Dear all
A couple of weeks ago I made a proposal on CF trac at
http://cf-pcmdi.llnl.gov/trac/ticket/17
concerned with cell_methods. The proposal aims to extend and modify
the use of cell_methods in order to be precise about the areas over
which averages (in particular) are calculated. We discussed this at the CF
day in Paris last June and several people expressed views then. In the light
of that discussion, what I have proposed now is not quite the same as what I
described in June. I hope it is somewhat simpler, in fact.
Nonetheless this is quite a long trac proposal, and I'd be very grateful if
some people could read it and comment on it. Despite its seeming complicated,
it is not at all "exotic" or theoretical: it refers to a problem we have quite
often come across, for instance in the IPCC AR4 (CMIP3) database. A common
example is the need to distinguish between sea-ice thickness averaged over
the whole grid box, or just over the sea-ice area. E.g. if sea ice 2 m thick
covers half of an ocean grid box, its thickness averaged over the sea-ice area
is 2 m, and over the whole grid box is 1 m. Both of these definitions are in
use in datasets, and they need to be distinguished. At present CF doesn't have
a standard method for distinguishing them. That's what the proposal is about.
Please comment on the trac ticket rather than by replying to the email list.
Thanks for your help.
Jonathan
Received on Tue Nov 06 2007 - 01:11:24 GMT
