Namelist definitions not being added to lnd_in file if `group` value is mispelled

### Brief summary of bug

If the `group` value in a namelist definition does not correctly reference any defined group the namelist option will not be set.

### General bug information

**CTSM version you are using:** ctsm5.3.038

**Does this bug cause significantly incorrect results in the model's science?** [Yes / No] No

### Details of bug

I discovered with #3063 that the `fates_electron_transport_model` wasn't showing up in the lnd_in file or the lnd.log file. This was due to the `group` value being incorrectly set to a non-existing group name:
https://github.com/ESCOMP/CTSM/blob/38e25b80eeedd1b524d8026cff0ba275c8a4ada0/bld/namelist_files/namelist_definition_ctsm.xml#L783-L788

I'm assuming this hasn't shown up previously since it is likely that other namelist options would result in other failure modes or unexpected B4Bness due to the option not being exercised. Since this option is passing values to FATES for an as yet unimplemented feature, there was nothing to notice being wrong.

### Important details of your setup / configuration so we can reproduce the bug

Tested with #3063 PR branch

### Important output or errors that show the problem

Example lnd.log with no output for the electron transport model option:
```
941 ED/FATES:
942 use_fates = T
943 fates_spitfire_mode = 1
944 fates_harvest_mode = no_harvest
945 fates_stomatal_model = ballberry1987
946 fates_stomatal_assimilation = net
947 fates_leafresp_model = ryan1991
948 fates_cstarvation_model = linear
949 fates_regeneration_model = default
950 fates_radiation_model = norman
951 fates_hydro_solver = 2D_Picard
952 fates_electron_transport_model =
953 fates_paramfile = /glade/campaign/cesm/cesmdata/inputdata/lnd/clm2/paramdata/fates_params_api.38.0.0_14pft_c250226.nc

```