Further analysis (and help from Steve Hankin) revealed that I had misunderstanding on the constraints associated with the values elements in a coordinate variable could have. Just as long as successive values are increasing or decreasing, the coordinate variable element values are CF compliant.
Once I got past this issue, the rest of CF compliant data declaration fell into place. Here are the key aspects of the CF compliant solution
(1)
The vertical coordinates associated with the data set is in fact dimensional. The coordinates represent increasing pressure levels in hectopascals (hPa).
(2)
The vertical pressure levels reported are not contiguous, and, in fact, represent discrete samples based on need. Thus, the new conventions associated with discrete sampling geometries applies. The feature type associated with our product is "profile".
(3)
The axis attribute needed to be set to "Z" to ensure it is clear that the pressure level variable is the vertical dimension coordinate variable.
Following is an example:
attributes:
:featureType = ?profile?;
dimensions:
time = 1;
y = 10;
x = 10;
pressure_levels = 101;
variables:
short profile_data (time, y, x, pressure_levels);
.
.
:coordinates = "time y x pressure_levels";
.
.
float pressure_levels (pressure_levels);
:standard_name = ?...?
:long_name = ?pressure?
:units = "hPa";
:axis = "Z";
Note that in this data declaration, the profile_data at the 101 levels in the atmosphere at each x,y in the grid varies most rapidly. This is as a result of the inherent characteristics of the data processing required to generate the product file.
On Jan 24, 2012, at 1:00 PM, Randy Horne wrote:
> On the program I am working, we have temperature and pressure profile products where data values are generated for multiple pressure levels in the atmosphere (i.e. the z-axis). Dimensionless vertical coordinates as defined in paragraph 4.3.2 of the CF metadata document appears to be the best way to establish coordinates for the data values at the different pressure levels. Unfortunately, the various formula options defined in Appendix D. Dimensionless Vertical Coordinates do not work for the pressure levels that are reported in our temperature and pressure profile products. In fact, because of the non-linear and non-exponential characteristics of the reported pressure levels, it is unlikely any type of simple formula can be constructed to provide a mapping between monotonically increasing integers and reported pressure levels.
>
> Any thoughts as to what other options we have to handle this ?
>
>
>
> ..............End of Message ...............................-->
>
>
>
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
____________________________________
Randy C. Horne (rhorne at excaliburlabs.com)
Principal Engineer, Excalibur Laboratories Inc.
voice & fax: (321) 952.5100
url:
http://www.excaliburlabs.com
Received on Wed Jan 25 2012 - 17:16:00 GMT
