District scale simulation - forcing with weather data only on the borders?

This is the place to post and discuss your project results, ideas, visions and other things in your mind (...-related to ENVI-met ;)...) with other users.

Remember: This is a forum on technical Q on ENVI-met. We can not/ will not analyse your project results for you(although we might give hints).
pusztaipter
Posts: 3
Joined: Thu Jun 04, 2020 8:10 am

District scale simulation - forcing with weather data only on the borders?

Postby pusztaipter » Thu Jun 25, 2020 11:48 pm

Dear ENVI-met board!

I am Peter Pusztai, an EMJMD student in Smart Cities and Communities: http://www.smaccs.eu, and my thesis will be to explore the effects of vegetation on thermal comfort using your simulation software's capabilities - on a district scale. Please see my thesis proposal here:https://drive.google.com/file/d/15krWl8XDf36exmUCvNbEKGLzA6E8DBfL/view?usp=sharing.

... So the input area size is planned to be 4x5 km, 8k x 10k cells, with a resolution of 0.5m. Looooong simulation.

What I would like to simulate (inside that area) are the Dependent variables: Air Temperature (°C) [+ Mean Radiant Temp. (°C), Air Temperature Change (K/h)], Relative Humidity (%), CO2 (ppm), Wind Speed (m/s) etc.

Therefore I would like to input this weather data only at the bordering cells of this input area file / simulated district.
Do you think it is possible, or do I have to assign any of those dependent variables to all the grid elements/cells?

Thank you for your time and consideration,

Best wishes,
Peter Pusztai

Tim
Posts: 1565
Joined: Mon May 15, 2017 3:46 pm

Re: District scale simulation - forcing with weather data only on the borders?

Postby Tim » Fri Jun 26, 2020 2:58 pm

Hi Peter,

There is no possibility of assigning those values to all grid cells. Every forcing type applies the values to the border. The simulation will then calculate its own values and flows etc within the area.
You will not be able to run such a large model area. 400x500 grid cells would be rather near to the maximum amount of cells I would recommend you to run... everything else will take months. 750x750 is the largest model area I have seen so far. You will also get problems regarding to analysis in Leonardo etc. NetCDF outputs strongly recommended for that. However, 400x500 cells would mean to use a resolution of 10 m instead of 0.5 m. Also for that amount of grid cells, a high end desktop PCs with 16 or 32 cores, a lot of RAM and disk space is needed.
0.5 m resolution is not necessary in general... for high resolution, you would run it in 2 m or 3 m resolution. Still with 2m or 3 m resolution, you would then have to create multiple model areas and run them partially because of the size.

Best regards,
Tim

pusztaipter
Posts: 3
Joined: Thu Jun 04, 2020 8:10 am

Re: District scale simulation - forcing with weather data only on the borders?

Postby pusztaipter » Fri Jun 26, 2020 4:15 pm

Dear Tim!

Thank you very much for your quick and detailed answer! Oh, I understand my misunderstanding about the borders now, thanks for the clarification!
I see your other points as well, a question if I may: If I would proceed with the 10m resolution, how could I describe a 3D tree object for example?
My study is about the effect of vegetation, it would be important to model each tree instance according to the real life scenario.

Any idea on how to overcome this resolution issue?

Thanks,
Peter

Tim
Posts: 1565
Joined: Mon May 15, 2017 3:46 pm

Re: District scale simulation - forcing with weather data only on the borders?

Postby Tim » Fri Jun 26, 2020 4:34 pm

Hi Peter,

In the 10m resolution, there will be most likely only one shape of tree: being one full 10x10xX meter cell filled with an LAD value.
If it is about single trees, the shapes are very important obviously. You should then try to locate those areas of interest where the trees need to be analyzed in detail and simulate them separately.
It is all about the scope of your work. If you need to get the results for the whole area in one single large scale simulation, you will have to choose a coarser resolution. If specific smaller areas with specific trees are of particular interest, I'd rather simulate several high resolution areas.

Best regards,
Tim

pusztaipter
Posts: 3
Joined: Thu Jun 04, 2020 8:10 am

Re: District scale simulation - forcing with weather data only on the borders?

Postby pusztaipter » Fri Jun 26, 2020 5:31 pm

Dear Tim,

thank you again for your detailed explanation, it all makes sense to me now!
First I will proceed with a coarse simulation then (4 of them), and based on the results' differences, I would select a couple with specific area of interest and more detailed resolution.

Best regards,
Peter


Return to “Projects & Results & Ideas”

Who is online

Users browsing this forum: No registered users and 0 guests