"Surface" temperature/wind doesnt show earth's surface values
Thanks for the link. That's exactly the same issue in fact
The actual altitude for Garmisch is 707m, but the ECMWF's reference altitude for G-P is 1369 m !
Also for Zugspitze (actual=2962 m , model's reference alt.=1700m).
Difference 1262m, temp. adjustment - 4 C (aprox).
Also for Zugspitze (actual altitude=2962 m , model's reference alt.=1700m).
Difference 1262m, temp. adjustment - 8,2 C
(as 6.5/1000 = 8.2/1262)
Could we know where do you find these graphics ?
Yes it is a weakness in mountain areas, but ECMWF and GFS are global models designed to provide weather forecasts covering the globe, difficult to ask them for a better resolution. At least NEMS-4 must be more precise in alpine area... as claim Meteoblue. But difficult to see a difference with ICON-7. The best should be COSMO-1 (1.1km) covering the alpine arc, or AROME (1.3km) covering France and a little more.
My concern is more about the Freezing Altitude which is considered « above the ground » but not above the sea (amsl) by ECMWF and ICON. In mountain, in Germany, what is the Freezing Altitude reference, local ground or amsl?
From ECMWF's application named "Forecaster" but it is not for public use,
you have to log in.
You said: "We have here the same problem, with -9ºC in Chamonix (1050m) and -11ºC on top of Mont-Blanc (4850m) !".
I made a small research and I provide you the data and conclusions
(for ECMWF model).
Chamonix: model's "surface" =2137m, elevation=1036m (Δ= 1101)
Mont-Blanc:model's "surface" =2265m, elevation=4808m (Δ= - 2543)
Mont-Blanc: -16,5 C.
Those "adjustments" take into account Standard Atmosphere's lapse rate (6,5/1000m).
In the real world, the lapse rate may vary from 9,8 °C/km in dry air
to around 5 °C/km for moist air (into clouds).
Hope it helps
... in the same way (adjustments in ºC) ...
for Vienna: +0,2 Innsbruck:+3,9 Praha: +0,4
Olympus Mt.peak (2917m), Greece: -10
Wouldn't it be even more accurate to use the result of the model for the exact elevation?
So for example Mont Blanc: The delta between ECMWF model elevation and real elevation:1101m:
So to get the estimated temperature/wind at the summit: use the closest calcuated value to 1101m above model surface. That's certainly also not 100% correct, because it doesn't take surface effects into account, but maybe more accurate than to assume ISA.
It don't have access to the ECMWF data, but here a snapshot from NEMS4 sounding for Mont blanc:
The NEMS4 model surface seems to be at ca 3200m according to this sounding, temperature around -5°C.
At 4800m it shows like -15°C. So a delta of -10°C in this case.
For Mont Blanc the Δ is 2543m
At the "sounding" :
3200m ... -5°C
4800m ... -15°C
4800-3200=1600m ..... ΔT=10
10/1600 = 6,25/1000m = ISA (almost !)
Nice sounding! Is freely available ?
Yes almost ISA in this case, but just an example. But in most cases the ISA temperature delta would be good enough problably.
This graph is from meteblue, it's only available for point+ subscriber i think.
Thank you. I took the example of Chamonix and Mont-Blanc temperatures just to underline the wrong temperature display by global models in mountain areas. You give the way to know the exact temperature, fine. But for me, just a quick look to the airgram, as you explained before, it’s enough precision.
What I have found through these questions, that’s the worse terrain resolution of the global models compared to there atmospheric resolution. And this is not obvious, as Windy show a very precise terrain resolution, we expect the same precision from global models, but that’s impossible !