A quick exploration of forecast evolution using MOGREPS-UK data. We've developed a normalised error metric for ``forecast uncertainty'', and taken a more heuristic, visualisation-based, look at 2015/16 winter storms Desmond and Gertrude.
The most illuminating notebook to look at to get the gist of what we've done is
ShinyForPresentation.ipynb