The Coldest Days on Record

[John Hall]

I recently bought a second-hand copy of HH Lamb's The English Climate, published in 1964. This was the second, heavily revised, edition of CEP Brook's book published in 1954. Appendix 2 is a "Calendar of Historic Weather Events since 1950". I realised that with its help I could probably compile a short list of the coldest days in the London area since reliable thermometers became available in the early 18th century. By "coldest day", I mean having the lowest maximum temperature. So here they are in chronological order:

11 January (New Style), 1740. "Afternoon temperatures in Holland -20C, in London about -9C. E gale." If I correctly remember what Gordon Manley wrote about this day, there were no outside thermometers in the London area at the time. They were still too expensive and fragile to be risked outside and were only hung indoors at this time. The London temperature was estimated based on reading in Holland, where there were some outside thermometers. To me, given the strength of the wind that -9C looks as if it's a fairly conservative estimate.

20 Jamuary, 1838. "Lowest [minimum] temperatures of 19th century in London: -16C reported at Greenwich about sunrise, -20C at Blackheath, -26C at Beckenham. Temperature in Greenwich -11C at midday." It seems likely that the day's maximum would have been several degrees high than this. With winds probably light and skies clear, judging by the low minima, the fact that proper screens for thermometers were yet to be in use may have depressed the values obtained.

14 December, 1890. "Coldest known December day in London, temperature never rose above -6C."

19 January, 1963. "Coldest day [of that winter]: afternoon temperatures of -7C in Surrey with E gale."

12 January, 1987. Maxima between -6C and -8C over much of England, with -9.1C at Warlingham.

From this short list, it looks as though the very coldest days mostly occur around the middle of January, and that the lowest possible maximum in the London area is close to -10C. That 1987 value at Warlingham may the lowest maximum in the London area that we can be absolutely certain is accurate, since the lower value in 1838 was a midday rather than a maximum temperature.

[Scott W]

Luke Howard recorded in Plaistow on February 10th 1816 a high of -6.7C (converted from Fahrenheit). 

https://wansteadmeteo.wordpress.com/2016/01/06/the-night-temps-fell-to-20c-in-east-london/

[John Hall]

Thanks, Scott. Very interesting.  Lamb seems to have overlooked that cold spell, as there aren't any entries at all for February 1816. The coldest ever February day in London?  And with a SW wind on the day in question, though I imagine it must have been very light.

Do you know what units he was measuring rainfall in? Not mm, surely? It looks as if he has fallen into the trap of quoting a value to a totally spurious degree of precision, unless someone has converted his original entry from one set of units into another and fallen into the same trap. I suppose that's likely, as the temperatures must have been converted from Fahrenheit to Celsius (and the -6.7 can't be relied on to the nearest tenth of a degree).

I wonder what rounding policy Lamb used for the temperatures he quotes - to the nearest degree? As he gives the Fahrenheit values as well, I could check.

[Scott W]

John, he used inches (1.38in on the day in question) - it was my own clumsy conversion. I've attached the original printed return 

[Tudor Hughes]

    The -9.2°C on Cold Monday 12 Jan 87 is my 15 minutes of fame.  I had the day off work (for other reasons) and was out to the screen "every 5 minutes".  Why doesn't it go up?  Brilliant sunshine - light wind, but a snow cover (about 10 cm).  My guess is the max was even lower a few miles to the SSE at the top of the Downs (877 ft) such was the lapse rate.

       This figure is so exceptional that it could statistically be regarded as not part of the set and therefore caused by a different mechanism from a routine easterly bone-chiller but the synoptics were not particularly abnormal - just "helpful".  The next lowest max I have recorded is - 5.8°C on 7 Feb 91.

       Strictly speaking my house is not in London by half the width of a road  although my phone is an 0208 number.  The boundary's a bit fuzzy.

Tudor Hughes, Hamsey Green, Warlingham, Surrey, 557 ft, 169 m.

[John Hall]

Scott, thanks. Comparing the original with the "converted" version, it looks at though in your converted version the temperatures have somehow slipped down a row, as the lowest Fahrenheit temperatures seem to be on the 9th rather than the 10th.

I wonder how Howard measured rainfall, especially when it was melted snow. That heavy fall probably equated to over a foot of snow, so even today it would be quite hard to get an accurate rainfall equivalent, especially if it was drifting.

It belatedly occurred to me that Lamb might have been aware of Howard's temperature readings, but decided for some reason that they couldn't be relied on. Maybe he had reason to think that the thermometer was inaccurate or the exposure unsatisfactory even by the standards of the time (too close to the ground, perhaps?).

I have a copy of the daily CET series in spreadsheet form, and am tempted to analyses that to see which days come out coldest. I think only the mean of max and min is given for each day, though, so they wouldn't be directly comparable to the London values I've given even apart from the difference in location.

Tudor, I too was struck by how cold that 1987 day was in spite of the wind from the Continent being fairly light. I wonder what the temperature might have been had it been stronger, as it was two or three days later.

When I referred to the "London area" I intended it to include the inner Home Counties as well as the modern region of Greater London. I suppose the height of Warlingham may have reduced your temperature by a bit, but temperatures of 7-8C seem to have been widespread throughout SE England.

[Scott W]

On Saturday, February 11, 2017 at 10:42:13 AM UTC, John Hall wrote:

Scott, thanks. Comparing the original with the "converted" version, it looks at though in your converted version the temperatures have somehow slipped down a row, as the lowest Fahrenheit temperatures seem to be on the 9th rather than the 10th.

I wonder how Howard measured rainfall, especially when it was melted snow. That heavy fall probably equated to over a foot of snow, so even today it would be quite hard to get an accurate rainfall equivalent, especially if it was drifting.

It belatedly occurred to me that Lamb might have been aware of Howard's temperature readings, but decided for some reason that they couldn't be relied on. Maybe he had reason to think that the thermometer was inaccurate or the exposure unsatisfactory even by the standards of the time (too close to the ground, perhaps?).

Thanks, John. I'll look into it. Having read the opening section of Howard's Climate of London his method of data collection was painstaking - he had several gauges at Plaistow and Tottenham and I believe it was his suggestion that the Royal Society's gauge at Somerset House be moved to a more suitable location, together with the exposure of their thermometers. Though he didn't have a set of guidelines at the time being a chemist by trade I'm pretty sure he would have been as careful with measurement as possible, to the point of obsession?

[Richard Dixon]

On Saturday, 11 February 2017 02:23:11 UTC, Tudor Hughes wrote:

    The -9.2°C on Cold Monday 12 Jan 87 is my 15 minutes of fame.  I had the day off work (for other reasons) and was out to the screen "every 5 minutes".  Why doesn't it go up?  Brilliant sunshine - light wind, but a snow cover (about 10 cm).  My guess is the max was even lower a few miles to the SSE at the top of the Downs (877 ft) such was the lapse rate.

Presumably very cold (snowy) surface and lack of mixing led to a daytime surface "cold pool" continuing from the night in the absence of any strong heating? What were the temperatures the night prior?

Richard

[John Hall]

Wouldn't your suggestion imply that Warlingham would be warmer than low-lying spots rather than being colder, as it actually was?

[John Hall]

Scott, I hope I wasn't too rude about the lack of rounding in the table of data that you provided. I'd have been more tactful had I realised that it was you who had done the conversion. I'm afraid that quoting values to a far greater precision than their accuracy warrants is a particular bete noir of mine. ISTR that my physics master at school impressed on us that it was a cardinal sin. :)

I've just been looking at Manley's two papers on the CET, in which he described the data sources that he used. I'd hoped that he might mention Luke Howard, but unfortunately he doesn't, although he mentions various other sources of data from London. In Climate and the British Scene, Manley mentions only his work on classifying clouds. Is it possible that Howard's journal with his readings had been lost, and only rediscovered within the last few decades? It doesn't seem very likely, especially as Manley himself unearthed a lot of old weather logs.

[Tudor Hughes]

On Saturday, 11 February 2017 13:15:49 UTC, Richard Dixon wrote:

      The min the previous night was -12.3°C with a 10 cm snow cover and it was my second lowest temperature.  I don't think it was a lack of mixing but rather the opposite.  There was an almost dry-adiabatic lapse rate from the surface upwards, an extraordinary situation for such a cold day.  From "Weather Log" ( in Weather) I have extracted the following temperatures from the midday Crawley ascent:

                          Surface       -7

                          1000 mb    -10

                            850 mb    -20              Computed 1000-500 mb thickness  498 dam.   The partial thickness 1000-700 mb must have been new record but I haven't worked it out.

                            700 mb    -30

                            500 mb    -39

Tudor Hughes

[Scott W]

On Saturday, February 11, 2017 at 7:07:19 PM UTC, John Hall wrote:

Scott, I hope I wasn't too rude about the lack of rounding in the table of data that you provided. I'd have been more tactful had I realised that it was you who had done the conversion. I'm afraid that quoting values to a far greater precision than their accuracy warrants is a particular bete noir of mine. ISTR that my physics master at school impressed on us that it was a cardinal sin. :)

I've just been looking at Manley's two papers on the CET, in which he described the data sources that he used. I'd hoped that he might mention Luke Howard, but unfortunately he doesn't, although he mentions various other sources of data from London. In Climate and the British Scene, Manley mentions only his work on classifying clouds. Is it possible that Howard's journal with his readings had been lost, and only rediscovered within the last few decades? It doesn't seem very likely, especially as Manley himself unearthed a lot of old weather logs.

Not at all - I regard any feedback as a positive. I think the digitisation of volumes by Google and others has enabled access to long forgotten (or at least incredibly hard to access) books. Original copies of Howard's Climate of London would set you back hundreds of pounds - but you can now leaf through it on Google Books.

In terms of his work being used I noticed that Tony Chandler dedicated his 'Climate of London' to the memory of Luke Howard - I would have thought that he considered a lot of Howard's data when reaching his conclusions.

I believe there's a mountain of old weather logs in the Met Office library, including many clergy rainfall records, that would make fascinating reading from a local climate point of view.