Gordon:

Ray asked me to send you some timegcm1.2 results for Resolute Bay
for days 260-365 of 2004, and days 1-105 of 2005. I post-processed
30 days at a time, saving ps plot files, and ascii data files 
containing the time coords and data corresponding to each plot frame.  

The files are at the following location:

http://download.hao.ucar.edu/pub/foster/gordon_shepard

Each ps file (and the corresponding ascii data file) contain 10
frames each with day of the year on the x-axis (hourly data),
vs height-integrated values of the field over Resolute Bay on the 
y-axis:

Frame	Variable
1	O2 O-O band emission (Rayleighs)
2	O2 O-O doppler TN (K)
3	O2 O-O doppler UN (m/s)
4	O2 O-O doppler VN (m/s)
5	O2 O-O doppler U+V vectors

6	OH-83 band emission (Kilo-Rayleighs)
7	OH-83 doppler TN (K)
8	OH-83 doppler UN (m/s)
9	OH-83 doppler VN (m/s)
10	OH-83 doppler U+V vectors

I think the .dat files give the vector sum of U+V for frames 5 and 10.

As you know, the model uses integer triplets day,hour,minute to specify 
the time of a history. The .dat files use number of hours since model
time 0,0,0. So time is as follows for each .dat file (this table is also
available in times.tbl at the above url:

File                            Model-time              File-time
(ascii data)                    (day,hr,min)            (hrs since 0,0,0)
-----------------------------------------------------------------
2004:
timegcm1.2_sy4j.d260-289.dat    260,0,0-289,23,0        6240-6959
timegcm1.2_sy4j.d290-319.dat    290,0,0-319,23,0        6960-7679
timegcm1.2_sy4j.d320-349.dat    320,0,0-349,23,0        7680-8399
timegcm1.2_sy4j.d350-365.dat    350,0,0-365, 0,0        8400-8760

2005:
timegcm1.2_sy4j.d001-030.dat      1,0,0-30,23,0           24- 743
timegcm1.2_sy4j.d031-060.dat     31,0,0-60,23,0          744-1463
timegcm1.2_sy4j.d061-090.dat     61,0,0-90,23,0         1464-2183
timegcm1.2_sy4j.d091-105.dat     91,0,0-105,23,0        2184-2543


Unfortunately, neither the model nor post-processor handle year-boundaries
very gracefully at this point. In this case, model times 365,1,0 to
365,23,0 of 2004 are missing. Ray may be willing to make a run to fill these
in later, but at present we are missing these 23 hours.

There is also a short IDL procedure at the url, rdascii.pro, as an example
of how to read the .dat files and make simple plots of each frame.

Let me know if you need help or have questions..

--Ben