Remove `FATES_ERROR_EL` from the `FatesColdST3` testmod

[glemieux]

Per NGEET/fates#1155 update, FATES_ERROR_EL does not have relevance to this test. FATES has been updated to remove this from the history output during this run mode and as such, will fail during RUN if this variable is included in the user_nl_clm history output.

[glemieux]

This is address for the FatesColdST3 test via ebf5f26.

[glemieux]

This is address for the FatesColdST3 test via ebf5f26.

Actually this doesn't address the issue given #2867 (comment).

The short term fix here is to clear the tapes and re-add the Fates testmod base variables to the FatesColdST3 user_nl_clm file, leaving the FATES_ERROR_EL out.

[glemieux]

Reopening this as my attempted fix in #2782 didn't really work.

[ekluzek]

The problem here is that FATES_ERROR_EL is explicitly not included when use_fates_ed_st3 is turned on. So adding excludes for it doesn't work as it'll die with an error when you try to exclude something that hasn't been added with a hist_addfld call.

The error seen looks like this:

lnd.log:

 the history field: FATES_ERROR_EL
 was requested in the namelist, but was
 not found in the list of fates_hist%hvars.
 Most likely, this is because this history variable
 was specified in the user namelist, but the user
 specified a FATES history output dimension level
 that does not contain that variable in its valid set.
 You may have to increase the namelist setting: fates_history_dimlevel
 current fates_history_dimlevel:            2           2
 ENDRUN:
 ERROR in clmfates_interfaceMod.F90 at line 769

An easy way around it is to remove it from the excludes. But, it has to be removed from both the fexcl line in FatesColdST3 AND the fincl line in Fates.

So if it's needed for other cases, it would need to be added to the cases that need it, and not mentioned in the ones that don't.