Surface files can give different soil profile settings depending on the host land model

[mpaiao]

This is related to a discussion in E3SM-Project/E3SM#7297. I recently found out that using the same surface files may give different soil settings between CLM-FATES and ELM-FATES when depth to bedrock (zbedrock) is provided.

Currently, only CLM uses this variable, and in ELM this means that the default soil grid is used. This grid sets the hydrologically active zone to be about 2.8 m (if considering the middle of the soil layer) or 3.8 m (if considering the bottom of the soil layer). I am not sure how serious this issue is, but I thought I should note it in here.

[mpaiao]

Just to follow up on this, I made a pull request that should address this issue for ELM. I think CLM already had more ways for changing the soil depth from defaults but maybe there are few things that could be nice to implement in CLM for cross-compatibility of the surface files:

1. ELM actually had the same capability for reading depth to bedrock from the surface files, however the expected variable was zbedrock in CLM and aveDTB in ELM. My pull request in ELM now looks for both variables.
2. In ELM the number of active soil layers in the surface file (nlevsoi) had to match the number defined in the code (10). Now the code just makes sure that nlevsoi ≤ nlevgrnd and updates nlevsoi based on the surface file.
3. I added the option to read variable ZSOI from the surface file if this variable is available. In my PR this is used for both defining the actual soil layers and for interpolating. Perhaps in CLM it would make sense for interpolation only as there are other methods for defining the soil profile.