WRF 2D variables in NCVIEW

[chunchih]

Hi WRF support,

I am curious why am I seeing some strange-looking boxes in some of the 2D fields (e.g., 2-m air temperature, 2-m potential temperature, 2-m water vapor mixing ratio, and heat fluxes) when viewing them using NCVIEW? I have attached the screenshot of T2 and TH2 plotted using NCVIEW from one of my simulation output files as examples. The 'box' patterns are circled.

Also, are these fields used in the model time integration, or are they just time-independent variables derived from other time-dependent variables?

Thanks,

David

 

[chunchih]

I have attached the namelist.input I used for this simulation. I am using QNSE PBL and surface layer schemes and Noah LSM.

For now, I am not terribly concerned about this problem since it doesn't seem to have a significant adverse effect on my simulation results. Please let me know if this is actually a serious problem.

Thanks in advance,

David

 

[Ming Chen]

Hi David,
Would you please also take a look at IVGTYP and ISLTYP? I am suspicious that you many also find similar 'box' patterns in these fields. if so, that will explain why the surface variables display the box pattern.

 

[chunchih]

Hi Ming,

I just checked these two fields in my output files. ISLTYP demonstrates a similar box pattern but IVGTYP shows a rather smooth pattern.

Is this because ISLTYP is related to SOILCTOP and SOILCBOT variables in GEOGRID.TBL? Would changing the interp_options for these two variables in my GEOGRID.TBL make the box pattern go away?

Thanks,

David

 

[Ming Chen]

David,
You are right that the interpolation option will affect the final pattern of soil type (ISYTLP). I would suggest you run georgic.exe using the default interpolation option, because these options are well tested and give us nice result.
If for some reasons you need to change the interpolation option, please check the result shown in geo_em to make sure they look reasonable.
Ming

 

[chunchih]

OK. I was using the nearest-neighbor interpolation scheme for these two fields, which is the default. It gave me the blocky-looking pattern in the soil types fields. I tried using the four_pt interpolation scheme for these categorical fields, but nothing seemed to change for the dominant soil category.

I also found out that this blocky pattern only appears in Canada (please see the attached plot of ISLTYP)...Maybe it's because soil types are documented differently in Canada than in the U.S. Because my simulation domain focuses on southern Canada, the blocky pattern is really noticeable north of the border.

David

 

[Ming Chen]

Hi,
I agree. If the original static data is blocky, then WPS cannot make it smoothy.