.... to add to the analysis elsewhere in this thread ...

I maintain an 'index' of summer (JJA) against winter (DJF) based on a
combination of Kew Observatory (various screens) & Heathrow, using a
mix of maximum temperature & sunshine for summer and mean temperature
for winter.

Lots and lots of fudging I'm afraid, given the changes in
instrumentation, exposure (North Wall screen for a lot of the early
Kew stuff), and of course we have 'artificial' changes with regard to
sunshine due to the various clean air laws and the inevitable
urbanisation of the record ... but ... I find the indices give a
useful summary of each season as defined above.

(1): Looking first at the 10 coldest winters since 1900 in that
series:-
... one (1906/07) was followed by the one of the 'worst' summers in the
series, and in the CET record, it was the 4th coldest summer [ in that
part of the series where figures are logged to the nearest tenth
degC].
... one (1946/47) was followed by a summer that was more than one
standard deviation 'better' than the mean, *but* only just & certainly
outside the 'top-10' of such fine summers.
... the overwhelming majority (8 out of 10) were all within one
standard deviation of the all-series mean.

(2): Looking at the top 10% of 'fine' summers, a total of 14 events
(based, remember, on maximum temperature & sunshine - i.e. the factors
that in my view 'govern' how people respond to the season overall),
then :-
... 10 were preceded by winters with a mean temperature value within
one standard deviation of the whole-series mean.
... 3 were preceded by winters +1sd of mean
... 1 was preceded by a winter +2sd of mean

NONE of the 'fine' summers as defined was preceded by a winter with a
value lower than -1sd of the all-series mean (i.e., a 'noticeably'
cold winter).

Taking (1) and (2) together, it would be remarkable if the forecast by
PWS were to verify as presented - i.e., beating the 1976 mean
temperature (which is number 1 in my list, by a decent margin), and
might aptly justify the 'exception that proves the rule' cop-out :-)

Martin.


--
Martin Rowley
West Moors, East Dorset (UK): 17m (56ft) amsl
Lat: 50.82N Long: 01.88W
NGR: SU 082 023