On 19-02-15 18:56, Jim Biard wrote:
> Maarten,
>
> It would, instead be something like this:
>
> dimensions:
> lon = 360;
> lat = 180;
> layer = 18;
> vertices = 2;
>
> variables:
> float lat(lat);
> lat:long_name = "latitude";
> lat:units = "degrees_north";
> lat:bounds = "lat_bnds";
> float lon(lon);
> lon:long_name = "longitude";
> lon:units = "degrees_east";
> lon:bounds = "lon_bnds";
> float pressure(layer, lon, lat);
> pressure:long_name = "pressure grid";
> pressure:units = "hPa";
> pressure:bounds = "pressure_bnds";
> float lat_bnds(lon,vertices);
> float lon_bnds(lat,vertices);
> float pressure_bnds(layer,lon,lat,vertices);
> float O3(layer, lon, lat);
> O3:units = "1e-9";
> O3:coordinates = "pressure";
>
> You can have a layer index coordinate variable if you require/desire
> one, but it isn't necessary. The bounds attribute is applied to the
> pressure variable, and a coordinates attribute is applied to the O3
> variable naming the pressure variable. The pressure variable is an
> "auxiliary coordinate", and must be explicitly associated with the O3
> variable in this way.
Thanks, but my colleague has a remark: "this is very inefficient"
(although it is CF compliant).
Only the pressure boundaries are of interest, the pressures themselves
are pretty much a dummy variable, although a pretty large dummy
variable: 18*360*180 elements. Then the interfaces: 18*360*180*2. so
instead of storing (18+1)*360*180, the storage requirement jumps to
3*18*360*180, or 2.84 times as much. Not a fantastic score. Especially
for vertical coordinates this is a common situation
For regular coordinates the overhead is there, but at least it is a lot
smaller. For irregular grids, this overhead is large, and adds variables
no user will ever need. I'll think if I can think of a neat solution
that can be used in CF2.
Best,
Maarten Sneep
--
KNMI
T: 030 2206747
E: maarten.sneep at knmi.nl
R: A2.14
Received on Fri Feb 20 2015 - 04:05:39 GMT
