Question about Urock wind speeds

[lgzdbf]

Hi,

I am currently analyzing each Urock output for every hour studied. On some pixels, I am observing rather high values compared to the initial wind speed (double the initial wind speed value in the example) provided in the forcing data.

Do you think this is normal?

I am attaching a case for a specific day and tile. Let me know if you would like more examples.

Also, I am not sure if this observation qualifies as an "issue". Feel free to tell me if you would prefer that I open it as a discussion instead.

Thanks in advance,

c-408_grenoble.zip

[j3r3m1]

I would not be surprise that the wind speed can greatly increase in some areas. However, the location of your high wind speed is probably not realistic. However I do not think it is a bug but more the result of the use of Röckle based models which are not (yet ;-)) perfect for all situations. I suppose you do not have the vector output ? It is often a good idea to keep it to better interpret the results.

[lgzdbf]

Thanks for your answer. If you want I have Rockle zones associated data. Let me know if it is vector data you are talking about.

Rockle_zones_c-408.zip

[j3r3m1]

No I am talking about the output wind saved as a vector file. Having the wind field with wind direction is easier to interpret what we can observe as wind speed.

[lgzdbf]

Ok, I see, maybe that ?

urock_output.zip

[j3r3m1]

Yes this is it. The acceleration is due to the strange building shape. All the wind coming from the north part of the cavity zone is ending in the small zone located between the building and the vegetation, leading to a quite high increase of the wind speed.

I need to check the initialization of the wind because it seems that the resulting wind is quite strange around the complex building geometry.

[j3r3m1]

Indeed bug in the vegetation located in built zone: the wind is set to 0 instead of being only deccelerated. I need to investigate why this behavior is now happening...

[j3r3m1]

OK no problem the reason is simple. When z<z0 the wind speed of points located in vegetation polygons is set to 0. For this zone and wind direction, z0 = 1.98. The previous image is for z=1.5m.
Above 1.98 m the problem disappear (below is for 5m height).

[lgzdbf]

Okay, thank you for your feedback; so ultimately, is this wind speed of around 8 m/s consistent?

[j3r3m1]

In summary:

• there is no bug since the algorithm does what it is supposed to do when initializing the wind speed
• However, having wind speed equal to 0 in some area results in a quite high wind speed after mass balance (and in a direction which is far from obvious in my opinion...) --> it affects the wind speed in some area nearby (your 8 m/s is part of this area)
  ==> I have updated the code in order to limit this phenomenon (UMEP-dev/UMEP-processing@58d4c77). The wind in this area is now 6 m/s. The high value is still due to the complex geometry associated to the Röckle model type. I would not expect the wind being that high at this location. However, difficult to be more objective except using CFD simulations.

With the new version, at 1.5 m/s:
Initialisation:

Resulting wind field after mass-balance:

[lgzdbf]

Great, thank you very much, that is very clear! Thanks again for your help.