Usually ECMWF provides better forecasts in Europe than GFS and I know that convective weather is very often hard to pinpoint exactly. But today ECMWF missed Austria completely while GFS is spot on.
ECMWF
GFS
Actual radar image
I’m at home until Sunday and I don’t have something smarter to do 
so I’m checking the weather. While looking at the extent of the thunderstorm right now in my area, I noticed the same situation as yesterday – pretty poor ECMWF forecast for Austria while GFS correctly forecasted what was going to happen, comparison picture below.
However, I believe I solved the mystery: ECMWF data is 4 hours old while GFS data is 17 minutes old practically stating current atmosphere state. I’ll make a comparison of forecasts tomorrow morning for the evening and see which one is more precise.
I would have never expected a wx forecasting service to be of much use for TS location.
It should be able to predict regions where the temperature lapse rate etc is sufficient for TS generation, but you are talking about merging that with actual wx radar or sferics data, and presenting it as a “forecast”. I would be surprised if any wx model does that.
but you are talking about merging that with actual wx radar or sferics data
No. The first two pictures are forecasts for 7 p.m. (ECMWF and GFS respectively) while the third one is actual radar data. However, GFS starting point is some 15 min before actual time which means based on more recent data while ECMWF was updated 4 hours ago.
I would have never expected a wx forecasting service to be of much use for TS location.
You’re right but missing a complete country known for TS activity in conditions where all surrounding in littered with TS is kind of failure.
I repeated exercise but this time in the morning getting ECMWF, GFS, and WAFC weather chart forecasts as well as actual radar image few minutes ago.
Today’s conclusion is that ECMWF is more consistent with what actually happened while GFS forecast matches WAFC but pretty much missed some TS areas. One more proof how hard is make precise forecast in these atmospheric conditions.
On a slightly different note, my observation is that the blue and green shades in windy are not precipitation at all, both regarding the forecast and the actual weather. Yellow and red is. I’ve seen this numerous times. Not sure what the background of this is. Maybe an uncertainty factor? Anyone?
aart wrote:
On a slightly different note, my observation is that the blue and green shades in windy are not precipitation at all, both regarding the forecast and the actual weather. Yellow and red is. I’ve seen this numerous times. Not sure what the background of this is. Maybe an uncertainty factor? Anyone?
Weather radar can’t really see precipitation as such – only water. So the coloured areas are really clouds. Of course the more water in the clouds the more likely it is that there will be precipitation.
I also think windy.com does not forecast fog at all.
Windy does predict fog (short term) and is in fall/winter pretty good.
I concur wirh Emir that on Friday (again) windy was a total fail w.r.t. TS risk.
Only showed the South of Austria with TS risk & precipitation, while there was already heavy activity around Vienna…
What was predicted OK (too low, thus consistent), was the strong & very gusty wind – which may be enough to make you reconsider.
Otherwise I am usually pretty satisfied with windy for shorter term expectations, but checking its various weather model versions where available.
