[CF-metadata] Use of number_of_days_with_air_temperature_below_threshold

From: Jonathan Gregory <j.m.gregory>
Date: Fri, 9 Aug 2013 16:47:34 +0100

Dear Martin

When these names were introduced, we took the view that the application of the
threshold is an operation which creates a new geophysical quantity. In that
respect it is analogous to the transformations mentioned in the standard name
guidelines, such as differentiation, integration, and calculating a histogram
or probability density function. Hence this operation does not have to be
described by the cell_methods, which makes it simpler. I think the number_of_
days data are quantities that have simple time bounds that specify the time
interval within which the number of days is calculated. They are extensive in
time i.e. they apply to the time interval as a whole, not to a point in time.
Hence the cell_methods should be just "time: sum". In fact this would be the
default for an extensive quantity, but CF recommends that cell_methods should
always be explicitly stated.

I hope this looks sensible to you, or at least acceptable.

Best wishes

Jonathan

----- Forwarded message from martin.juckes at stfc.ac.uk -----

> I just wanted to check on the usage of the standard name "number_of_days_with_air_temperature_below_threshold" for "frost day" and "ice day" counts. The definition of this standard name implies, I think, that cell_methods should be specified as "time: minimum within days time: sum over days" for frost days and "time: maximum within days time: sum over days". This usage would convey some of the relevant information, but I'm not sure it is consistent with the usage of cell_methods described in the CF standard itself. In the standard it is implied that a variable with cell_methods="time: maximum within days time: sum over days" would be the sum of the daily maxima. Here there is a thresholding operation in between which is omitted from the description only because "threshold" is not an available operation in table e.1.
>
> If "threshold" was added to table e.1 of the standard I suppose that "number_of_days_with_air_temperature_below_threshold" would become redundant and we could use "air_temperature" with cell_methods="time: minimum within days time: threshold within days time: sum over days". This would also allow us to deal with other indices (dry days, windy days) without creating new standard names. An additional attribute would be needed to describe the threshold in some way.
Received on Fri Aug 09 2013 - 09:47:34 BST