The Central England Temperature [CET] record is a meteorological dataset originally published by Professor Gordon Manley in 1953 and subsequently extended and updated in 1974, following many decades of painstaking work. The monthly mean surface air temperatures, for the Midlands region of England, are given in degrees Celsius from the year 1659 to the present.
This record represents the longest series of monthly temperature observations in existence. It is a valuable dataset for meteorologists and climate scientists. It is monthly from 1659, and a daily version has been produced from 1772. The monthly means from November 1722 onwards are given to a precision of 0.1 °C. The earliest years of the series, from 1659 to October 1722 inclusive, for the most part only have monthly means given to the nearest degree or half a degree, though there is a small ‘window’ of 0.1 degree precision from 1699 to 1706 inclusive. This reflects the number, accuracy, reliability and geographical spread of the temperature records that were available for the years in question (intro courtesy of Wikipedia).
Autumn 2016 in Central England was slightly warmer than average with a mean temperature of 8.57°C (based on provisional figures), which was +0.42°C above the 1961-1990 long-term average. There were a couple of highlights during the Autumn 2016.
- A cold spell during the first two weeks of November with a minimum anomaly of -7.2°C on the 8th.
- A mild December from the 6th, and exceptionally so during the second week, with two new extreme records broken, marked on the above graph with a red star for a new highest maximum on the 9th, and a blue diamond for a new highest minimum on the 10th.
Six air frost’s so far this Autumn, which may not sound a lot but it’s not far of the average, and certainly higher that in 2013/14.
In central England the mean temperature of 9.0°C in the last week (7th to 13th December) made it the seventh mildest week in the daily CET series since 1772, with an anomaly of +4.3°C.
Of course these are still provisional CET values from the Met Office, but significantly last week was a full 1°C warmer than the same week in 2015, which as you probably know was not only exceptionally mild and the mildest December since 1659, it also the highest anomaly of any month as well. This table is ranked on mean temperature anomalies for all months since 1772, and as you can see December 2015 beat them all into a cocked hat.
Going back to the last week in the CET series and you compare the mean maximum you’ll find it was the joint second mildest week (7th – 13th December) since 1878, with a mean maximum temperature of 11.9°C and +4.3°C anomaly.
I make the CET mean anomaly up to the 13th of December +1.77°C, and I can’t see the remainder of the month being a lot different judging by the available NWP from the American GFS model. It would take a very intense period of severe cold indeed to even reduce the final CET anomaly for this month back to zero.
This current very mild spell has been threatening to break a daily CET record all week and it finally managed it yesterday (9 December 2016). Yesterday’s maximum of 13.5°C (provisional) exceeded the previous warmest 9th December by 0.8°C in the daily CET series that started in 1878. Yesterday was just over +6°C warmer than the 1961-1990 long-term average for that day.
The maximum anomaly for the last three days of CET (7-9 December) have all been round the +6°C mark, and yesterday the minimum anomaly was a massive +8.6°C above the long-term average. The last five days have seen a remarkable turnaround in daily CET values from cold on Monday to record mild by Friday.
Yesterday was (provisionally) the second warmest 7th of December since the CET daily series commenced in 1879, the mean temperature of 13.8°C was +5.8°C above the long-term average, and was only pipped by the 7th of December of 2015 into second place. It was an exceptionally mild day across the whole country, in fact it seemed to me that the mildest areas seemed to lie outwith the CET area.
Yes, a rather cold November 2016, but not exceptionally so in CET series, with a mean temperature for the month of 5.5°C, which was -1.0°C below the 1961-1990 long-term average and in the 36th percentile in the monthly series that started in 1659. That mean anomaly 0f -1.0°C made it the coldest November since 2010 (not much of a headline I know, but I challenge the Met Office Press office to make any more of it than that) and only the second month out of the eleven this year with a negative anomaly, the other month being April. Interestingly the minimum anomaly of -1.34°C was much lower than the maximum anomaly of +0.68°C, reflecting the predominantly anticyclonic nature of the month and the cold nights. Thanks, as always, to the Met Office for the data, and to Gordon Manley for dreaming the whole thing up in the first place.
Winter mean temperature anomaly maps courtesy of the Met Office
It’s the last day of what people are now calling ‘meteorological’ autumn, so I thought I would put some code together to display daily Central England Temperature [CET] series for any season as a grid of weekly deciles, along with the corresponding mean temperature and anomaly. I did this so that I could take a closer look at winters in the series since 1772, and rather surprisingly I found no winter has been consistently cold throughout the 13 weekly temperatures deciles (1), although the winter of 1794-95 did come close with 12 out of 13 decile 1 weeks. Another surprise, was that the winter of 1978-79 tied with both 1946-47 and 1962-63 with 10 weeks of decile 1 mean temperatures. One thing that I always seem to be being reminded of is how late the winter of 1946-47 started, and if you look closely the first fours weeks of the winter were indeed average or mild. The code for this has been more than a little awkward to get right, mainly because I’ve added functionality to allow for meteorological seasons and also output the results as quintiles as well as deciles.
It’s seems a long time since the last cold winter in 2009-10, and the weekly deciles since the late 1980’s have been predominantly of the mild 3 variety. The mildest winter that I can see from my results looks to be 1989-90 with 12 weeks of decile 3, followed by 2013-14 with 10 weeks of decile 3, in fact 9 out of the top 10 mildest (astronomical) winters as far as I can see have occurred in the last 30 years if you rank winters by their weekly mean temperature deciles.
I think the best way to visualise Central England Temperatures [CET] is by means (excuse the pun) of a 365 day moving average. Any shorter and it’s a little too ‘noisy’, any longer and you lose some of the more subtle nuances that are present in the underlying daily values. Here is a chart of the last 30 years, spanning the period from November 1986 to November 2016. I have annotated the graph to remind both you and me of what month or season probably caused the various upturns and downturns in the anomalies during those thirty years. At the moment we look like we are possibly at the start of a downturn on the roller coaster ride that the CET series is on, and who knows where that will lead in the coming months.
Data courtesy of the Met Office
By the way, the series in the graph with a black line is a moving 365 day average of a 365 day average and that’s why the line is so smooth. It does take a bit of thinking about, but it seems to work just fine!
Knightshayes near Tiverton in Devon (31 October 2016)
The Central England Temperatures [CET] for October 2016 ended up with a bit of a flourish, and what looked liked being a very average month in terms of temperature, ended up a little bit on the warm side. The monthly mean was 11.05°C, which made it +0.54°C above the 1961-1990 long-term average, and the 36th warmest October since 1659 in Central England. I have got egg on my face with this month ending up warmer than average, because I was convinced when it started that it was going to be a cold month in general – my advice to anyone who is reading this is – never trust an analog.
There have been many warm October’s since the 1950’s and this is just one more of them. October is quietly becoming less of an autumn month as the years go by and more of a summer one.
As you can see in the chart below the last week of October was not only warm by day, it was even warmer by night, which pushed the overall mean temperature up that bit higher.
I get the distinct feeling that October 2016 is going to be an unusual month as far as weather is concerned in the British Isles. The reason I say this is down to the early intense Scandinavian anticyclone that initially formed on the 3rd of the month, and the first substantial block we’ve had in the circulation pattern across the British Isles for some considerable time. As usual I always call on the aid of the central England temperature [CET] series to help me out , and this time is no different. Here’s a table of the coldest October’s in the CET series. You’ll notice that I’ve highlighted 1881, I did this because 1881 was the best match that I could find to the start of this October in the Objective Lamb weather type [LWT] series. The October of 1881 was the eleventh coldest since 1659 in the monthly CET series, with a percentile of 4 (i.e. one of the top 4% coldest) and a mean temperature of 7.06°C, which was -3.45°C below the 1961-1990 long-term average.
This is a graph of all Octobers since 1659. There has been a slow warming over the centuries in October’s, with many more warm than cold since the World War II. As as you can see the last notably cold one was in 1992.
Below is a more detailed look at the CET of Autumn 1881 and the one thing that striked me about it how quickly a long period of cold can suddenly give way to a long period of mild, before flipping back to cold again.
The month of October across the British Isles was very anticyclonic in nature. It started in a very similar way to October 2016 with a large anticyclone over Norway, which eventually gave way by the 10th to allow a low pressure system to run and intensify from Iceland and into the Norwegian sea and into the Baltic before filling and being replaced by a transient high pressure cell on the 16th. The high migrated to Scandinavia and before you know it the 19th of October looked a lot like the first day of the month again. That high remained for the rest of the month, it retrogressed NW’ward from the 20th to allow low pressure to take up residency across the SW of the British Isles and produce another spell of E’lys from the 19th to the 30th of the month. The low over the SW gradually filled as high pressure built from the north from the 25th, and by the end of the month an anticyclone sat over the country.
Here are the 0000 UTC weather charts for October. Please note that the LWT table uses 1200 UTC data so they are not perfectly in synch.
I would have included a graph of rainfall for October 1881 but the daily totals only go back as far as 1931, and the regional monthly totals only go back to 1910 so they’re of no use, so we are left with the total from the monthly England Wales [EWP] rainfall series of 82 mm, which was 94% of average. It’s surprisingly difficult to glean much information about past weather across the country, and 1881 is just a little too early to be included in the daily and monthly weather reports of the Met Office.
Well what was all that in aid of? Well it’s been almost 25 years since we experienced a really cold October, and I suppose I’m still vaguely looking to find some kind of analog between the October of 1881 and 2016. I know I’m totally delusional, but what the hell.
September 2016 ended up the fifth warmest in the central England temperature [CET] series that began in 1659. I reckon the mean temperature for the month was 16.05°C, and the anomaly +2.53°C above the 1961-1990 long-term average. It never quite matched 2006 and almost caught 1949, and after all’s said and done that only makes it the warmest since 2006, the warmest September in the series.
Both maximum and minimum temperatures were almost always above average for the entire month, with no real cold spell in evidence. Two new CET maximum records were set on the very warm spell on the 13th and 14th, and a new high minimum was also set on the 14th as well.