[CF-metadata] Units: absolute vs difference

[John Caron]

Jonathan Gregory wrote:

>Dear John
>
>  
>
>>is there any convention for specifying that a quantity is a differences 
>>(like delta temperature) rather than absolute ?
>>    
>>
>
>There is not yet such a convention. I think the units should be unaltered.
>This is really an aspect of the description of the quantity, I'd say - it has
>to do with how it has been derived. A common case is where anomalies are
>expressed relatively to a climatology; this might be done with a large range
>of quantities and we may need a systematic way to indicate it. But we would
>need to be able to say what the climatology was as well.
>
>Could you describe the example you are dealing with?
>
>  
>

we're looking at WRF model output with temperature differences in 
Kelvin. unfortunately, our software package converts to Celsius by 
subtracting 273. We could fix if we knew it was a temp difference 
instead of an absolute temp.

we're trying to nudge WRF into using CF, which is why the question came 
to this forum. thanks for any thoughts.