I know its a long way out, but I find this chart a bit concerning.
If you look at the forecasts of this model from previous years, the temperature forecasts have been, IMO, very poor, at least for our latitude anyway. However, looking at the forecasts of the 500z geopotential height, they are significantly better than the temperature ones.
The model correctly predicted the blocking that occured in 2009/2010 from as early as September:
and the November to December 2010 blocking:
as well as the lack of 2011/2012 blocking:
However, the temperature forecasts arent so great, for example, 2009/2010 again at the same time as the first chart:
So I do think that this model has performed well at modelling heights, just take no notice of those temperatures!