[Dart-users] major update to wrf model_mod

[Nancy Collins]

dear dart-users,

if you are not a wrf/dart user, you can toss this note now.

i've just committed a set of files under the $DART/models/wrf
directory which contain the recent updates to the wrf model_mod
and associated converter utilities.   Here is the contents of the
subversion update message:

First of several expected major updates to the WRF model_mod and 
related programs. THIS UPDATE IS NOT EXPECTED TO CHANGE THE OUTPUT
compared to previous runs.  It rearranges the initialization code,
makes the field selection namelist driven, and moves the bulk of
the functions in the old dart_tf_wrf program back into the model_mod
where it should be easier to maintain.

The dart state vector contents are completely namelist driven now;
the wrf field names, and order they are copied into the state vector,
are based on the values in the 'wrf_state_variables' namelist item.
The updated input.nml contains a namelist which is consistent with
previous model_mod defaults; if no namelist is specified, the dart
state vector will consist of:  U, V, W, T, QVAPOR, PH, MU.

dart_tf_wrf was split into two programs: dart_to_wrf and wrf_to_dart.
If 'wrf_state_bounds' is specified in the model_nml namelist, the
values will be clamped within the given range.  NULL means infinity.
The programs can also error out if a value is outside a valid range.

The advance_model.csh script was updated to call dart_to_wrf and
wrf_to_dart instead of dart_tf_wrf.  The older mkmf and path_names
files were removed and new ones added for the new converters.

Thanks to Josh Hacker for doing the code for this update.


This has been regression tested on several platforms with several different
wrf configurations, include V2 vs V3, regional vs global, different physics
options, and on different platforms.     as always, however, please email me
(nancy at ucar.edu) if you see something 'funny' after your next subversion
update.

thanks,
nancy