Temperatures in central England are still flatlining this month (fig 1). Provisional till the 15th, indicate that 2017 has been the coldest start to an August since 1993. It seems that August can all too easily get stuck in a rut in some years, and it looks like 2017 is one of those years. The current mean maximum anomaly is 1.14°C lower than the 1961-1990 average.
The coldest August since 1910 was that of 1912. It was both wet, cold and dull. At many coastal stations around the UK the sea temperature was higher than the air temperature. In the monthly weather report for 1912 the review of the month concludes by saying
Observers in various parts of the United Kingdom noticed in the rare intervals of fine weather the sky seldom assumed its ordinary blue tint, but appeared to be covered with a hazy film “producing grey whiteness of the unclouded sky, and extreme weakness of all sunshine”. A similar appearance was noted by several continental observers.
The mean temperature for the UK was just 11.7°C which is 3.26°C below the 1981-2010 long-term average. The mean maximum anomaly for the month was even lower at 3.94°C below average. Looking at the regional temperature anomalies (fig 2), the cold was across the board.
As well as being very cold, it was also very wet, especially in more southern regions, and particularly in East Anglia which saw over three times the monthly average rainfall (fig 3).
In fact August 1912 was and still is the wettest August in the entire EWP rainfall series that started in 1766 (fig 4).
As you probably noticed in the daily charts for the month (fig 1), August 1912 was a very cyclonic month as you can see in the Lamb Circulation types for the month (fig 5).
As regards the CET for the month, I can’t remember ever seeing a summer month as cold as this one (fig 6). August 1912 was, and still is the coldest August on record back to 1659, beating even the cold year of 1695 into second place.
Not only was it the coldest August on record, it was also ushered in the start of a three-month cold spell in central England, with a cold September (mean anomaly -2.5°C) and October (mean anomaly -2.4°C) to follow (fig 7).
At this point I would like to produce some statistics to show that August 1912 was also the dullest on record, but I can’t, the Met Office maintain that they only began measuring sunshine from 1929. The MWR comes to the rescue though, because it says about sunshine:
Sunshine was very deficient, a large number of stations situated in nearly all parts of the kingdom recording considerably less than half the average amount. In the Channel Isles and at a few places in the extreme southeast of England the mean daily duration ranged between 4 and 4½ hours, and was equal to about 30 percent of the possible. Over Central and Southern Scotland and at a few places in the northeast of England the daily duration was less than 2 hours; at Crathes, Glasgow and Eskdalemuir it amounted to only 1o percent of the possible.
What caused it?
Here’ a graph of 12 month rolling CET values for around that time, forget the date in the subtitle, another bug for the programmer to fix. I’ve overlaid the volcanic dust index events that were greater, or equal to 4, on the VEI on top of the line series, and as you can see the Novarupta event (VEI 6) looks like it may well have been responsible for global cooling that also affected our own CET series back in 1912 across, and fits well with the reports of a greyish white haze from the Monthly Weather Report for August 1912.
I had never heard of the Novarupta eruption until I started researching this article today. I never even realised it was the most powerful volcanic eruption of the 20th century, here’s what the Wikipedia article had to say about it:
The eruption of Novarupta in the Aleutian Range began on June 6, 1912, and culminated in a series of violent eruptions. Rated a 6 on the Volcanic Explosivity Index, the 60-hour-long eruption expelled 13 to 15 cubic kilometers (3.1 to 3.6 cu mi) of ash, 30 times as much as the 1980 eruption of Mount St. Helens. The erupted magma of Rhyolite, Dacite, and Andesite resulted in more than 17 cubic kilometers (4.1 cu mi) of air fall tuff and approximately 11 cubic kilometers (2.6 cu mi) of pyroclastic ash-flow tuff. During the 20th century, only the 1991 eruption of Mt. Pinatubo in the Philippines were of a similar magnitude; Pinatubo ejected 11 cubic kilometers (2.6 cu mi) of tephra. At least two larger eruptions occurred in the 19th century: the 1815 eruption of Tambora (150 km3 (36.0 cu mi) of tephra), and the 1883 eruption of Indonesia’s Krakatoa (20 km3 (4.8 cu mi) of tephra).
There’s a interesting article on the Geology.com website about the eruption that you might find useful.
The latitude of the Novarupta eruption was just about perfect at 58° north for maximum effect across the northern hemisphere, as was the timing of the 6th of June to coincide with the cold months of August, September and October of 1912 in the CET series. I should imagine that the ash from the eruption would have taken at least a month or so to completely encircle the northern hemisphere and reduce the amount of sunlight. I am of course completely guessing that this was the cause of the cold August of 1912, I’ll have to spend some more time looking at NCEP reanalysis surface temperature data for 1912 to see just what affect it had on other countries across the northern hemisphere to completely be sure of my assertion.
Finally here are the daily CET values for the Summer of 1912 (fig 11), which I think says it all. This is one of the better articles that I’ve put together for my blog, I found little evidence of any link between the cold August of 1912 and Novarupta in any of my climate and weather books, and even though Philip Eden does mention the poor summer of 1912 in his book ‘Great British Weather Disasters’, he doesn’t make the link with the volcanic ash of Novarupta. August 1912 does get a short mention as the worst on record in the book ‘The Wrong Kind Of Snow’, but again no mention of why. So if you thought that August 2017 has been cool in its first ten 10 days, the mean temperature for the first 10 days of August 1912 was 2.5°C colder still.
I notice that the latest provisional mean maximum daily temperature for the first eight days of August in Central England of 18.89°C (-1.38°C) are the lowest for that period since 1987 (fig 2). The 10 year average mean temperature for this period, has been on a downturn since 2000, and even the linear trend shows only limited warming since 1772 (fig 1).
Apologies for the delay in publishing the chart of daily CET values for July 2017 (fig 1), the situation was out of my control, as it usually is when it comes to climate data for the UK! After all that, July was a very mediocre month as far as temperatures were concerned in Central England. I made the mean temperature for the month 0.73°C above the 1961-1990 long-term average. It’s quite unusual in my experience, to see daily temperatures so close to the daily mean as they were from the 20th to the end of the month.
After a warm start to the year up till the end of June, 2017 was vying for the warmest year (to date), but thanks to this lacklustre performance in July, it’s now slipped back to 5th position, with a mean (to the 7th of August) of 10.73°C, which is 1.52°C above the mean to that date, 2014 still heads the pack.
I have a number of obsessions as you’ve probably noticed, most of them can be summed up as weather, climate and blogging. I first started my obsession with climate stats back in 1983 when I took a redundancy package from British Steel, and rejoined the Met Office. That package gave me enough money, the first in a few lean years after the steel strike, to buy a BBC micro computer, and one of the very first applications that I wrote using it was a BASIC program to visualise monthly CET data. The first job was to type in over 300 years of monthly mean temperatures, which took a while. Since then I have written a Windows application that does very much the same thing, but with a few extra bells and whistles. One of the first things I do after starting my PC each day is to download the latest CET from the Met Office, but recently the daily estimated data that I religiously download hasn’t been updated, usually this is for just a day or so, but occasionally it can go on for a week or more. No big deal for most, but it bugs me so I fire of an email and get the usual reply:
Of course, if the Met Office did free up and make their NCM climate data public, then I could calculate the CET values myself, I could add the three numbers up and then divide by three, for the maximum and minimum temperatures.
I can’t afford to pay the Met Office to access the climate daily data for Rothamsted, Malvern (I had always thought it was from nearby Pershore College) and Stonyhurst, which are the three sites that are used to calculate the composite temperature that we know as CET. Knowing my luck, even if I did pay for the service, it would prove to be just as flaky as the present CET system is and fail to deliver that data as well.
I don’t know why I’m that worried so much about July 2017, because I already know that it wasn’t in any way exceptional, and to most people this rant will seem even more petty. But to me, their attitude with CET exemplifies how casually they treat the climate data they collect on our behalf. Sure they share some of the monthly data to comply with the data.gov decree of a few years ago, but the holy grail to anyone interested in climate, is daily data, for all UK sites, both the latest and the archived, and that they jealously guard.
I am afraid it looks like that I’ll be forever beholden to the Met Office for access to the latest CET data.
It’s now four days since the latest estimated CET series was updated, I’m betting that normal service won’t return till at least Monday of next week, we’ll see. Of course very few if anyone at the Met Office will ever read this article, but as far as I’m concerned, one of the main reasons that you write a blog is to get things of your chest, and that I have now done.
I did get some good news from them this morning concerning this problem which I informed the Met Office of in June. It wasn’t wrong as I first though, just a bit misleading, because you don’t always expect Sunderland to turn up in the Lake District even if there is a small hamlet with that name there. Anyway that shouldnt happen again, I’ve been reliably informed.
The summer has just been going down hill since the end of the first week of July sunshine wise. The sunshine accumulations for Jersey, which is currently the sunniest spot in the British Isles this summer, traces its decline quite well (fig 1). And if you thought it was bad where you are, spare a thought for those in Aviemore Scotland (fig 2), they’ve only seen 211 hours of sunshine compared to the 481.9 hours in Jersey, so perhaps the cooler and cloudier conditions further north and west have now spread across the entire country.
Temperatures are not doing much better either, here as a guide are the daily CET values so far for this summer (fig 3). The heatwave in the middle of June now seems just a distant memory, and there has been a definite cooling trend in the second half of July.
In Central England, 2017 still remains as the second warmest year to date (28th July) since at least 1772. July started so well, and early high CET values pushed 2017 briefly ahead of 2014 (fig 1), but as the CET cooled, 2014 overtook it is now 0.12°C ahead of it, both having now overtaken 2007, which had been warmest for a good part of this year till now. It’s hard to keep anomalies as high as +1.77°C for the rest of the year, and although 2014 did a good job, the running mean did drop back to less than +1.5°C by the end of August. But the way the summer temperatures have crashed over northwest Europe in recent weeks, has made me think that 2017 will find it very difficult to take the crown from 2014 as the warmest year since 1659, unless we see a very mild Autumn, and a December with temperatures as high as those in 2015.
The Met Office have just released the third edition of their State of the UK Climate for 2016. I always find that this a strange time to publish such a document, why not by the end of January? But once you have taken a look at the 59 page document, you’ll see that it must take a lot of work to put it all together, and I am just being my old curmudgeonly self.
There’s an interesting correction at the foot of the article about CET, which as most of you know is something I like to talk about at length. I find it amazing how people scour the climate datasets to pick out a new extreme, be it high or low, this one’s no exception – decades that span 10 years but don’t start on a year that ends in a zero aren’t what I would call a decade. It’s time to see what the Collins dictionary definition of a decade is:
A decade is a period of ten years, especially one that begins with a year ending in 0, for example 1980 to 1989.
So they are strictly correct, but personally I still like to think of a decade as a period that starts with a year that is divisible by 10 without a remainder, as the Collins dictionary definition suggests it is. Anyway back to the correction, what they were initially said was that the period 2007-2016 was the warmest ‘decade’ in the whole series back to 1659. Obviously they noticed that something was wrong, and said it was incorrect. I wonder why they didn’t just delete the reference in the HTML? I was certainly late on this one and never saw the original.
Just out of interest, here is a centred 10 year moving average of monthly CET values since 1900 (fig 2), and as far as I can see, the warmest decade was the one from 1997 to the end of 2006, when the mean 10 year anomaly for the first time just exceeded +1.0°C, since then, Central England has cooled, until about 2012, when annual anomalies increased sharply again. I think the reason they corrected the original article could have been for a typo, because 2007-2016 and 1997-2006 are just ten years apart. They can always ask me to provide them with graphs, I have a graph of CET values to suit just about any occasion, and the bonus is that I work cheap.
Here for completeness is the full 10 year series of moving averages since 1659 (fig 3).
And finally the infographic that the graphics team have produced to advertise the publication of the 2016 report is a bad idea to my mind. I love a well designed infographics – you only have to look at my blog to see that – and I do realise climate statistics are not the most exciting things to try to visualise, but a clear well constructed table or graph is really all you should need. Most of us have moved on from kindergarten I would have hoped, but they may have been instructed to produce a simple infographic for politicians the like of Michael Gove to understand what’s going on.
The cold few days at the end of June set 2017 back a little, but the warmth of the last week has now allowed it to move into second place in the warmest years to date, with a mean temperature of 9.81°C (1 January to 7 July) in the CET series. 2007 will continue to weaken in the coming weeks, so there’s every chance that it will get into first place, but it’s still being chased hard by 1846 and 2014. But 2014 is the year to beat though, because we already know that it’s the warmest year in the series that started in 1659, the two years are almost inseparable at the moment though, with just 0.03°C between them.
June 2016 could have finished in the top six warmest June’s since 1659 in the CET series, but was scuppered by a record cold end to the month. The month ended with a mean temperature of 16.04°C, with an anomaly of +1.99°C making it the 18th warmest since 1659. There were two new record high maximums for the CET series during the month, one on the 19th, and the other on the 21st, and a record low maximum on the 28th.