FMRC run times for SABGOM ROMS

[jbzambon]

The run times and forecast times are a bit wonky...

• For SABGOM ROMS the TimeCoverage start time is "Start: 1858-11-17T00:00:00Z" (attached screencap). Does this mean that one of the history files has _CoordinateModelRunDate set to this date? I can hunt this down, but I'd like to know what I'm looking for...

Current SABGOM ROMS NF Catalog

[rsignell-usgs]

@jbzambon , actually, that funny-looking date in 1858 is fine -- it's the start of Modified Julian Date, see here:
https://www.myroms.org/forum/viewtopic.php?p=225
and here
http://tycho.usno.navy.mil/mjd.html
It can be convenient to write all your netcdf files with the same base date, so that's a convenient one to use. Way better than days since 01-01-0000 that some data centers use. But seconds since 1970-01-01 00:00 or whatever is fine also. Just as long as it's units since some-date-that-is-after-most-of-the-world-agreed-on-gregorian-calendar 😸

[jbzambon]

So does this mean the ocean_time or dstart on some file is 0 since those only use "units: seconds since 1858-11-17 00:00:00", and this is polled to come up with the TimeCoverage on the catalog page? ocean_time/dstart reference the MJD calendar, the others reference a start time of 2010-12-31 (the actual first date of available model data). Sorry, I guess I'm not understanding where "TimeCoverage" is derived for the catalog? Its correct for the USEast COAWST-ROMS/SWAN catalog page and didn't require any changes to the xml or output files.

Thanks for the explanation of the MJD origin! I always show my classes this video from computerphile in the hopes no one will ever mess with time in their code 😀