
This directory contains the Polvani-Scott-Thomas
primitive equation baroclinic instability test case.

To run the test, use the analog to "mpirun" that is
appropriate for your system on the preqx executable,
redirecting the standard input from the namelist file
"lmp.nl" that is contained in this directory.

The test case integrates for 12 days with 20 levels.  The l-2 norm
of the zeta field is printed in stdout, and it should
should converge to approximately

	2-norm relative vorticity = 0.78E-05



after 12 days.  It is possible this number will vary by
1-2%, which is still within the acceptable range for
validation.

The maximum semi-implicit time step for ne=9 (9x9 elements per 
cube face) is 700 seconds, and the explicit time step for ne=9 
is 60 seconds.  

ne9np8L20, strong-form viscosity:
original, non-conservative explict:  .77997e-5    (4/12/2007 mt)
semi-imp code: dt=600                .76793e-5    (9/23/2010 mt)
semi-imp code: dt=300                .78868e-5    (9/23/2010 mt)
semi-imp code: dt=60                 .79253e-5

ne9np8L20, weak-form viscosity
conservative, explicit code          .80310e-5     (9/23/2010 mt)

****************************************************************
HOMME regression test
Mark Taylor 12/2010

The script 'baro.job' can be used to run regression tests 
to test the semi-implicit and explicit codes, with PIO or
PIO_INTERP output.    NE=9, NV=8.  Semi-implicit/explicit is controlled by
the semi_imp=1,0 in baro.job.  PIO or PIO_INTERP output is controlled
by editing Params20.inc.  


baro.job   w/ semi_imp=1        2-norm relative vorticity (stdout)  0.788e-5
baro.job   w/ semi_imp=0        2-norm relative vorticity           0.803e-5

for PIO_INTERP, compare the zeta as plotted by zetall.ncl
with zeta_latlon.png.  For PIO, compare the zeta as plotted by 
zeta.ncl with zeta_cube.png





****************************************************************
mt 5/15/2007  

Added ncl scripts which will plot the Temperature and Vorticity
fields to match those in Polvani-Scott-Thomas MWR 2004.   
(same contour interval, plot the same zoomed in region.  
the colormap does not match)


T.ncl       plot temperature, from cubed-sphere netcdf file
zeta.ncl    plot zeta, from cubed-sphere netcdf file

Tll.ncl     plot temperature, from lat-lon netcdf file
zetall.ncl  plot zeta, from lat-lon netcdf file

PNG plots are also included showing these fields from a run:

NE=9 NPTS=8
weak-form viscosity
explicit time stepping
non-conservative SEM 
Robert = 0.0025

The plots are from two identical runs, one using PIO output on the
native grid, and one using the PIO_LATLON output.  Note that using
HOMME to output lat-lon netcdf files produces much nicer results.

(png's were created with by using ctrans to convert to postscript,
and then: pstoimg -crop lrtb -antialias -density 150 file.ps)

****************************************************************
mt 10/23/2008

Adding plot 
zeta_spherew.081104.png 
from SPHEREW (conservative) explicit code

NE=9 NPTS=8
weak-form viscosity
explicit time stepping
SPHEREW (conserative SEM)
Robert = 0.05

Initially I thought this version was producing poor results,
with very wigly contour lines in zetall.ncl.  But this turned 
out to be an interpolation issue:  Interpolating to a very
fine lat/lon grid (512x1024) produces excellent smooth results
while interpolating to a lon/lon 128x256 grid produces noisy
results.

Also verified that the semi-implicit code also produces similar results.



****************************************************************
mt 9/23/2010

Adding plot 
zeta_spherew.100923.png 
from SPHEREW (conservative) explicit code

NE=9 NPTS=8
weak-form viscosity
explicit time stepping
SPHEREW (conserative SEM)
Robert = 0.05












