It was a crisp, misty winter day in London first thing this morning (actually yesterday, Weds 20th, by the time I completed this post). The trouble is, it’s the Spring equinox, when it’s normally very mild. The Guardian reports the lateness of Spring, with “nature lying dormant”.
I checked the weather forecast, as I normally do, and was amazed to discover that the temperatures later in the week (i.e. for Friday 22nd and Saturday 23rd) were predicted to be fully 5C colder than I’d remembered they’d forecast just the previous day! Yet more snow is now expected over much of the country.
The result is that March 2013 is now likely to be the coldest for fully 40 years.
I’ll cover the forecasting difficulties first – a side-issue really, but notable, nonetheless – and then look at how cold this March has been.
A Knife-Edge Winter Forecast
Luckily, when I switched on my work computer this morning (Wednesday 20th) the forecasts made by the Met Office on 19th were still cached by Firefox, so I was able to do some screen-grabs. Here’s the one for the weather at Heathrow on Friday 22nd, as of 11:00 Tuesday 19th:
when they were expecting the temperature to reach 9C in the afternoon.
Curiously, as I complete this post on Thursday, I see that the forecast is now for even colder weather on Sunday. This is worth including as the London temperature is expected to be around the same as on 11th, which was billed as the coldest March day for almost 30 years (since 1st March 1986, just after that record-breaking bitter February I might have mentioned before) for the country as a whole:
Why this sudden change in the forecast?
The answer is the difficulty in predicting the outcome of what is often termed the “battle” over the UK between cold easterlies and mild westerlies. The following charts show an an Atlantic weather system pushing in, bringing the rain seen in the forecasts above for Friday in particular. On Tuesday the system was expected to move across much of the country:
I’ve indicated the warm front, behind which the air is milder with a red arrow and the cold air with a blue arrow.
At first glance it seems little has changed, but in fact the warm front (red arrow) has pushed much less further north and east.
For completeness we see that the front, now indicated with a green arrow because it is occluded (hey, I’ve got a colour palette and I’m going to use it), is expected to be stuck over the country by Saturday:
So what seems a minor detail in the great scheme of things has a significant effect on the weather in one particular place, in this case much of the UK. Quite a bit of snow is now expected, though not in London, alas.
Clearly the Met Office need even bigger computers! They were big enough to ‘fess up in their weather-warning:
March 2013 in the CET
The difference between forecasts on Tuesday and on Wednesday is sufficient to change the predicted temperature for the whole of March. As in previous posts, most recently for this winter up to February, I prefer to look at the Central England Temperature (CET) series, because it gives a decent historical perspective.
The temperature so far for March is 3.7C. With cold weather expected until at least 28th, it is unlikely that the mean temperature in the CET will exceed 4C for the month as a whole. This will make it the coldest March for over 40 years. For once it seems the media have understated rather than exaggerated (though maybe measures other than the CET are turning out differently).
It looks almost certain we’ll beat March 2006 at 4.9C, and extremely likely the 4.5C in March 1996 will not be exceeded.
Curiously March 1996 and March 2006 are the only Marches (apart from 2013) averaging less than 5C since a run of 4 from 1984-7. One wonders if it is more than coincidence that this run followed the El Chichon eruption in 1982. Anyway, the coldest March of the 1980s was 1987 at 4.1C. It seems likely we’ll be colder than that this year as well.
Marches colder than 5C seem to have been more frequent back in the day, occurring in 1971, 1975, 1976, 1979 and 1980. Many would attribute the much reduced frequency in the 1990s, 2000s and so far in the 2010s to global warming, making this year even more exceptional (like the summer of 2011).
But we have to go back to 1970 for the most recent March with an average CET of less than 4C, when 3.7C was recorded. It’s possible this March could even beat that mark.
But March 1969 was even colder at 3.3C. I doubt the figure for this year will come out below that. Most likely the headlines will be “coldest March for 44 years”.
For the record, March 1962 was even colder at 2.8C, there was a 2.7C in 1892 and 1.9C in 1883. The record seems to be 1.2C in 1785, though really we should discount those clearly influenced by eruptions, in that case Laki, which must surely explain the run of 4 cold Marches from 1783: 3.3C that year, through 2.7C in 1784, 1.2C in 1785, as mentioned, to 2.1C in 1786. Of course, it was generally colder in that period as well.
March 2013 has been so cold that I thought I’d produce one of my CET graphs for the “long winter”, that is December to March inclusive:
In general including March makes surprisingly little difference (compare the similar chart for December through February in my last weather post). The longer perspective makes winter 2012-13 closer in coldness to 2009-10, but it’s still not historically very significant. Further back, 1962-3 becomes less notable, struggling to hold on to 4th place in the all-time coldest winter stakes, rather than challenging for 2nd, with 1813-14, famous for the last Frost Fair on the Thames, colder on a 4 month comparison. 1683-4 remains the outlier.
The running means (green, red and black lines) show that our “long winters” are not yet as cold as around 1980, which was milder than in the early 1960s, which in turn was less cold than in the late 19th century and the mid 18th. The late 17th century was colder still. The late 20th century and early 21st century mildness is attributable to global warming, of course.
CET Anomaly Annual Running Mean Goes Negative
What seems most significant about this March to me, though, is that it completes a run of 12 months colder in the CET than the 1961-90 mean.
I provided the monthly data to date for this year and for the 3 previous years earlier in the post. You can see that 2012 was only 0.2C warmer than the historical average. The first 3 months this year will have been significantly colder than in 2012, especially March which will be colder by more than 4C, maybe approaching 5C! The result is that the average of April 2012 through March 2013 is well below the 1961-90 figure (just average the anomaly column). (Sorry, no time just now to produce a CET 12 month running mean graph – another time maybe).
2010 was also significantly (0.6C) colder than the long-term average, as a result of the record-breaking December and, to a lesser extent, the cold January that year.
There were, previous to 2010, some 13 years of temperatures above the 1961-90 mean:
Since the global temperature data annual mean shows temperatures remaining as high in the 2010s as in the 1990s and 2000s, it does rather seem as if a change in weather patterns is now counteracting some of the effect of global warming on UK temperatures apparent through the two decades of the 1990s and 2000s.
Finally, it’s now looking unlikely that the year 2013 as a whole will approach the mean temperature seen in many of those years in the 1990s and 2000s. Then again, 2011 had a very mild summer, yet, taken as whole, wasn’t far off the warmest year on record. My understanding is that, because of global warming, we’re tending to get exceptionally warm months, such as February, April and November 2011, far more frequently than in the past. I expect we’ll see one again soon, though I have a hunch the next time the weather is notable enough for me to put together a blog post, it’ll be flooding I’m taking about, again.