December is currently looking like a very similar set up to November in terms of extreme warmth.
The output this morning has the CET at 8.9C by the 12th. This is 3.8C above the 1981-2010 mean.
However, it is not unprecedented. Indeed 2000 was a degree warmer. Here are the warmest first 12 days of December on record:
2000 9.86C
1898 9.77C
1934 9.40C
1979 9.33C
2015 8.90C EST
1831 8.88C
1918 8.88C
In recent times most years that have had a warm November and December have gone on to have either a very mild rest of the winter of something closer to average. This has happened quite a few times in recent years - 2011, 2006, 2002, 1997, 1994 (1997 was the only El Nino year but it was a very strong one).
Going further back there are a couple of exceptions. 1986 being the obvious one when Jan, Feb and Mar 1987 all ended up cooler than average. 1986 was a moderate El Nino year. 1953 is the other which saw Jan and Feb 1954 quite cold. There was a weak El Nino that year.
The best match is probably 1986 which saw a cool September followed by very warm months in Oct - Dec. This pattern is very similar to 2015.
However if we look at the QBO we see that 1953 and 1986 were in the easterly phase. All the warm years were either in a westerly or neutral phase apart from 2011 which does stand out at somewhat odd.
As Gavin's model roundup this morning on his website shows, there are mixed messages from the models. Some want to go with a very mild winter whereas some others want to go with a progressive set up starting mild in December and turning colder in February. There is support for both scenarios from the analogues. Very difficult to know which one is more likely.
On balance my inclination would be to favour the milder solution simply because of the QBO phase. In 2013/14 we saw a very mild winter with a very positive QBO. Last year the QBO was a record negative and the winter was much cooler, albeit it was not cooler than average just average. This year the QBO is looking likely to swing from the record negative last year to a record positive this year. The October reading was +13.4 which is higher than the +11.7 in 2013. The previous highest October reading was +13.1 in 1980.
However, a positive QBO does not necessarily mean a mild winter. 1985 is a good example of an exception. A very warm December was followed by four very cold months in early 1986.
So I think the best that can be said is there is probably about a 50% chance of a very mild winter and a 50% chance of a very mild first half of the winter with possibly a very much colder second half which could extend into March.
I am surprised that 1974 isn't in there.