Special Issue Article
1684: The coldest winter in the English instrumental record
Version of Record online: 26 APR 2011
Copyright © 2011 Royal Meteorological Society
Special Issue: Sixty five Years of Weather
Volume 66, Issue 5, pages 133–136, May 2011
How to Cite
Manley, G. (2011), 1684: The coldest winter in the English instrumental record. Weather, 66: 133–136. doi: 10.1002/wea.789
- Issue online: 26 APR 2011
- Version of Record online: 26 APR 2011
This article was previously published asManley (1975).
In England systematic daily meteorological observations, with instrument readings, were recommended soon after the foundation of the Royal Society in 1660; the barometer and the so-called ‘thermoscope’ were known and used a few years earlier. A brief account by the present author, with an illustration of the ‘scheme’ devised by Robert Hooke, can be read in Endeavour XXI for January 1962. Knowledge of the comparative intensity of heat or severity of cold, so far as it could be registered by thermometers, began to be acquired; and, however imperfect the instruments and techniques of measurement, once one can obtain a sequence of daily entries with wind and weather, standardisation can be attempted, with the result that a representative sequence of monthly means for England can be carried back to 1659 (Manley 1974, with sources and references – many used in this article).
Not that earlier efforts at comparison were lacking. From early medieval times we find accounts of the length of winter frosts in terms of weeks, or of the Church festivals, e.g. ‘a frost from the feast of St. Catherine to St. Valentine (25 Nov.–14 Feb.) 1433–34’. But we must always question just what was meant by ‘frost’. My own inclination is to think that for many it began when night minima in the open country fell, at least for several nights, noticeably below the freezing point; otherwise it is hard to understand such phrases as ‘a great frost which lasted for 15 weeks’ (1407–08).
Intensity began to be noted for example when ‘wine froze’ in Scotland in 1435; extent and persistence of ice began to be noted for example when the Thames was frozen and ‘bore waggons as far down as Gravesend’ in 1435 (Holinshed). Sometimes there might be a note on the persistence of lying snow; for example snow is said to have lain at Edinburgh from 18 March to 22 April (N.S.) in 1595. It begins to be possible to estimate the character of winter by comparing the reported dates of formation and break-up of ice on our larger inland rivers on which trade was increasingly important; for example, Trent at Nottingham, Ouse at York, Severn at Worcester, Thames at Oxford or London and we can compare these with similar events in northern France or Belgium. If only we had more records of the freezing of our lakes such as Windermere or the rarer occasions of ice on Loch Tay, Loch Lomond or even across the head of Loch Fyne, we could compare them with the occasions when the Bodensee (Lake Constance) or the Lake of Zürich were frozen. But as we do not hear of our men of Westmorland and Lancashire organising a religious procession across their frozen lake, as they did between Switzerland and Germany, records are lacking. Occasionally someone notes effects on vegetation, especially the more delicate evergreens (-rosemary, lamented by Evelyn in 1684, and, near Penzance, by Borlase in 1740).
Gradually the documents become more numerous; more men record more facts more precisely. Barker in 1740 noted that his pond at Lyndon froze 3in. thick overnight, and that this had only once been surpassed in ‘the Long Frost’ of 1684 when his grandfather (Whiston) recorded 3¼in. in a night. But we do not know the dimensions of the pond; presumably it was where the stock were watered. In 1740 ice was noted 22in. thick on the Eden in Cumberland, ‘two feet thick’ on the tidal Lune at Lancaster, and 26in. at Leyden in Holland. Ground frozen to a depth of ‘three-quarters of a yard’ was noted at Northenden near Manchester in 1684 and ‘three feet’ in Kent; again we could do with more detail. For comparison, ground was frozen 24 to 30in. deep in a number of places in February 1895, where there was little snow, and ice was noted upwards of 15in. thick; on the Cambridgeshire fens a measured thickness of 2ft. was reported (Bayard and Marriott, 1895).
Much can be learnt about 17–18th century standards of comfort from the indoor temperatures, frequently observed with the primitive thermometers of the day, some of which were over 2ft. long with bulbs upwards of an inch in diameter. These were understandably precious to their owners and were often hung indoors beside a window, or perhaps in a hallway. We can provide approxi-mate Fahrenheit equivalents, and from the early MS. journals we learn that in January 1698 at Oates near Ongar, John Locke, more familiar to us as the philosopher, recorded 38° in his parlour at about 10am. With big open fires and wide chimneys, one can understand high-backed chairs and footstools. During the phenomenally bitter easterly gale early in January 1740 (Gregorian dating) Jurin observed 34° in his London study at about 9am, and Short at Sheffield on the same occasion read his instrument on his staircase at 25°, and was very properly impressed by the fact that ‘even within a nobleman's library’ the reading was 29°. Frequently it was recorded that milk froze overnight indoors; more rarely ‘it froze beside the fire’, and one readily imagines the housewife or one of her maids coming down in the dark morning hoping that she could blow up the embers, and finding a skim of ice on the pan left ready for the cook beside the hearth. At Oxford in 1684, Wood recorded that his bottle of ink froze at the fireside on 12 January.
But in the popular mind the real test of 18th century severity was when the clergy and the more polite travellers were able to record that vessels were found frozen within the bedroom. The forbears of today's outspoken scientists were then to be found among the medical men who were wont to define the vessel with greater precision. Such events indeed are not quite unknown today. In December 1938 we found ice in the bedroom washbowl in a well-built and well-found hotel beside the Great North Road; and in 1947 after a night when the temperature in my fireless study fell to 19°, one of those memorably vigorous ex-service undergraduates came in to the morning's lecture with the news of the ultimate 18th century event that had just befallen him in his Cambridgeshire village lodging.
In the endeavour to establish for this country the longest representative series of temperatures in the form of monthly means that so far appears practicable, search of many sources enables one to carry a sequence back to 1659. Over England generally, the coldest winter since the beginning of instrumental observation was that of 1683–84, taking as ‘winter’ the three months December–February (all months and dates here are those of the Gregorian or ‘New Style’ calendar). It was appropriately known as ‘the Long Frost’; at London the Thames remained passable on foot for seven weeks, 2 January to 20 February. It does not appear that such persistence prevailed in any of the later renowned winters such as 1740 or 1795. And while each of the three months was not by any means the coldest on record (about nine Decembers, two Januaries and four Feb-ruaries were colder) the three together gave an overall mean about 1 degF below that of any other winter season in the series. Not the least difficulty in such seasons arose from the stopping of many water-mills; hindrance to working of ships' rigging in the bitter wind, and to water transport was a serious matter. We read much of this in 1740, when ships could not leave the frozen Tyne and the price of coal in London multiplied by eight; a further problem was that of getting the laden coal carts up the icy streets rising from the river. For 1684 perhaps the most familiar general account comes from John Evelyn's diary; we read how he ‘took coach to Lambeth’ along the frozen river.
We have enough in the way of daily weather observations for 1684 to go some way towards a reconstruction of the meteorology. For London we have John Downes' daily notes on the weather with thermometer readings; he was physician to Christs' Hospital. To them we can add the regularly-maintained journals by Gadbury and Ashmole. For Oxford there are the daily observations from 11 January onward by Robert Plot (in Phil. Trans. R.S.), and from the north of England, the valuable daily note of wind and weather kept by Christopher Sanderson of Eggleston in Teesdale, where he had his small estate 700 feet above sea-level. With these we can make use of a good many scattered descriptive notes, from various places.
Of the preceding autumn, September 1683 was an average month, unsettled westerly at first, improving and becoming anticyclonic 15th–20th with much cloud off the North Sea, rather warm and thundery 16th–18th in the south-east. Later, anticyclone-margin westerly weather until about 3 October, becoming quiet with cool nights and some morning fog in the South.
4–5 October saw more wind as a low came down with an arctic outburst behind it; in Teesdale snow on 6th, with hard frost to follow. From 9th–17th unsettled, cool, mainly north-easterly to east; moist westerly, 22nd–31st with a frosty night or two. As a whole, October was decidedly cooler than average, as it was, indeed, in 1974; but with very different consequences.
November was unsettled, cool west to north-west at first; snow both in Teesdale and London on 4th; thence unsettled and rather windy in north, quieter in south; brief snowfalls in Teesdale 14th–15th, then frost; sleet on 18th in south, then milder. Fog and frost in south-east at end of month, which as a whole was slightly cooler than average. Some of the general accounts, however describe ‘the frost’ as beginning in mid-November.
December began cold and frosty with a cutting east wind and snow showers in the north; early on 3rd, there was a northerly blizzard in Teesdale. Thereafter persistent cold northerly weather in the north; on 7th ‘a melting snow all day’ in London. On 10th to 11th very windy in Teesdale with ‘great thaw and rain’, followed on 12th by sudden hard frost, and snow most of the night on a strong northerly wind with much drifting. This snow reached London, and cold north-east to easterly weather prevailed 14th–21st; London had snow laying 15th–18th with ‘pinching hard frost’. Thaw however reached Teesdale 19th–22nd but ‘not enough to clear the snow from the higher hills’; these were cold rainy days in the south. And then a brief west wind spell brought thaw and fair weather in Teesdale 22nd–23rd, cloud and rain in London; but then, snow showers on a north wind ushered in a prolonged spell of fair north-wind weather with hard frost at night, in Teesdale from 24 December to 10 January, with two days on which snow showers fell. How well those who know Durham can picture the sequence today. Further south, Wood reports ‘a great deal of snow fell at Oxford’ on 25th; perhaps this provided the very heavy ‘Lorna Doone snowfall’ on Exmoor. Some snow fell on that night in London, but it does not appear to have been heavy. At Bristol on 19th, ‘very wet and cold with snow and sleet’.
1–10 January was described in London as ‘bitter hard frost, snow lies, some blew mist’; Evelyn has ‘the air so very cold and thick’. From 11th–19th, ‘hard frost, cloudy, mainly easterly winds’. Indeed in London from 10 January to 13 February the wind is noted as easterly on 22 days, north-easterly on eight, northerly on one and south-easterly on one; in Teesdale mainly northerly or easterly. Thaw in Teesdale 21–23 January, in London and Oxford on 20th–23rd give the impression of a low over France bringing in milder air for a time; but on the night of 23rd–24th cold cloudy weather returned, with snowfalls on 27th and showers thereafter; about 6in. fell in Teesdale. 4 February gave heavy snow in Teesdale, ‘about 27 inches in the plains’, after which the anticyclone built up again with sunshine in Teesdale and fog in London. Frost ap-pears everywhere to have been intense, 2–7 February. Bristol, recorded severe cold on 30 January.
On 15 February ‘began to thaw’ in Tees-dale (wind west to south); ‘rain and great west wind’ in London. We can compare the break-up of the 1947 frost, about 10–17 March, with the Fenland floods, the Furness snowstorm, and later the great westerly gale (at Mildenhall for example) while the cold persisted in Scotland. In 1684, the thaw about 15 February saw the break-up of the ice 15th to 16th on the Trent by Nottingham where the bridge collapsed. Cold returned to Teesdale on 17th–19th (north-east wind, with some snow) backing to south-east on 20th with ‘renewed thaw and great wind’. But in London and Oxford the 17th–19th gave rainy westerly weather, although ‘some rain and snow’ fell at Oxford on the night of 19th, before the general south-wind thaw on 20th. Such a sequence brings to mind the meteorology of the break-up of the 1947 cold. Ice on Tees broke at Eggleston on 21st, but not till 23rd at Barnard Castle six miles down river; and despite the thaw and rain, Thames was still ‘hard’ on 19th, although at Nottingham ice had broken on 15th to 16th in the earlier thaw.
According to the daily thermometer readings kept by Dr. John Downes, who had studied at Leyden, the greatest cold occurred on 15–16 January; and it was almost as cold on 2, 3, 4 and 6 February. His thermometer hung indoors, probably adjacent to a window following the prevailing custom, used a degree whose equivalent was about 2.4°C (4.3°F) and these were read above and below a zero which in 1684 was probably about 48°F. The fact that in mid-January the spirit fell ‘within the ball’ indicates something of the order of 25°F indoors (mid-morning), and probably below 10°F outdoors. This is of the same order as the lowest daily means recorded at Greenwich (e.g. 15 Jan. 1838, 6·2°F; 20 Jan. 1838, 10·7°F and 8 Jan. 1841, 12·8°F). Lowe, writing in 1870 quotes for 1684 a record of –8°F at ‘London’, but we do not know the source, the conversion factor, or the location of the instrument. For comparison there are reports of –6°F in Marylebone on Christmas Day 1796, and a more acceptable –4°F at Chiswick on 20 January 1838. We can read that in 1684 ‘large forest trees were split by the frost’ near London in Enfield Chase. From a garden at St. Boswells in the Tweed basin it was stated that ‘large trees were giving off reports like the firing of a gun’ in January 1881 when –16°F was observed at Kelso and stagnant cold must have lasted for many hours.
In Teesdale Sanderson noted now and then the thickness of ice formed ‘where John Aislaby gets his water’; on 12–13 February ice formed ‘two inches thick overnight’. He then adds ‘it was in my gazette that at the Downs the water was frozen a mile into the Sea, which was never known before’. This would probably have occurred during the latter part of January. From other sources we learn that for 14 days packet-boats with the mails could not leave the Belgian coast and that ‘the sea was frozen for two leagues off Caen’. Near Manchester ‘it did freeze ice more than half a yard thick, and some ice continued till 25 March’.
We learn that the cold air was still persisting in the north; on 24 February it became cold again in Teesdale with snow and hard frost at night, although in the south it became a rainy afternoon after morning frost. But from 27 February to 6 March there was persistent thaw with mild winds and a good deal of sun-shine; the last of the ice on Tees was carried away on 28th and evidently the ground (at 700 feet) became clear of snow. But it was freshly powdered again on 7 March and thereafter until 26th, north-east winds prevailed, with snow showers noted on eight days, a 6-in. fall on 22nd and the Tees frozen over again. Sanderson also noted a violent storm in Yarmouth Roads on 31st; evidently he had news of shipping through Stockton and Yarm. His son, as an undergraduate at Cambridge (St. Johns), went by sea to King's Lynn and thence by river.
April of 1684 seems to have been a coolish to average month, cold at first, then generally mild, though no marked warmth. Teesdale had snow showers on 26th; a quite normal occurrence around the time of what Westmorland knows today as ‘the gesling blast’.
Reviewing the season as a whole, over much of the country the mean tempera-ture of October 1683 was fully 4 degF below the average prevailing in this century; November 2 deg below, December 7 deg below, January 10–11 deg, February 8 deg and March 4 deg below. Towneley's observations near Burnley indicate very little precipitation in December and January as we might expect, followed by a very wet February (Old Style months); but we also cannot be sure how much snow his roof gauge caught, or how far across the Pennines the frequent north-easterly falls carried. From the depth of frozen ground near Manchester there must have been an appreciable period with little or no snow cover, and the account of the intensity of the frost at Lancaster supports the likelihood of a good deal of clear, dry weather west of the Pennines.
We know that severe cold prevailed over much of west Europe; not only the greatest rivers, but the Lake of Constance itself was frozen. But in France the winter of 1709 was regarded as more extreme and is so rated by Easton. Moreover Derham recorded in early January 1709 a lower temperature than on any previous occasion in his own observations and thought it colder than on any day in 1684. On the other hand, in 1709 it was particularly noticed that in Cumberland and Scotland the lakes and rivers did not freeze, and at London the Thames was not frozen for anything like the spell in ‘the long frost’ of 1684.
The weight of the evidence suggests for 1684 a situation like 1963 but developing a little earlier; with high pressure at first probably towards Iceland, then devel-oping and forming a ‘blocking high’ giving clear skies and light northerly winds for that critical fortnight, in late December–early January when there was a widespread snow-cover at least over Scotland and the east and south of England, if not further. During the short-lived anticyclone over Scotland in mid-November 1919 the maximum at Braemar on the 16th was 12°F and the minimum was –10°F: this gives an indication of the possibilities, given quiet weather and clear skies after a snow-cover during the period roughly November–February when out-going radiation is in excess. The anticyclonic cooling over the Midlands in mid-January, 1940, although there was no great snow-cover, can also be cited. In 1947, the onset of the cold came later in January, the wind throughout February was more persistently easterly and the month was very cloudy. But in a clear-sky interval after snowfall the minimum of –6°F at Houghall near Durham on 6 March 1947 still stands as the lowest yet recorded for England in that month. We can be fairly confident that temperatures of that order would have occurred in Teesdale in January–February 1684. The Kentish reference to frozen ground is less easy to explain unless the snow was either somewhat lacking compared with the west country, or was much drifted. Undoubtedly too the impression is gained that both the North Sea and the English Channel were somewhat colder than that which we should expect today; but we have yet much to learn about past vicissitudes of surface temperature over our northern seas.
There is some prospect of making a more detailed interpretation of the meteorology. The Admiralty archives include log-books, which were sometimes kept in ports, from 1669 onward; hence there might well be some reports from Aberdeen or Leith, Liverpool, Plymouth or Dublin for example. To these it looks as if one might add Dutch, north German and Danish observations. Indeed the meteorology of the whole period from about 1667 to the beginning of the next century appears worthy of careful study, comprising as it did some or the greatest extremes of which we have knowledge. There is still abundant oppor-tunity for the amateur with time, energy and a zest for the archives – not forgetting, as a by-product, the revelations of the personalities of his assiduous forefathers, who in the 17th century lived through what appear to be short-period, wide-amplitude oscillations about which we need to know more.
- 1895. The frost of January and February 895 over the British Isles. Q. J. R. Meteorol. Soc. 21: 141–160. ,
- 1974. Central England temperatures: monthly means 1659 to 1973. Q. J. R. Meteorol. Soc. 100: 389–405.
- 1975. 1684: the coldest winter in the english instrumental record. Weather 30: 382–388.