20010311: output

>From: Diana DeRubertis <address@hidden>
>Organization: UCAR/Unidata
>Keywords: 200103120009.f2C09NL13644

>Hi Steve,
>       I was away at a conference for two weeks, but downloaded that
>library and flstogem worked fine. I actually extracted sixty stations from
>the fsl data fr. the first CD, and plan to obtain these fr. the remaining
>CD's.
>       My goal, starting with the 1947-65 data first, is to calculate 7
>stab. indices for each sounding and gather these in some sort of output
>file, in which I can clearly access/identify the station, date and index.
>Should I run SNLIST for each station individually?  Ultimately, I want
>files that can be input into another program for statistical analysis
>(unless gempak can do this) on the entire period of record.
>
>I can give you more details on exactly what I'm trying to do, what my
>dissertation is about, etc.,  if that will help.
>
>Thanks again,
>Diana
>
>******************************************************************************
>
>Diana DeRubertis
>Dept of Geography
>University of California, Berkeley
>Office:  199 McCone Hall
>Email:  address@hidden
>
>


Diana,

You can run snlist with a list of stations in AREA, such as
AREA = @72469;72317;72304;etc
or
AREA = @DEN;GSO;HAT;etc

with DATTIM=all and OUTPUT=f/filnam will stick everything in one
file from the gempak file specified with SNFILE. With OUTPUT=f/filnam
in snlist, you will send the program output to "filnam" (by default,
without specifying filnam, the output will be a file called
the gempak programname.lis, eg snlist.lis). The output will overwrite
a file if it already exists, so you will want to set the filnam
accordingly for each time you run the program (probably 1 file
for each year).

But, I guess what you might want to consider is if your statistical
software (such as sas, splus, etc) would be easier to get all times
from one station first, rather than having times and stations
intermixed in an input file.

GEMPAK doesn't have any statistical analysis functions, so you most
likely will want to output and postprocess the data into some
other package. For simple statistics like means, standard deviations,
etc, you could either dump with snlist, or create a gempak program
specifically for that purpose- but most statistical programs have the
more complicated functions that would make it best to use
within them.

Typically, if you have to run a program lots of times for lots of stations,
you would set up a shell script like:

#!/bin/csh -f

set STNS=("72469" "72317" "72304")

@ YEAR = 47

while ($YEAR <= 65)
   foreach STN ($STNS)
      snlist << EOF
        snfile = data_19${YEAR}.gem
        dattim = all
        output = f/outfil_19${YEAR}_${STN}.dat
        snparm =
        stndex = lift;totl
        levels = 
        vcoord = pres
        mrgdat = y
        area = @${STN}
        r

        e
EOF
   end
   @ YEAR = $YEAR + 1
end


The above script would run snlist for ever year from 1947 to 1965
and for each station in the STNS list and create an outfil for
each station/year separately.



Steve Chiswell