Roundabout fix of "USING URBPARM_LCZ.TBL WITH OLD 3 URBAN CLASSES. SET USE_WUDAPT_LCZ=0" while using cglc_modis_lcz_global data

[Tanmoy]

Hello WRF community,

I'm recently trying to run a short simulation using Local Climate Zones (LCZ) using the CGLC_MODIS_LCZ_global (https://www2.mmm.ucar.edu/wrf/src/wps_files/cglc_modis_lcz_global.tar.gz) in WRF v4.5.2.

I have 3 domains, whose info in shared in the namelist files.
I have followed all the steps to run this simulation correctly and it was successfully executing upto real.exe. But while running wrf.exe, It was showing the error:


------------- FATAL CALLED ---------------
USING URBPARM_LCZ.TBL WITH OLD 3 URBAN CLASSES. SET USE_WUDAPT_LCZ=0
-------------------------------------------
Abort(1) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

So, I search the internet and found that the issue was my domain d01 being very large and not having any LCZ classes. So, I employed the fix shown in: Bug for determining the use of urban LCZ parameter table · Issue #1878 · wrf-model/WRF , and recompiled the WRF. But the compilation of WRFV4.5.2 failed. So, I recompiled WRF using the old code (not the fix).

But, now to solve this error, what I was thinking of doing is assigning a dummy variable of any LCZ class, in the edge (1,1) of my domain d01. This actually solved the error. and model successfully ran. But is this approach correct, Will changing a single cornermost grid of domain d01 have any impact on my study region which is well under domain d03??

Also, by what I understand use_wudapt_lcz=1 will only have one value, like it cannot be enabled for domains seperately.

 

[kwerner]

Hi, Making that small adjustment to the single grid cell on the edge of d01 likely won't make much difference. I'd recommend trying to run a short test, just to make sure things look reasonable, and if so, then proceed as planned.

 

[cenlinhe]

Also, we usually recommend users to activate urban physics for all domains to ensure model consistency and stability.
For this LCZ issue, you may also want to double check the LCZ map (LU_INDEX) generated in your geo_em file looks reasonable (i.e., LCZ should have numbers of 51-61)

 

[Tanmoy]

Hello everyone,

Sorry for the late reply. I have gone through much literature, and it seems that LCZ-based studies are better suited for the high-resolution domain. However, my study region is somewhat on the moderate resolution side, and it considers a larger area with multiple cities within it. Also, I observed that my simulation output had very minimal changes when incorporating LCZ, so I have opted to drop the idea for now, but I hope to work with these in the future.

Thank you to everyone who has replied to this thread. May it help other researchers in one way or another.

Best Wishes,
Tanmoy