[Dart-users] update to wrf model_mod.f90 and input.nml

[Nancy Collins]

dear dart/wrf users,

there's an important update i just committed which affects the
updated wrf/model_mod.f90 and the wrf/work/input.nml namelist file.

in the &model_nml namelist, there is a new logical, 
'default_state_variables'.
it defaults to .true. in the code, and i overlooked putting it into the
input.nml file in the wrf work directory, so it would be easy to be
completely unaware of its existence.  if left as true, the new code will
only run with a minimal set of wrf arrays in the dart state vector, namely:
U, V, W, PH, T, and MU.   this is regardless of which arrays are listed
in 'wrf_state_variables'.

if you're running with the revised model_mod, you will almost certainly
want to set this to .false. and list your own set of wrf arrays to include
in the state vector for assimilation.

if you have any of the output from your wrf_to_dart or dart_to_wrf runs,
you will see some print statements saying which fields are being
converted -- so you may want to check any recent runs to be sure that you
really have all the variables you expect in your dart state vector.

if you update from the subversion server, you will see 3 changes:

i've added 'default_state_variables' to the input.nml file in the wrf work
directory; i've changed the default list of arrays to match the defaults in
the model_mod, and i've added new messages in the model_mod init
code to print out explicitly which arrays are being put into the state
vector for each domain.

sorry for any problems this might have caused.

thanks,
nancy