This forum is about wrong numbers in science, politics and the media. It respects good science and good English.
I would take the 'traditional' definition of the seasons applicable to the UK myself as being that used in the 'Gaelic calendar', which is still apparently used in modern-day Ireland. There is a tendency to completely ignore Gaelic or Celtic stuff in England, but it should be noted that the Celts were the dominant people in England before the Roman invasion. For people not familiar with the Gaelic calendar, it is described here:
Basically Spring is February, March, April. Summer is May, June, July. Autumn is August, September, October. Winter is November, December, January.
The start of spring corresponds to a Gaelic festival called 'Imbolc' on February 1st, described in this Wikipedia article:
According to the Wikipedia article the Imbolc date may have been significant since neolithic times ('neolithic' is the final part of the Stone Age, from about 4000 BC to 2500 BC in the British Isles), based on the alignment of ancient stone momuments in Ireland. So the idea that February 1st (or whatever the equivalent of Feb 1st was in ancient times) is the first day of Spring may have have been going on for several thousand years in the British Isles. [It might be that the climate was actually warmer back in those days.]
The astronomical seasons, based on the solstices and equinoxes, do look as though they are scientifically based, but I have a suspicion that it may be more connected with astronomy's less respectable cousin, astrology. The astrological year starts at the Vernal equinox with the first star sign Aires, and runs through to the next Vernal equinox finishing with the star sign of Pisces. Somebody may have just split up the astrological year into four equal periods, and then identified those four periods as being the four seasons.
Evidence that the astronomical definition of the seasons has never really been traditional in England comes from the definition of the 'close season' used in English football (soccer), given on this FA website page:
The close season was defined by the FA in 1891 as being from May 1st to Aug 31st. The four month window, provided so as not to interfere with the established summer sport of cricket, is consistent with both the Gaelic calendar seasons and the meteorological defintion of seasons, but not with the astronomical definition of seasons. Nowadays the FA's close season has been whittled down to just one month, the month of June.
Out of the three possible definitions of the seasons - the Gaelic calendar definition, the astronomical (which is possibly astrological) definition, and the meteorogical definition, I actually prefer the meteorological defintion.
Very interesting and informative stuff.
I will however note that over my dead body do I accept a spring that starts at the beginning of February. ;-)
I noticed a bit of supporting evidence in a recent blog post in the "Not a lot of people know that" blog for the idea that it might have been warmer back in neolithic times.
A Danish academic who has investigated ice cores from sites in Greenland claims it was perhaps 2.5 deg C warmer about 4000 years ago than the present day. If it was 2.5 deg C warmer in the British Isles as well, it would make a lot more sense for somebody to think that spring started in early February.
My concern was not so much the definition itself, rather who has the right to impose it.