On 23 Nov, 09:17, "Will Hand" wrote:

1. Zonal and mild (poss cold zonal at first) 50%
2. Cold and blocked very cold easterly 20%
3. Uncertain, possibly no-mans land, neither fully blocked nor fully zonal 30%

I've often wondered when models diverge and are particularly chaotic,
which of the models do the best? It's all very well comparing the
models for, for example, 500hPa height error over the course of the
year or other such situations, but it would be interesting to see
which of the models cope the best when the inter-model variability is
huge and things are on a knife-edge. Maybe some work has been done on
this already?

Richard