GZIP step needs improvements

[DanIredell-NOAA]

A couple of issues with the gzip step(s):

• it fails if the forecast is not exactly 4 or 8 days. It cannot handle any other number. It should be able to run if forecast is 1 day, which is the minimum for NCODA (actually 12 hours, but let's not do fractional days just yet).
• the gzip00 step takes 17 minutes to run, which is longer than the 15 minute threshold that NCO has set for jobs to be 'restartable' instead of restarting.

[DanIredell-NOAA]

Changed logic to use seq instead of brace expansion so can gzip any number of files as specified by fcstdays. old way: {06..96..6} # 96 is fcsthours but it cannot be a macro in a brace expansion new way: $(seq -w 06 06 $fcsthours)

Modified the gzip00 step to use pigz (parallel implementation of gzip) and it now takes 12 minutes to complete (5 minutes faster). This puts it under the 15 minute threshold. Will need to also investigate on using pigz for the gzip06 and gzip12 stepes, which take 11 minutes each.

[DanIredell-NOAA]

When using pigz, jobs gzip06 and gzip12 improve from 11 minutes to 8 minutes each.