20010725: extracting data from GEMPAK

>From: Cheryl Lacotta <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200107260442.f6Q4ge112467

>Hi!
>
>I have a rather large amount of data stored in GEMPAK files, *.gem. 
>Since there is so much data, I separated it into smaller files, so that
>the GEMPAK programs wouldn't choke on anything.  Problem is, I want to do
>a time series across all of these files.  For example, I want to plot the
>values of PW at Tucson for the three days that I am studying.  Is there a
>way I can extract the hourly values of PW at Tucson so that I may use
>Excel, or some other program to create one continuous plot?  Or is there
>an easier way to plot it in GEMPAK?
>
>Thanks!
>Cheryl
>
>
>


Cheryl,

The limiting factor you have in GEMPAK is the LLMXTM (number of
times) in a GEMPAK data management file which is set at
compile time in $GEMPAK/include/GEMPRM.PRM. The default is 300 times.

The number of data base entries is currently 30,000 (MMHDRS) which
is LLMXTM (300) + LLSTFL (29700). You can use SNMOD to create a file where you
move Tuscon observations to a single file, but for 3 years @ 2 soundings
per day, you would need LLMXTM > 2190. So, you could compile GEMPAK
with LLMXTM=2500 and LLSTFL=27500 for example.

Or, you can use SNLIST to list out the data you want, typically by writing
a script to automate the process for every time, and then pull out the data 
value and place in a text file with date/time that you could use with a 
spreadsheet.

I have used programs like MATLAB which are particularly good at handling
large matricies, such as the 50 years of sounding retirevals I show in
Journal of Atmospheric and Oceanic Technology (October 1994, Vol 11, No 5)
pages 1253-1261.

Steve Chiswell