can’t say if there was a similarity there. At the start of summer 2009 we had also just exited from a minor La Niña event and just entering an El Niño, so completely the reverse of what’s happening at the moment in summer 2016. We will just have to see what the next couple of months bring and hope for the best.
A bit of a strange title I know, but I’ve recently written an application that displays climate data for the UK from three separate daily data sets for atmospheric circulation, temperature and precipitation, and hence the tri.
Daily Central England Temperature [CET]
Objective Lamb Weather Type [LWT]
UK regional precipitation series [HadUKP]
It’s not the first time I’ve merged weather data sets in a single application, but this is probably the first time I’ve managed to finish it and publish the results that it generates. The essential requirement of course is a source of regular daily weather data, and so the CET and LWT series were the ideal (and only) choice because they are both updated on a daily or weekly basis. The other daily weather set that fits was the HadUKP series that the Met Office maintain, but there are a couple of problems with this series, one being that is only updated on a monthly basis, and the other is that the series isn’t very long and only extends back to 1931, and not 1772 and 1861 as in the case of CET and LWT. The big plus for anyone interested in the climate of the British Isles is that you can explore the climate of a particular day, week, month or season very easily and quickly. Here is a screenshot of the application as it stands now:
Below are a few examples of some particular well-known periods and spells of weather from the past, starting with a look at last Autumn and Winter.
You can certainly see the lovely anticyclonic spell that we had in September 2015, and the record mild November and December that followed, in this six month overview. Next a four-month window and a look at the Winter of 1946-47, you can clearly see how the cold started with an anticyclonic spell in the second half of January 1947, with the snow following along at the start of February.
Here’s the summer of 1976 and the record warmth of late June and early July, notice also the preponderance of anticyclonic types up until the start of September, then the breakdown into more cyclonic weather and the rains that brought an end to the drought.
Here’s the great winter of 1962-63, in comparison to 1946-47 it’s clear that winter 1962-63 started much earlier (before Christmas) and finished earlier, but was also drier and more anticyclonic.
I could maybe add an extra chart in the shape of a ‘barograph’ because I hold all the mean pressure points in the LWT data. I could present that as a scatter graph of all the 16 MSLP grid values for 12 UTC and then plot a moving average. I could also highlight with a star the named storms, but that would only work for the very latest years. I could also colour the precipitation bar chart blue to indicate snow rather than rain when the CET was less than 1 or 2 °C (I have in fact now implemented that idea as you can see if you look at the screenshot of the application!). I do plan to add functionality to show a grid of archived weather charts for the selected period from Wetterzentrale. The one element that I think it does miss is daily sunshine data, but there is no source that I know of for daily sunshine values for a region, let alone for a single station, so that’s a non-starter. I must say that this really is an excellent tool for any climatologist with an interest into the weather of the British Isles over the last 150 years or so.