LU_INDEX and LANDMASK changed in wrfout file

[Yuanlin]

Hello, WRF team

I am using wps-4.1 and wrf-4.1 for a two nest simulation

The problem is that I found the LU_INDEX and LAKEMASK from the wrfout file is different from the LU_INDEX and LAKEMASK in the geo_em*.nc file, and other relevant fields (ALBEDO, SMOIS ...) changed as well. In particular, this problem occurs only in domain-1.

I tried to change surface_input_source from 1 to 3 (in namelist.input, physics), but that didn't work.

Would you please let me know if it is normal to have LU_INDEX changed in the wrfout files?

The attachment is geo_em*, wrfout*, namelist.wps and namelist.input

Thanks,

Yanlin

 

[kwerner]

Hi Yanlin,
When you tried setting surface_input_source = 3, did you re-run real.exe first? If not, you'll need to do that.

Also, just to note, with the files you sent, I see the differences in both d01 and d02.

 

[Yuanlin]

Hello, WRF team

Thank you very much for your reply！

Unfortunately, I have tried to re-run real.exe after setting surface_input_source = 3, but nothing changed. In order to describe the problem more accurately, I have attached the LU_INDEX figures of geo_em, wrfinput and wrfout (in LU_INDEX.docx).

From the figures, everything was fine in wrfinput, and lake grids were accurately replaced with water body (the default set of real.exe). When it comes to wrfout, d02 was well consistent with wrfinput (I am not sure what is the difference in d02 you mentioned above), but d01 was obviously changed.

In addition, I have setting sst_update=1, so the SST was mismatch with water body in d01 due to this problem.

To summarize, does the LU_INDEX problem in d01 affect the simulation results? especially in d02? How can I solve this problem?

Thanks again!

Yuanlin

 

[kwerner]

Yuanlin,
Thank you for the thorough explanation. The graphics were useful. I believe the reason for the changes to d01 are due to the fact that you have feedback turned on. This means the higher-resolution values for d02 will go back and overwrite d01 values. Because the resolution is higher on d02, it can cause small differences. Typically users are interested in the results for the higher-resolution domain, and are only using d01 as a sort-of buffer zone for input to d02. Assuming that is the case, the differences you are seeing on d01 should not be problematic.