Summer Index: 1929-2016

[xmetman]

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

Worst Summer

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!

[John Hall]

Thanks, Bruce.

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!

Yes, from 1976 onwards that odd-even summers thing seems to have broken down. Looking at your data from 1967, there seems to be a tendency for "good" and "bad" summers to occur in "clumps", though, particularly for "good" summers. You have to go back to 1973 to find the last time that there was a single-member run of better than average summers (if that can be called a run).

I hadn't realised until now just how "good" the summers of 2013 and 2014 were. (I put "good" in quotation marks because for those of us who don't like it too hot they weren't good at all.)  

[Phil Officer]

Bit confused how 1970 scores 0 summer index in the best summers list.

[xmetman]

As I point out in the article, the summer index scale runs from -48 (worst) to +48 (best). Zero would indicate an average summer. The list is just the top ranked summers from high to low.

[xmetman]

[Smoky Bacon]

I think you have to be a bit careful with lumping the whole of the UK into this "best summer" category.

I remember a holiday in Northwest Scotland during the summer of 1977, and it was very hot and dry with drought conditions in certain areas. Talking to the locals (admittedly only b&b owners, really) it was evident that 1976 had not been particularly good over Scotland, and certainly nothing like the weather over much of Southern and Central England.

1977 seems to have been Scotland's 1976.

Similarly, the 1983 summer in Wales (where I was living at the time) was hot and very dry, and felt like 1976 repeated. In fact our mains water was switched off between 10a.m. and 4p.m. in order to cut demand during the hottest spells in June and July.

[xmetman]

That's the one advantage of the MO 1910 data series is that it's split nationally and regionally (see map courtesy Of the Met Office). 

I can look at the Summer index for any region but I thought that I would look at the UK as a whole for the article in my blog.

The one thing I would urge the Met Office (not that the person in charge of climate services is reading this) to do is to agree with Met Éireann (the Irish National Meteorological Service) for permission to use their climatological data, the weather knows no boundaries, and after all this is the 21st century.

[Tudor Hughes]

    The summer of 1983 was hot everywhere.  In my own records, starting March 1983, July 1983 is my 2nd hottest month, beaten only by July 2006.  Part of the reason was that the hot spell coincided with the calendar month whereas in 1976 the hot spell included both June and July.  I wasn't recording in 1976 so I can't prove this but I suspect July 1983 was warmer than any month in 1976.

     One summer I can remember which showed a great diversity over the UK was 1968.  In the SE it could only be called dreadful.  Cool NE'lies and heavy rain whereas in the west and north of Scotland it was fine, dry and warm.  The circulation was upside down.

Tudor Hughes, Warlingham, NE Surrey, 557 ft, 169 m.

[xmetman]

Tudor 

Nothing wrong with your weather memory.

Quite correct the summer index for Southeast & Central Southern England for 1968 was -34.

I've attached a ranked list (mean temp) of July's in Southeast & Central Southern England

Bruce.

[John Hall]

An interesting table.

Going back to the original table, one thing it shows at a glance is how much our summers have "improved". In the most recent couple of decades, the dominant colour is red, whereas early in the period blue dominates.

[xmetman]

I'm surprised Dawlish didn't notice that...

[Martin Rowley [West Moors/East Dorset]]

... for many years I've kept a 'summer index' based purely on sunshine & mean maximum temperature - in my view the two parameters that the great mass of people will, consciously or otherwise, base their assessment of summer (JJA) on. It shows a distinct improvement in recent decades for the two areas I maintain an interest in, as John has noted.

http://booty.org.uk/booty.weather/metinfo/NFL/Summer_Indices.htm

Martin.

[xmetman]

Martin

I never realised that you had keep your own summer index. Yesterday I gave in and decided to suscribe to the Weather Magazine so I could read the full article that John had given me the link to. Did you come across Murray by any chance? On looking up his articles in a search of the archives, he seemed to be interested in the same kind of weather stories that I am.

Bruce.

[Phil Officer]

I was pointing out 1970 is ranked 26th best along with 9 others on 0. It is the only one in those 10 to have numbers in the temp quintiles, precip and sunshine terciles.
Presumably there is a cancelling out to give a score of 0.
Most unusual?

[John Hall]

The contrast between June and July in that year is very striking. Perhaps the weather gods were passing their opinion on the result of the General Election held in June! 

[xmetman]

Follow the link and do the maths hopefully I programmed it correctly☺

[Tudor Hughes]

I can remember waking up each morning in early August 1968 to the sound of heavy windless rain.  Also, in September there was the tremendous 2-day downpour with thunder and lightning on the 15th/16th that gave me about 175 mm according to the map but I wasn't recording then.  Bob Prichard once made the interesting remark that if 1968 had occurred 30 or 40 years later the media would have scuttling about like headless chickens yelling "the weather's gone mad" or similar.  ISTR that the anomalous circulation was put down to cold in NW Russia with a lot of unmelted snow.

      You make kind remarks about my weather memory but I was completely wrong about 1958.  We had the Horsham Hailstorm (5th Sept).  No hsil here but torrential rain and the best lightning display I have ever seen.  Also I have a large number of B & W photos of steam trains on the Oxted line.  All this combined to make me think it had been a fine summer, but the figures show it was cool, cloudy and wet.  15-yr-olds are indestructible and not deterred by poor weather.

Tudor Hughes 

[Dick Lovett]

Bruce

I recently came across this 1998 article about Summer indices in Weather that you will probably find of interest.

http://onlinelibrary.wiley.com/doi/10.1002/j.1477-8696.1998.tb06332.x/abstract

Dick Lovett

[xmetman]

Dick

Thanks for that article I'll have a read. 

Bruce.

[Phil Officer]

You programmed correctly Bruce, but I don't see the point of including any date pre-1929 when you can't find a summer index for those years.
Your tables should read best and worst summers 1929-2016.

That's why I was a bit puzzled at first to see 1970 pitched in the middle of years with no sunshine data.

Sorry to be a bit pendatic.

[xmetman]

Hi Phil

Yes you're right about the heading should read 1929 - 2016, but the sunshine data is solid back to 1929 and theres no missing data for 1970, its just a zero summer index for that year i.e. average.

I don't know when sunshine recorders started being used, but the glass sphere looks a Victorian idea to me. It probably starts so late in the MO data series because of the number of stations that measured it was small.

Bruce.

[Ron B]

According to Brazells 'Climate of London sunshine records began in 1881,so obviously the Campbell Stokes recorders were invented by then .....

RonB

[xmetman]

Ron

It maybe that there is plenty of available sunshine data before 1929 and the Met Office just need to digitise and grid it.

Bruce.

[Julian Mayes]

Hi Bruce / all,

I understand that it is indeed just a case of digitising and gridding the observations before 1929. The data is all sitting there published in the back issues of the Monthly Weather Report which also started up about the same time as the C-S recorder. Of course, if the Met Office digitised the main data table of the MWR as well as just the front covers as at present, we would have a kind of digitisation by the back door. It dismays me that the Internet is a source of such plentiful current weather info and so little really long term obs data.

Julian

[xmetman]

Julian

I totally agree with you. If the Met Office put just 10% of the money it spends on trying to forecast climate change in the future, into digitising climate data from the past (before 1910) we would be in a lot better position to assess recent changes in the UK climate by comparing them to a much longer and fuller climate series.

Rainfall is a perfect example. The British Rainfall Organisation encouraged and published detailed records that started in 1860. The Met Office have published the PDFs but I thought getting at the data would be far more important, and in these days of sophisticated OCR should make the task a little easier. It is a disgrace that the Met Office have sold the public a lie, by churning out the stock phrase "since records began in 1910".

Bruce.