[CF-metadata] newbie questions about degree_days

From: Jonathan Gregory <j.m.gregory>
Date: Mon, 5 Dec 2005 14:40:09 +0000

Dear Keith

I agree with what Bryan says about putting some of the metadata in other
places apart from the standard name. I would suggest that we introduce
standard names of integral_of_temperature_excess_wrt_time and
integral_of_temperature_deficit_wrt_time. This is the idea of degree-day, but
we shouldn't imply that the quantity should necessarily be measured in a
particular unit. It could be in degC day or degF day or K s, for instance,
and more generic names therefore would seem suitable to me. The two names are
for positive and negative degree-days, respectively.

To define the threshold I would propose that you use a single-valued coordinate
variable. As usual this can be done in CF with either a size-one dimension or
a scalar coordinate variable. We need a standard name for this too. It could
be called temperature_threshold, for instance.

The time period over which the integral is done should be specified with the
bounds of the time coordinate variable. The cell_methods doesn't need to be
supplied as the default interpretation (that it applies to the entire time
interval, since it's an extensive quantity) is correct. It could be shown
explicitly with a cell_methods of "sum".

Best wishes

Jonathan
Received on Mon Dec 05 2005 - 07:40:09 GMT