But how did the weather forecast do at predicting what the Christmas Day weather would be? I've been checking the BBC's online weather forecast for London every day for the last two weeks to see how good a 14-day forecast actually is. And it's not.
Saturday 25th December first appeared at the end of the forecast ribbon back on Sunday 12th. The forecast was overcast and dry with a maximum of 6°C and a minimum of 3°C (which was much the same as the forecast was predicting for every day in the week before Christmas). No hint of sunshine, which proved correct, but colder and drier than it would turn out to be.
Predicted temperatures remained similarly cold over the next four forecasts. The first hint of a Christmas Day raindrop came on Wednesday 15th, i.e. with 10 days to go. The first incorrect dabble with sunshine came on Saturday 18th, one week in advance, but fairly soon faded away. The forecast shifted significantly on Monday 20th with a leap in temperature suggesting a mild Christmas. The meteorological model that powers the calculations had come to a new conclusion regarding the advance of a tropical air mass and the location of a cold front, but that turned out to be too optimistic.
It was only on Wednesday 22nd that the forecast finally settled down around the correct values. It was still too overconfident about the amount of rain but got that right the next day, and was absolutely spot on by Christmas Eve.
Christmas Day's light rainfall turned out to be particularly difficult to predict because it was sandwiched between a very wet Christmas Eve evening and a very wet Boxing Day morning. The long-term meteorological model didn't have a chance of getting that right with a week to go, let alone two, and instead just churned out a suggestion of best guess average conditions that turned out to be off-whack. Britain's weather is all about battling air masses and not even supercomputers can reliably determine the outcome several days in advance.
I don't know why the BBC's meteorological service insists on putting out a 14 day forecast when it's invariably incorrect. Their previous provider, the Met Office, only risks 7 and even that's usually wrong to start with. It's all too easy to place far more faith in these figures than they deserve, forgetting that two weeks out they're much more fiction than science.
This Christmas it turned out the Christmas Day forecast wasn't worth checking until the big day was three days away. A maximum of 9°C, a minimum of 3°C, a smattering of drizzle and a total lack of sun. Which, incidentally, is the same weather they're predicting for two weeks' time so best ignore that too.
the dg weather review of the year
I normally do this with big colourful monthly tables but that's too much effort for too little reward so here's a breezier summary. All data is for Hampstead, as per usual.
temperature
rainfall
sunshine
Jan
cold
very wet
Feb
one v chilly week
3 snowy days
Mar
ended very mild
dry
Apr
cool
exceptionally dry
sunny
May
mostly cool
wet
Jun
1st half hot
wet
Jul
one heatwave week
exceptionally wet
Aug
never hot
dry
very dull
Sep
briefly hot
dry
Oct
fairly mild
wet
Nov
one cold week
very dry
Dec
unusually mild
wet
exceptionally dull
2021
typical
wet
dull
Apparently the sun's only come out for two hours at Hampstead during the last three weeks, and not at all for the last 17 days. But then it's not been a vintage year sunshinewise with only April managing more blue skies than average. Roll on 2022, or whenever the big yellow globe deigns to return.
