First of all, the following plot shows the observations (as they are from GTS). Top row is the threee individual 24h periods, borttom are the 48h and 72h accumulations. Sunday the rain was mostly in the southern areas, by Monday it moved to north, and Tuesday it stayed there but the amounts were less extreme. The 72h total shows up as over 100m even with this sparse coverage (especially in the flood affected region). Orange corresponds to 30-50 mm, red to 50-75 mm and purple to 75-100 mm.
On the following plots always 8 model runs appear (if available verifying on the corresponding period) starting from top left going to the right. The 1st and 3rd rows are ECMWF, the 2nd and 4th rows are UKMO forecasts. The bottom right corner shows the observations. The colour scheme is the simila in the forecasts and the observation. Orange corresponds to 30-50 mm, red to 50-75 mm and purple to 75-100 mm.
- For Sunday the MetOffice is better, the EC timing is quite bad, the rain goes too far north (execpt one UKMO run, which missed the event completely).
- For Monday it is more mixed, first the EC forecasts are better, showing consistently the large precip, even if too far north. Later, closer to the event, the EC forecasts are surprisingly bad, the precip is way too far north, and also not enough.
