In article ,
John Hall writes:
In article ,
Ian Bingham writes:
Thanks for these figures, John. Very interesting. What strikes me
about them though is how un-smooth they are for an average of 235
years. For example, in the second half of February the temperature
rises from 3.73 to 4.45 in just 3 days. This rather suggests that the
winter temperature profile has changed during that period, and probably
several times. For example it changed after Buchan's time, resulting
in the poor fellow getting an undeserved raspberry from the
meteorological cognoscenti.

Yes, I'm sure that there must have been changes. I'm going to try
looking at the last hundred years or so, and see what difference that
makes. (I think that a 30 year period, though fine when looking at
monthly means, would be too short for daily means, as there's a greater
possible range of daily values than there is for monthly ones.)

OK, here are the corresponding figures for the period from 1901, and
also the change in the means for that period with respect to those for
the whole period. (Clearly those changes would be much bigger if one
compared them with the pre-1901 period instead.)

Jan Feb Dec Jan Feb Dec
1 4.09 4.16 5.55 0.80 0.24 0.45
2 4.36 4.13 5.52 1.06 0.27 0.50
3 4.23 4.38 5.58 1.05 0.49 0.53
4 4.13 4.31 5.37 0.75 0.32 0.29
5 4.01 4.35 5.21 0.79 0.28 0.07
6 4.21 4.45 4.80 0.85 0.34 -0.01
7 3.97 4.26 4.87 0.79 0.03 0.28
8 4.17 4.25 4.82 0.96 0.27 0.45
9 4.31 3.98 4.62 0.97 0.05 0.44
10 4.40 4.06 4.73 0.98 0.14 0.47
11 4.21 3.89 4.77 0.93 0.09 0.58
12 3.84 3.78 4.72 0.79 0.07 0.35
13 3.76 3.58 4.84 0.48 -0.04 0.35
14 3.73 3.35 4.89 0.42 -0.37 0.40
15 3.96 3.74 4.66 0.77 -0.23 0.27
16 3.95 3.80 4.68 0.61 -0.09 0.36
17 3.76 3.72 4.79 0.43 -0.01 0.37
18 3.79 3.80 4.55 0.49 0.07 0.17
19 3.94 3.98 4.21 0.5 0.23 0.16
20 3.94 3.95 4.12 0.65 0.22 0.18
21 4.05 4.17 4.21 0.58 0.03 0.27
22 3.93 4.29 4.44 0.37 -0.04 0.57
23 3.72 4.40 4.67 0.22 -0.05 0.85
24 3.80 4.22 4.49 0.29 -0.17 0.72
25 3.75 4.16 4.24 0.36 -0.32 0.80
26 3.73 4.38 4.26 0.27 -0.07 0.85
27 3.76 4.58 4.29 0.14 -0.01 0.87
28 3.80 4.69 4.21 0.26 0.09 0.81
29 4.00 - 4.13 0.28 - 0.49
30 4.00 - 4.14 0.20 - 0.55
31 4.13 - 4.07 0.17 - 0.63

Early January has warmed by much more than late January. The period from
the 12th to the 28th is now the coldest, with a brief warmer spell from
the 19th-22nd. The 1st-11th is the warmest part of the month. So the
recent phenomenon of mid-January being milder than early or late January
is still not apparent in these figures.

February has warmed up by much less than the other winter months (some
periods even having cooled), showing a strenthening of the trend
established in late January. I wonder if that could be because
easterlies have perhaps warmed up less than northerlies (where less
Arctic sea ice compared to the 18th and 19th centuries could be a
factor); prolonged easterlies are much more common in late winter than
in early winter. The coldest period is from the 9th to the 20th, ie much
the same as before. That period is now comparable in coldness to the
coldest part of January, which used not to be the case. Indeed
Valentine's Day is now on average easily the coldest day of the winter!

December has also warmed substantially, though not as much as January,
and especially from about the 23rd onwards. The end of the month is
still the coldest period, though by less than before. There's still no
sign of a post-Christmas warming. Indeed the 22nd-24th is on average
warmer than the following days.

It's noteworthy that 4 of the 8 sub-zero CET Februaries have occurred
since 1901 (all since 1947), compared with only 3 of 20 sub-zero
Januaries (and now 1 of 7 sub-zero Decembers), which is consistent with
the trends identified above.
--
John Hall

"The covers of this book are too far apart."
Ambrose Bierce (1842-1914)