Rename monitor files to avoid name collisions.

[EdwardSafford-NOAA]

Expected behavior
Files generated by the DA monitors should not be overwritten by successive runs.

Current behavior
The rocoto default output location for the DA monitors is $NOSCRUB/monitor/*mon/stats/$suffix/gdas.$pdy. There is no hour subdirectory. Most of the files produced by the monitors include a full cycle time, so they will are unique. However the files below do not include a cycle time and are overwritten with each cycle:

OznMon:
stdout.horiz.tar.gz
stdout.time.tar.gz

RadMon:
stdout.validate.tar
radmon_angle.tar.gz
radmon_bcoef.tar.gz
radmon_bcor.tar.gz
radmon_time.tar.gz

Machines affected
All machines.

To Reproduce
Run a cycled experiment for two or more cycles. The files indicated above will be overwritten with each cycle if the rocoto default output location is used.

Detailed Description
Make these name changes to the output files:
OznMon:
stdout.horiz.tar.gz ==> stdout.horiz.$pdate.tar.gz
stdout.time.tar.gz ==> stdout.time.$pdate.tar.gz

RadMon:
stdout.validate.tar ==> stdout.validate.$pdate.tar.gz
radmon_angle.tar.gz ==> angle.$pdate.tar.gz
radmon_bcoef.tar.gz ==> bcoef.$pdate.tar.gz
radmon_bcor.tar.gz ==> bcor.$pdate.tar.gz
radmon_time.tar.gz ==> time.$pdate.tar.gz

The RadMon changes will require corresponding changes in GSI-monitor, which will be done in NOAA-EMC/GSI-Monitor#37.

[WalterKolczynski-NOAA]

Is there a reason the path doesn't have a cycle after gdas.${PDY}? I think that would be the preferred solution.

[KateFriedman-NOAA]

@RussTreadon-NOAA added $cyc subfolders into the $TANKverf_rad* paths in the GFSv16.3.0 package @ 4335ef2 . Perhaps a similar change is needed in develop?

[EdwardSafford-NOAA]

OK, I'll add a $cyc subfolder instead then.