CESM3_alpha02a FV3 build problem

[lamaliang]

Hello,

There is a bug in the FV3 case build. Due to CESM3_alpha02a removing MCT, "comp_interface" should be removed in components/cam/cime_config/buildlib at line 114.

By the way, if we would like to use FV3 for CAM7LT or CAM7MT, could you provide initial data for the simulation? If not, do you have any suggestions?

Thanks

 

[lamaliang]

also should removed in components/cam/cime_config/buildlib at line 55

 

[jedwards]

Thanks - this fix will be in alpha02b - sorry for the problem.

 

[jedwards]

I'm not sure when we will be able to supply FV3 initial data for cam7LT and cam7MT simulations.

 

[lamaliang]

Thanks for your reply.

Due to our model TaiESM1 that is base on CESM, in next version, we would like try to testing FV3 dynamic core instead of FV. Could you have any suggestion about initial condition data.

Thanks

 

[aherring]

You could start by remapping the L58/L93 ncdata files (in namelist_defaults.xml) for SE to FV3.

The remapped lat/lon arrays won't be identical to the ESMF mesh file values and so you will need to overwrite the mapped lat/lon values (ncks -A -v) with mesh file values.

 

[jet]

I was just going to answer along the same lines. There are a number of options for remap but in the past I've used ncremap and the appropriate SCRIP files on Derecho under /glade/p/cesmdata/inputdata/share/scripgrids. If the model crashes right away you may also have to shorten the dynamic timestep to get past the initial imbalance due to the interpolation. There are other options to increase damping as well but I'm a bit out of my depth with any recommendations along those lines. See https://www.gfdl.noaa.gov/wp-content/uploads/2020/02/FV3-Technical-Description.pdf

 

[lamaliang]

Thanks so much. I will try it.

 

[lamaliang]

BTW, I have already remapped the ne30 initial data to C96 for LT and MT, and the simulation has started.