Initialising WRF with CESM

[Emily]

Hello,

I am trying to use CESM2 data to initialise the WRF model. I was planning to use CMIP6 output, but was advised in this post to use the CESM2 data directly from the model.

To initialise WRF I need to interpolate to pressure levels. I was using pyngl vinth2p, which gives strange values when interpolated below surface pressure. I didn't think this would matter as these values shouldn't be used by WRF, however I am using nudging in WRF and getting very high nudging values for geopotential even high up in the atmosphere, along with very high values for 'mu' ("PERT COLUMN DRY MASS FOR FDDA GRID NUDGING").

I am not sure why the vertical interpolation below the surface is causing a problem, but I assume this is the issue. I am now trying to use geocat "interp_hybrid_to_pressure", but am unsure what the values for 'geopotential/temperature at the lowest model level' should be. Are these taken from the 3D geopotential/temperature?

Any help you can give would be much appreciated as I cannot work out what's going wrong here. I have also tried using the cmip6 output (already interpolated to pressure levels), but I get the same issue, although to a smaller degree. The result is impossibly high precipitation.

Can anyone help with this?

Many thanks,

Emily

 

[slevis]

@Emily @nusbaume would it make sense to move this thread to the CAM forum? @Emily is there a similar WRF community where you may also ask these questions?

 

[Emily]

Thanks, the CAM forum is probably a good place. Is that within this forum or external? If it's possible to move the entire post please do, if not I will repost it.

I have also asked on the WRF forum, but could particularly do with advice regarding vinth2p/geocat which I think people here are likely to have more experience with.

Many thanks for your help.

Emily

 

[nusbaume]

Hi Emily,

Apologies for the slow response! I have asked a couple scientists here and they recommended using geocat's interp_hybrid_to_pressure function, with either TBOT or T at the lowest model level (which in CAM is the largest vertical index value) as the input for temperature, and PHIS as the input for the lowest model level geopotential. They said that produced realistic extrapolated pressure values for them, so if you are still getting incorrect values then please let me know, along with possibly a list of which CAM variables you are using as inputs into the python routine.

Also, I was told that the following manuscript describes below-ground interpolation in significant detail, in case you need to do create something yourself:

https://opensky.ucar.edu/system/files/2024-08/technotes_168.pdf

Finally, I have gone ahead and moved this thread to the CAM forum in case anyone else has any insights.

Anyways, I hope that helps, and please let us know how it goes!

Thanks, and have a great day!

Jesse