Amendments to CF Conventions documents

[ros]

Whilst correcting some bugs in the CF checker I came across some errors/omissions in the CF Conventions document and Conformance Document. I propose the following changes.

2.5.1 Missing Data
The table in Appendix A incorrectly states that missing_value and _FillValue are Numeric attributes. This is inconsistent with CF 2.5.1, which says, "The scalar attribute with the name _FillValue and of the same type as its variable is recognized by the netCDF library as the value used to pre-fill disk space allocated to the variable."

The correction required is:

In Appendix A change the type from N to D for missing_value and _FillValue in the table.

3.3 Standard Name Modifiers
The requirements and recommendations for standard name modifiers are missing from the Conformance document.

Change Section 3.3 to

3.3 Standard Name
Requirements:
 * The standard_name attribute takes a string value comprised of a standard
   name optionally followed by one or more blanks and a standard name
   modifier.
 * The legal values for the standard name are contained in the standard name
   table. 
 * The legal values for the standard name modifier are contained in Appendix C,
   Standard Name Modifiers.

5.6 Grid Mappings and Projections
The requirements and recommendations for grid mappings are missing from the Conformance document.

Add a section

5.6 Grid Mappings and Projections

Requirements:
 * The type of the grid_mapping attribute is a string whose value is a single
   variable name.
 * The specified variable name (known as a grid mapping variable) must exist
   in the file.
 * The grid mapping variable must have the grid_mapping_name attribute. The 
   legal values for the grid_mapping_name attribute are contained in
   Appendix F.
 
Recommendations:
 * The grid mapping variable should have 0 dimensions.

Rosalyn

[jonathan]

Dear Ros

Thanks for spotting these. Although they are larger than typos, I agree that they should be regarded as defects rather than substantive changes. No doubt we just hadn't thought of character-type data variables as regards missing data, and the grid mapping was probably omitted from the conformance document by accident (the conformance document having been put together after most of the current standard).

Cheers

Jonathan