HOMME contains an executable "homme_tool" which can be used
to get access to a variety of HOMME features without running
a simulation.

scripts:
test.job       runs all tools at NE4 (similar to hommetool regression test)
template.job   create GLL SCRIP and subcell grid metadata files
cori.job       perform parallel interpolation of EAM history file output


Tools currently include:

================================================================
Topo converter  tool = 'topo_pgn_to_smoothed'

Use a pure-physgrid non-smoothed topography file to create a smoothed
GLL-physgrid topography file. This is part of the physgrid topography full
tool chain.

The input is an unsmoothed pg4 topography file output by
cube_to_target. The output is a GLL-physgrid file with smoothed topography
and GLL-physgrid consistent PHIS_d and PHIS fields, respectively. This
file is then input for a second run of cube_to_target, this time to
compute SGH, SGH30, LANDFRAC, and LANDM_COSLAT on physgrid.



================================================================
GLL topo smoother.  tool = 'topo_gll_to_smoothed'

Apply HOMME's laplace operator to smooth a topography data set.

Note: this tool can also be used to output (with no smoothing)
the topography used by HOMME's various build in test cases.
This is useful for generating topo files as par of tool
regression testing.

The newer topo_pgn_to_smoothed tool
should be used instead, as it will produce both GLL and pg2 data.


================================================================
Topo converter  tool = 'topo_convert'

Convert a pure-GLL topography file to GLL-physgrid topography file. This
is useful if you have an existing v1 GLL topo file and want to run an
equivalent physgrid-based simulation without having to run the topography
tool chain from scratch.



================================================================
Interpolation.  tool = 'interpolate_tool'

HOMME can read in an EAM native grid "ncol" or "ncol_d"
file and interpolate this to a lat lon file, with either bilinear interpolation
or native SE interpolation.  (no downscaling).


================================================================
Mapping files. tool = 'gll_mapping_file'

Output a mapping file that will reproduce the HOMME bilinear or native SE
interpolation algorithm.


================================================================
Template files.  tool = 'grid_template_tool'

SCRIP files are for use with SCRIP/ESMF mapping
utilities to generatye mapping weights.  The SCRIP files compute a control
volume around each GLL node, and perform a crude iteration to get the
control volume area to match the GLL weight. This iteration is exact for cubed-sphere
meshes, but only approximate for RRM meshes.  For RRM meshes the control volumes
can be star-like and are not recommended.  LATLON files give the connectivity
of the GLL "subcell" grid: the grid of quads whose corners are the GLL nodes.
This is useful for plotting native grid output.   This tool uses HOMME's
native grid output, and then the user has to run an NCL script to post-process
the output into the correct format.


================================================================
Template files: (alternative):  tool = "gll_subcell_grid"

This is the old code from CAM that would produce the SEMapping.nc file
for every run (since removed from EAM).  This is identical to the
LATLON file from the tool above.  This tool has the advantage that the
output does not have to be post-processed.  But it has the
disadvantage that the entire grid is stored in memory on MPI task 0 so
may not be scalable to ultra high resolution.

