[cf-pointobsconvention] Draft 2

[Jonathan Gregory]

Dear John

> >   3. z coordinate required, with values in any vertical coordinate system.
> >   4. z coordinate required, may be a "nominal" value (just a string description)
> >   5. No z coordinate required.
> > 
> > I would lean towards 4, but could be convinced of 5 or 3. We 
> > should however recommend the data provider add as much info as 
> > possible, and make sure there is a standard way to do so.
> 
> I think it should be 5, but if there is a Z value, it should be 3.
> In the atmosphere, raobs measure pressure and it is converted to
> height so the "real" z coordinate is not convertible to meters.
> And as I mentioned earlier, if you require 1 or 2, you cannot store
> satellite derived winds in that format because you have no height value.
> 
> ... if there is no Z value,
> what do I do if you go with 4?  Just have a string called "unknown"?

I agree with these arguments of Don's. Also, in a case where the data is on a
"special" level which can only be named, rather than given a numerical z-coord,
that should be indicated by the standard name (though I note John Graybeal's
reservation about this, which is that the same instrument might be producing
data of different stdnames when used for different purposes).

I agree with you, of course, that if the z-coord exists meaningfully it should
be provided, to make the data more useful. However, our usual kind of policy in
CF would be to make this a recommendation, rather than a requirement. There is
lots of optional metadata that good practice would include when relevant. Too
many *requirements* are a deterrent to adoption of the standard.

Best wishes

Jonathan