This is the data set requested by Ron Gelaro
 and is obtained from '/archive/users/drholdaw/e5131_fsoi_DONE/obs:
 /archive/users/drholdaw/e5131_fsoi_DONE/obs/Y2014/M12/DXX/H15
 /archive/users/drholdaw/e5131_fsoi_DONE/obs/Y2014/M12/DXX/H21
 /archive/users/drholdaw/e5131_fsoi_DONE/obs/Y2015/M01/DXX/H15
 /archive/users/drholdaw/e5131_fsoi_DONE/obs/Y2015/M01/DXX/H21

 DJF_2014/2015 and JJA_2015 are new directories cre4ated for 
 txe & twe components:
 DJF_twe_2014_2015/
 DJF_txe_2014_2015/
 JJA_twe_2015/
 JJA_txe_2015


Karki
Mar 22, 2016

The change in organizing the data file structure was made after the following 
mail from Ron Gelero on March 22, 2016 and data was placed accordingly:

On 03/22/2016 10:30 PM, Ron Gelaro wrote:

     Hi Karki,

     Thanks for setting this up.  However, I'd like to change the placement
     of the files in the Dxx directories slightly if possible.  Currently,
     Dxx corresponds to the first date node in each filename.   In other
     words, the file:

     e5131_fsoi.imp3_txe_conv.obs.20141231_15z+20150102_00z-20150101_00z.ods

     falls in the directory M12/D31 because of the first node, 20141231_15z,
     in the filename.  I know this is the way the files are archived, but
     I've always thought this unfortunate, especially for outside users,
     because the valid time for the obs impact actually corresponds to the
     **last** (or third) date node in the file name, or 20150101_00z in this
     case.  So it would be much nicer for the users if the file above fell
     under the M01/D01 directory, and so on.

     Would it be possible to organize the files this way?  It would be a lot
     more intuitive for folks, especially for the filenames crossing month
     boundaries, i.e., the ones valid on the fist of the month.  Let me know
     if this overly complicates things for you.

     Thanks,
     Ron
---------------------------------------------------------------------------
---------------------------------------------------------------------------
Ron Gelaro asked again on Aug 23, 2016 to do the same data sataging on
Train but this time 12Z & 18Z period (set H03 & H09 from archive) from

#----------------------------------------------------------
# Copying 12Z & 18Z data for AUG 2015 from dirac under:
# /archive/users/drholdaw/e5131_fsoi/obs/Y2015/M08/D01/H03/*txe*
# and
# /archive/users/drholdaw/e5131_fsoi/obs/Y2015/M08/D01/H09/*twe*
# as requested by Ron Gelero
#
# NO NEED TO ADJUST DAYS FOR 12Z & 18Z UNLIKE 00Z & 06Z
#
# Karki
# Aug 25, 2016
# -------------------------------------------------------------------------

script for doinfg this is in dirac under: /home/mkarki:

get_obs_impact_data_to_Train_for_12_and_18Z_AUG_2015_Ron_Gelero.bat


Karki 
Aug 26, 2016