- 0 -
Hams, R. 0. (1970) Notable British gales of the past fifty years. Weather, 25, pp. 5748
McCallum, E. (1990) The Bumsâ Day storm. Weather. 45, pp. 168-173
Smith, S. G. (1982) An index of windiness for the United Kingdom. Meteorol. Mag., lU, pp.232-247
_ _ _ _ _ Smoothed using 7-year binomial filter
I 1 â , . . I * . . . I * . J * I . â S ~ I ~ ~ * . â ~ J - J I . , a * I . , - * â â ~ â , â ~ * ~ ~ 1 * ~ * ~ 1
1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990
FROM THE ARCHIVES
Some food for thought from our December 1963 issue.
âDOES THE WEATHER FORGET THE PAST?
By L. C. W. BONACINA
N Readerâs Forum (Weather, January 1962) I raised a question arising out of Mr E. I Knightingâs article (Weather, Septemkr 1961) on âNumerical weather predictionâ. Mr
Knighting says that for numerical prediction purposes the weather may be considered
probably independent of the past at long range, but partially dependent at medium range,
say from 5 to 30 days ahead, and closely dependent at short range.
As the idea of independence in this context had always been at variance with my
thinking on the subject I asked for some elucidation of the argument but the reply did not
meet my point.
More recently, at the Symonsâ Memorial Lecture, presented at the Royal Meteorolo-
gical Society on 27 March 1963, Professor J. Smagorinsky remarked that âthe weather
forgets the pastâ. But does the weather really forget its past? I simply cannot believe that it
does. Nor do I think that the question has been answered.
Suppose we define the weather of a particular day in terms of statistics such as
maximum and minimum temperatures, amount of precipitation, duration of sunshine, etc.,
and consider the time-scale for Wday forecasting. The argument according to Mr
Knightingâs article would be that the weather on the first day of a month has had time to
forget (after about 30 days or so) what it was on the first day of the previous month. But, on
this time-scale, it will still be âmindfulâ of what was happening in the middle of the previous
month, which in turn will have been influenced by the events of the first of the month. Thus
the argument is paradoxical; it means that some measure of dependence is passed on in
Suppose, on the other hand, we describe a dayâs weather as a sequence of events. Such
a sequence would be part of a longer sequence and, since we could never call any particular
phase an initial state, we again find an element of perpetual dependence on weather of the
past -even though we may not be able to trace the linkages in the complicated atmospheric
So, whether we define weather as a statistic or as a sequence of events, it appears that
the weather is always influenced by a vast âmemoryâ.
Is a âmemoryâ responsible for the curious tendency for weather excesses to recur in the
same geographical area after days, weeks or even months while in the intervals there may
be a complete reversal of type? Such repetitions can only be explained by hidden but
persistent patterns which Sir Graham Sutton calls âcentres of actionâ in his book The
Challenge of the Atmosphere. These patterns, which are masked by the more obvious
systems of greater concern to the short-range forecaster, may also account for the late and
sudden amval of a big storm after its initially obvious portents have declined and given the
impression of a false alarm.
I well remember during cold snowy weather in mid-December 1927 an authority
telling me that real winter was at hand and a foot of snow was not unlikely in London. In
fact, after a hazardous glazed frost, a south-west wind ushered in a thaw just before
Christmas. But the blizzard, not to be cheated of its prey, swooped down on southern
England on the afternoon of Christmas Day as heavy rain in an east wind suddenly turned
to snow which caused widespread chaos.
I also recall that late in May 1911 a heavy thunderstorm had been foreshadowed in the
London area. This receded for a time, as though to muster its forces, before breaking out
on the hot afternoon of the 31st. I shall never forget the temble darkness before it struck.
Several persons were killed by lightning near London and the storm figured prominently in
the meteorological literature of that period.
Great storms have a curious way of reaching a climax in the same geographical area.
The tempest which devastated millions of trees in the state forests of north-east Scotland at
the end of January 1953 had a formidable forerunner in December 1952.
We all remember the Lynmouth disaster of August 1952, but probably few remember
that in April of that year the Barnstaple district had suffered serious flooding after thundery
rain. Was this sequence an accident or were the two storms connected by a hidden link
spanning the four-month gap?
Was it mere chance that the memorable snowstorm which struck western and south-
em Ireland in late January 1917 was followed by a still more furious snow-storm early in
April of that year?
Tocite more recent events: the hard winter of 1962-63 was notable for its succession of
paralyzing blizzards in south-west England and as early as mid-November an unusually
heavy fall of snow had occurred in parts of Somerset and Devon. Again, was this just a
coincidence or was a controlling pattern for the winter established in November, well
before the onset of the main cold spell?
These particular incidents are but a few of the countless sequences which suggest that
weather patterns have some sort of âmemoryâ -even though we mortals d o not know how it
SEEN FROM ABOVE
From cdour photographs by R. E Saunden
(Top) The upper surface of a sheet of altostratus seen from an aircraft at about 9km over Austria on
4 October 1989. (Bottom) Close-up at 3 feet of 'mini-scallops', about 2cm wide and Icm deep,
formed in calcium carbonate calcite-laden water as it flows downhill near Hieropolis (Turkey) on
2 October 1989.