Simulation error - vegetation problems?

Discussions about the Version 4 concerning the core model simulations
Post Reply
Phil Re
Posts: 4
Joined: Wed Aug 19, 2020 7:32 am

Simulation error - vegetation problems?

Post by Phil Re »

Hi Tim,

for my master's thesis I am currently creating a climate model for each of three city districts. Two of them run without problems, but one of them always breaks off after 3% for reasons that I cannot explain. The model has an area of 200x200x30 grids with a horizontal resolution of 2 meters and a vertical resolution of 3 meters. The outermost 25 grids were left free on each side, so that sufficient distance to the first buildings and vegetation is ensured. There is also sufficient space to the top of the model. The model is driven by full forcing with data from a nearby weather station.
I had already tested various modifications of the model and smoothed the input data (wind direction and speed) but always unsuccessful. If I remove the 3D vegetation, the model runs without problems - so the error seems to be located here. Where exactly could the mistake be with the 3D-vegetation? I only used trees from the system database, so I didn't need any vegetation I created myself ...

All model data as well as the generated error message are attached in the appendix.
Many thanks for the help!

Phil
Attachments
Modeldata.zip
(1.28 MiB) Downloaded 11 times

Tim
Posts: 1707
Joined: Mon May 15, 2017 3:46 pm

Re: Simulation error - vegetation problems?

Post by Tim »

Hi Phil,

I would recommend you to run the model without DEM. You only have small height differences, so it will not make much of a difference. In contrary, it may cause simulation issues and furthermore problems in the analysis.
I'd also reduce the amount of greenery if running without DEM does not help. Trees seem to be very large and very dense, maybe there is too much LAD in your model area.
FullForcing looks good to me. However, cloud cover should not exceed 8 in total. In your files, low, medium and high clouds always show the same value leading to a sum high above 8.

Best regards,
Tim

Phil Re
Posts: 4
Joined: Wed Aug 19, 2020 7:32 am

Re: Simulation error - vegetation problems?

Post by Phil Re »

Hi Tim,

thank you for your reply! Sounds good - I'll try all of the improvements you suggest and hope that the simulation finally works. Thanks also for the advice regarding cloud cover - I was not aware that the total amount of cloud cover (low + medium + high clouds) should not exceed 8.

Greetings
Phil

Phil Re
Posts: 4
Joined: Wed Aug 19, 2020 7:32 am

Re: Simulation error - vegetation problems?

Post by Phil Re »

Hi Tim,

unfortunately I have to open up this topic again. After I had significantly reduced the LAD in the model as you suggested (use of lower trees with a less dense canopy), the model ran well - until it broke off again at 42% (see appendix). I guess the only option left is to try to remove the DTM to create more stability?
My question is: is there a way to remove the DGM without having to transfer the model back to the 2D-Concept-Design? Otherwise, unfortunately, all adjustments to the buildings (roofs etc.) will be lost, which would then have to be reworked manually and in a time-consuming manner. Is there possibly some way to avoid this?

Best regards
Philipp
Attachments
Fehlermeldung_newRun.png

Tim
Posts: 1707
Joined: Mon May 15, 2017 3:46 pm

Re: Simulation error - vegetation problems?

Post by Tim »

Hi Philipp,

Yes, that is possible in a texteditor such as Notepad++. You need to delete everything that is between the tag <dem3D> </dem3D> and the tags themselves. Additionally, the terrain heights in the 2D DEM field need to be 0 for every cell (found at the tag <dem>). There are multiple ways to achieve both things. One way would be to export the Monde area again without topography (simply uncheck the box). Open it in Spaces, convert it to 3D and then copy paste / replace the DEM sections within a copy of your old perfect file (where DEM was included) by using Notepad++.

However, you might also check the wind speeds / directions again. Maybe you need to correct the forcing file somehow: Less wind direction changes, more wind speed for the timesteps around the time it crashed.

Best regards,
Tim

Phil Re
Posts: 4
Joined: Wed Aug 19, 2020 7:32 am

Re: Simulation error - vegetation problems?

Post by Phil Re »

Hi Tim,

I had already tried editing the Inx-file in the text editor - but never with convincing results. I think it was because I didn't set the terrain heights to zero in the 2D DEM field. Now, it worked with your instructions - so a very big thank you!

I will definitely check the wind speeds/directions around the time of the crash as well. Maybe that will add stability.

Greetings
Philipp

Post Reply