Dear Jonathan,
Sorry, that was me being a bit slow on the uptake! Of course we don't want
to define the actual variable against two dimensions as the data itself is
still only really defined against one time dimension.
I have just tried the following example in CDMS to see how it deals with the
time dimension and it tries to encode "time_index" as the dimension and
loses the time:units attribute from which it understands time. Might we be
safer in defining the "time" dimension as the index that is also used for
"forecast_reference_time". Eg:
dimensions:
lat = 181 ;
lon = 360 ;
time = 10 ;
variables:
double time(time) ;
time:standard_name = "time" ;
time:units = "hours since 1991-01-01 00:00" ;
time:axis= "T" ;
double forecast_reference_time(time) ;
forecast_reference_time:standard_name =
"forecast_reference_time" ;
forecast_reference_time:units = "hours since 1991-01-01
00:00" ;
CDMS will then pick it up. I would argue for this version in line with our
previous discussion about using 'time' instead of 'forecast_validity_time'.
It means that the description will be more compatible with current (non
CF-compliant) software.
Kind regards,
Ag
> -----Original Message-----
> From: Jonathan Gregory [mailto:j.m.gregory at reading.ac.uk]
> Sent: 23 July 2003 17:28
> To: cf-metadata at cgd.ucar.edu
> Subject: [CF-metadata] Example of forecast data
>
>
> Dear Ag
>
> > I hadn't thought about the introduction of an index
> dimension to describe
> > the length of 'time' and 'validity time'.
>
> I thought this was our intention for the case where you had various
> combinations of analysis and forecast time e.g. 0, 6 and 12 h
> forecasts
> from an analysis at 0Z, 0 and 6 h from 6Z and 0 h from 12Z. That makes
> six altogether. Since you don't have every combination of analysis and
> forecast time, you can't have two orthogonal time axes. What
> I meant was
>
> dimensions:
> time_index = 6 ;
>
> variables:
> double time(time_index);
> time:units="hours since 2003-7-23";
> double forecast_reference_time(time_index);
> time:units="hours since 2003-7-23";
> float myvar(time_index, lat, lon);
> myvar:coordinates="time forecast_reference_time";
>
> data:
> forecast_reference_time=0, 0, 0, 6, 6, 12;
> time=0, 6, 12, 6, 12, 12;
>
> Best wishes
>
> Jonathan
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://www.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
Received on Thu Jul 24 2003 - 01:35:45 BST
