Re: NetCDF with GrADS


Subject: Re: NetCDF with GrADS
From: Brian Eaton (eaton@bearmtn-e0.cgd.ucar.edu)
Date: Mon Jan 25 1999 - 09:05:19 MST


In the netCDF files produced by ccm2nc the time coordinate is determined by
the current day and seconds variables (NDCUR,NSCUR) in the integer header
record, and has a value of type double which is equal to NDCUR +
NSCUR/86400. The units string has the value "days since yyyy:mm:dd
hh:mm:ss" where the date and time come from the base date and seconds
variables (NBDATE,NBSEC) in the integer header record. The smallest time
increment possible is of course 0.0 if the CCM history file contains more
than one time samples that have identical values of (NDCUR,NSCUR) in their
integer header records. The next smallest increment would be .00001157
days which is 1 second.

Hope that helps.
Brian

Brian Eaton
NCAR, Climate Modelling Section
P.O.Box 3000
Boulder, CO 80307
eaton@ucar.edu

>
> I used the ccm2nc program to convert a CCM history tape to NetCDF
> format (DJFdata.nc). When I attempted to read it into GrADS 1.7 (using
> the sdfopen command), the following error message was printed:
>
> Scanning self-describing file: DJFdata.nc
> Found 72 data variables in SDF file.
> Time unit has too small an increment for use with GrADS (min. 1 minute)
>
> Presumably the unit is seconds. Has anyone run into this problem before?
> It seems like either the ccm2nc program should have some kind of relevant
> option, or GrADS, but I don't know which. Any help would be appreciated.
>
> Bill Campbell
> UMD Dept. of Meteorology
>
>



This archive was generated by hypermail 2b27 : Thu Jun 01 2000 - 09:26:38 MDT