Dirk:

Ok, I've got a version of tigcmloc + my_doptuv working. There may
be a bug or two (see below), but I'm not sure if its mine or yours,
so I want you to try this out and see how it goes.

The code is in ~foster/tigcmloc/lumm.  (there will also be a copy of
this message in there). The tarsrc in that dir gets
the usual *.h *.f M* *.inp from the cwd, then also gets 6 source
files from /aurora/d/lumm/red/tigcmloc. (I had to make some changes
to my_doptuv.f, so it is in ~foster/tigcmloc/lumm). 

I advise starting out with a virgin directory, say ~lumm/tigcmloc
(or whatever). Copy *.f, *.h, M*, and *.inp from ~foster/tigcmloc/lumm
to the new dir (~lumm/tigcmloc or whatever). Then set up a tarsrc
file like mine (just reset the dest to whatever). You may also want
to make a couple changes in the Makefile.  Then I would try
running it interactively on shavano (rcp the tar file, tar xf, make,
pshell, etc). Note that the code requires 3 libraries from my
home on shavano (see the script ~foster/tigcmloc/lumm/tigcmloc):

/u0/foster/lib/libplt.a
/u0/foster/lib/libutil.a
/u0/foster/lib/getgcm.a

So you can just cp these over to $TMPDIR (copies are also on mss 
/FOSTER/lib). These are in addition to your disort.a, of course.
Then if you want to run batch jobs, set up a script similiar to
~foster/tigcmloc/lumm/tigcmloc.

I have completely rewritten pltdop, but it produces more or less
the same type of plots. As I recall, you already started working
with a tigcmloc I had here before you left, so you are already
familiar with the input file tigcmloc.inp and its differences from
locn.inp.  The doppler flag works as before, just set DOPPLER=1.
Note that you can control the y-axis, getting subsets of the pressure
scale, and/or specifying your own linear height scale (see ZP_RANGE
and HT_SCALE). This should work for the VER contour, but of course
does not apply to the doppler line plots.

Note that it will now interpolate between grid points if BILIN=1
(I have not tried this with DOPPLER=1).
Note also that you can specify zonal means or global means for
locations, using the special value 999. in LOC.
e.g., LOC = 57.5,-70., 57.5,999., 999.,999. This may be where I'm
getting tripped up in my_doptuv.  You'll notice that my_doptuv now
uses a getloc routine (in mkfields.f) that calculates and returns
zonal or global means if necessary (i.e., if ix=ifix(gmflag) for
zonal means, or ix=jx=ifix(gmflag) for global means). However, I
did not go into solred.f which references gcmlat(jx) and gcmlon(ix)
which would produce unknown results if ix or jx = 999. In short,
avoid zonal and global means in your locations, or else fix solred.f 
and whatever else needs fixing for that.

This all worked fine with subsets of the ut's you were working with.
But when I tested it with all the uts (13.2->13.12 every 15 minutes)
something weird happened at about ut 8.5 and 11.75 or so. See the plots
in /sting/d/foster/tigcmloc/lumm/gmeta.weird. But when I reran it
with DDHHMM = 13.7->13.12 (every 15 min), it appeared to be ok again.
That's when I decided to let you try it out. It may very well still
be my bug.

I may log in some this weekend, so send email with problems/questions.
In the end I think this will be a much better code to work with than
locn, and when your doppler stuff is ready, it will be easier to
add it to the permanent tigcmloc code for general use, etc.

Good luck,

--Ben