[THREDDS #UQE-222785]: year 0000 broken in TDS 4.3.17

1) Looking at 

SeaWiFS_Chla_Climatology_Monthly_1997_2010.nc

the TIME variable has

double TIME(TIME=12);
  :units = "1-JAN-0001 00:00:00";
  :time_origin = "hour since 0001-01-01 00:00:00";

problem is that "units" attribute is not a valid udunit time unit although 
"time_origin" is. Since this uses CF COnventions, its not recognized as a time 
coordinate.

2) Looking at

PF5_SST_Climatology_Monthly_1982_2008.nc

double TIME(TIME=12);
  :units = "hour since 0001-01-01 00:00:00";
  :time_origin = "1-JAN-0001 00:00:00";
  :modulo = " ";

this is ok as time unit, the coordinate values are:

hour since 0001-01-01 00:00:00
TIME =
  {366.0, 1096.485, 1826.97, 2557.455, 3287.94, 4018.425, 4748.91, 5479.395, 
6209.88, 6940.365, 7670.85, 8401.335}

 0001-01-16T06:00:00Z
 0001-02-15T16:29:05.999Z
 0001-03-18T02:58:12Z
 0001-04-17T13:27:18Z
 0001-05-17T23:56:24Z
 0001-06-17T10:25:30Z
 0001-07-17T20:54:36Z
 0001-08-17T07:23:42Z
 0001-09-16T17:52:48Z
 0001-10-17T04:21:53.999Z
 0001-11-16T14:51:00Z
 0001-12-17T01:20:06Z

is this incorrect? Anyway, I would expect that WMS would work with it, although 
i havent tested it.

John


> Hi John(s)
> 
> thanks for looking into this. Please find the staged netcdf files at:
> http://xfer.nmfs.hawaii.edu/lmoxey/temp/
> 
> Thanks,
> 
> Lucas
> 
> 
> 
> address@hidden> wrote:
> 
> > Hi John:
> >
> > can we get a copy of one of the actual files? thanks
> >
> > JOhn
> >
> > > Dear THREDDS developers,
> > > After NOAA OceanWatch upgraded their TDS on Friday to 4.3.17, we noticed
> > > that their satellite monthly climatologies were broken. These datasets
> > use
> > > dates with year 0000 as a convention to indicate that they are monthly
> > > climatologies: because they average over several years, there is no
> > > appropriate single year so 0000 is used to indicate this.
> > >
> > > TDS was working fine for these datasets prior to the 4.3.17 upgrade:
> > > OPeNDAP, NetCDF Subset Service (NCSS), Web Map Service (WMS), and Web
> > > Coverage Service (WCS) all worked (with the small hiccup that NCSS
> > returns
> > > different dates than WMS does since TDS and ncWMS use different
> > underlying
> > > time libraries, is what I've been told in the past).
> > >
> > > After the upgrade, OPeNDAP worked, NCSS did not provide any time
> > dimension
> > > options, and WMS GetCapabilities did not list any times either (so
> > Godiva2
> > > didn't work either). We use these monthly climatology services in our
> > > integrated map viewer (Voyager) at PacIOOS and hope that Unidata can fix
> > > this year 0000 issue. I have been working with Lucas Moxey at OceanWatch
> > on
> > > this upgrade (cc'd here). As a test, we tried switching the data to year
> > > 0001: while NCSS and WMS now recognize the time dimension, Godiva2 fails
> > to
> > > display anything. I would prefer to get it working with year 0000 if
> > > possible.
> > >
> > > Here are the datasets in question:
> > >
> > > http://oceanwatch.pifsc.noaa.gov/thredds/catalog.html?dataset=sw_clim
> > > http://oceanwatch.pifsc.noaa.gov/thredds/catalog.html?dataset=pf5_clim
> > > http://oceanwatch.pifsc.noaa.gov/thredds/catalog.html?dataset=aq_clim
> > >
> > > Thanks for giving this your attention!
> > > Cheers,
> > > John Maurer
> > > Pacific Islands Ocean Observing System (PacIOOS)
> > > University of Hawaii at Manoa
> > >
> > >
> >
> >
> > Ticket Details
> > ===================
> > Ticket ID: UQE-222785
> > Department: Support THREDDS
> > Priority: Normal
> > Status: Open
> >
> >
> 
> 

Ticket Details
===================
Ticket ID: UQE-222785
Department: Support THREDDS
Priority: High
Status: Open