[CF-metadata] change_in

Jonathan Gregory j.m.gregory at reading.ac.uk
Mon Mar 20 01:38:47 MST 2006

Dear Bryan

> Can someone explain to me 
> (possibly again) why this isn't a tendency of something due to advection? )

It's got different units. change_in_X = tendency_of_X * time-interval

> 1) I don't like the chances of people working with data appreciating the 
> difference between intensive and extensive variables. 

It's just the same distinction as between precipitation_amount (kg m-2) and
precipitation_flux (kg m-2 s-1), for instance.

> 2) Limiting the use of change_in in this way is non-intuitive, again, while
> we might want to define it that way, it may get used in other ways. Not now 
> perhaps, but ...

That was my main reservation about it. Can you suggest a phrase which is less
likely to be generally interpreted than "change in"?

> 3) Whether or not the thing is intensive or extensive should
> be independent of whether or not it is tendency otherwise we're complicating
> our vocabulary. If there we need to introduce a new qualifier that would be 
> cleaner ... 

It really is a different quantity, so it needs its own standard name. It's not
a statistical description (like the cell_methods) nor a kind of ancillary
data (like standard error). We have plenty of other intensive/extensive
distinctions, because, as in this case, they have different units. For a
spatial example, consider sea_ice_area (m2) vs sea_ice_area_fraction (1).
The intensive/extensive distinction has been built into cell_methods for a
long time, where it determines the default interpretation.

Best wishes


More information about the CF-metadata mailing list