Comparison of two near-identical runs of the UM on different clusters

This is a comparison that I have done with the help of Doug Smith. We have both run as near as possible (he uses UM version 4.7, while ESSC runs UM version 4.5) indentical transient runs. Doug at the Hadley Centre and I on the ESSC cluster. One reason for doing this was to reassure ourselves that the UM is not doing anything unexpected on the ESSC cluster (at least compared to the Hadley Centre). You can find some plots from the comparison here. The main result is that there do not seem to be any statistically significant differences between the models in the long term.

Questions? Comments?

-- LeonHermanson - 20 Dec 2006