Road widths unrealistic in urban LCZ Table

[bfkg]

Different to the classification in URBPARM.TBL where the road widths are 8–10 m, in the LCZ parameters file, the default widths are betweeen 39 and 108 m.

https://github.com/wrf-model/WRF/blob/3dd1546ba425b5e77d682dd92253738416b3cb0a/run/URBPARM_LCZ.TBL#L42

This leads to occasional model crashes and unrealistic surface temperatures of > 350 K.

I suggest to provide more realistic default values with respect to the WUDAPT method, or even multiple exemplary files for certain cities or typical for countries.

[weiwangncar]

@cenlinhe Can you take a look at this report? Thanks.

[cenlinhe]

These road width values do seem to be too large. I am not sure the history of these values. @andreazonato Do you know where these default values come from? Do you have recommended (more realistic) values to replace them? I did not find any road width data in Stewart and Oke's LCZ paper.

[cenlinhe]

I discussed this with Alberto Martilli (expert on WRF-urban modeling). He does not know where these default values come from, but he suggested to make them equivalent to those used in the multilayer scheme (see STREET PARAMETER section towards the end of the file). Same thing for ROOF_WIDTH and ZR values. He will provide me with an updated URBPARM_LCZ.TBL file and I will create a PR for it once I have the updated file.

[cenlinhe]

I submitted this PR (https://github.com/wrf-model/WRF/pull/1969) based on Alberto Martilli's suggestion to solve this issue.

[weiwangncar]

@bfkg Can you see if proposed change fixes your issue?

[bfkg]

Yes. The new table seems to work properly. Thanks.

On Tue, 2024-01-23 at 09:55 -0800, weiwangncar wrote:

@bfkg Can you see if proposed change fixes your issue?

[weiwangncar]

@bfkg Thanks for your response.