IIRC some model runs this time last week, especially those from the GFS, were indicating the zonal spell lasting until possibly the last week of January. There really has been quite a turnaround in the operational runs since last weekend.
Originally Posted by: David M Porter
It's worth mentioning again how well the ensembles have handled the end of the mild zonal dross. Two weeks ago they first picked up on a decent chance of it and although sometimes it was as low as 35% of the runs it was generally around 50% of the runs which showed a transition to cold weather. In the last week the colder cluster gained more and more members and, well, you know the rest!
It's a good reminder that using clusters, as the Met Office themselves use, is the best way to use the ensemble output. The operational runs are fun but time and time again you'll find some folks looking at them to the exclusion of the rest of the suite. There's a reason the Met Office have their own ensembles (MOGREPS) and ECM cluster ensemble product (DECIDER) after all!
Talking of the latter, look what I've found! I guess that explains why Fergie mentions Shannon Entropy, as it's a graph within the DECIDER suite! (This is sample data rather than the real thing, which I suspect costs megabucks...)