I think that the important factor in the text posted by Jonathan: '''Generic applications should treat the data as missing where any auxiliary coordinate variables have missing values; special-purpose applications might be able to make use of the data.'''
is that it puts the onus on interpretation on the downstream application, with some advice; this is better than requiring the data provider to I feel that having the capability to have missing data stored in auxiliary coordinates is important. Whilst the interpretation of the meaning of such encoding needs some care, I feel that the specification should not be too restrictive on this. I have a use case to consider: Phenomena as Auxiliary Coordinates The distinction between a data variable and an auxiliary coordinate can often be quite an arbitrary one. A data variable may be used as an auxiliary coordinate for another data variable which shares the same coordinate variables. Consider - coordinates x, y, z and t - relative humidity data with respect to x, y, z and t: - pressure data with respect to x, y, z and t I would not want to alter the specific humidity data in any way as a result of adding the pressure data as an auxiliary coordinate. A post processing application may choose: - to interpret the pressure data auxiliary coordinate missing data as missing data indicators for the relative humidity data to enable data regridding onto specified pressure levels; - to compute the variability of the humidity measurements, ignoring the pressure data; these are operational choices, made by the software. I think CF should enable data to be stored to enable such processes to take place, and not mandate that one or other storage method is correct. -----Original Message----- From: cf-metadata-boun...@cgd.ucar.edu on behalf of Jonathan Gregory Sent: Sat 31/03/2012 21:19 To: cf-metadata@cgd.ucar.edu Subject: Re: [CF-metadata] CF-1.6 Conformance Requirements/Recommendations Dear all John Caron proposed > "Applications should treat the data as missing where the auxiliary coordinates are missing" and Steve proposed (an hour later, I think) "Application writers should be aware that under some (rare) circumstances data auxiliary coordinate values may be missing, while other parameters at the corresponding indices remain valid. While special purpose applications may be able to glean useful information at these indices, most applications will want to regard data as missing where the auxiliary coordinates are missing " I could agree to either of these. I prefer John's, because it is simpler, but it's more severe than Steve's. A compromise might be possible, e.g. Generic applications should treat the data as missing where any auxiliary coordinate variables have missing values; special-purpose applications might be able to make use of the data. Any good? Cheers Jonathan _______________________________________________ CF-metadata mailing list CF-metadata@cgd.ucar.edu http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata _______________________________________________ CF-metadata mailing list CF-metadata@cgd.ucar.edu http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata