SHORTCOMINGs

1- Figuregen input files: The time header of the coarse mesh inputs is the cold start day
                          The time header of the fine mesh inputs is the end of the hindcast (spin-up)
   These two dates need to be written automatically. This feature is not supported yet.
   The user needs to write these dates manually whenever the job is submitted for the 1st time.

2- Check Coastal Process Research Group. 
   If the timesnap for Hs is 14:15:00, the cprg-HS-download.sh adds 4 or 5 hours to it based
   on the time zone (EST or EDT). Time zones are detected by the code and they are no the problem.
   if EST => 14:15:00 + 4 = 18:15:00
   if EDT => 14:15:00 + 5 = 19:15:00
   Time intercal for Hs is 3 hours.
   The next time step would be 22:15:00
   The next would be 01:15:00

   genvalidation_s1.sh script subtract 2 days from the 01:15:00 (local time converted to UTC)
   This 01:15:00 is constant. It is not a variable and it is not corrected by the system when
   time zone changes. Therefore, the user has to update it to 00:15:00 when time zone gets back
   to EST.

   Anther problem is posed when CPRG website operator changes the time format to 18:45:12!
   User has to detect it and change the constant in genValidation_s1.sh

   If the cycle 12UTC is used for validation, the start time of the validation model would 
   be 12 UTC while the observation start time is 00 UTC! 
   It is not a big deal, but!

   If genvalidation_s1.sh subtract 2 days from the beginning of the current cycle (not from a constant)
   and use a smarter way to cut the last two days of obs from the file (e.g. postProc/CPRG/obs-cur) 
   that contains the obs time series of all the time, these problems would disappear.