Dear Richard I think it is fine to encode these as a coordinate variable. In fact one of reasons why CF bounds are (n,2) rather than (n+1) is so that the bounds can overlap like yours do.
Best wishes Jonathan ----- Forwarded message from "Hattersley, Richard" <[email protected]> ----- > Date: Mon, 23 May 2016 12:40:31 +0000 > From: "Hattersley, Richard" <[email protected]> > To: "[email protected]" <[email protected]> > Subject: [CF-metadata] Overlapping bounds on coordinate variable > > We have a dataset that represents accumulations from a single, common start > time to successively later times. As such, the time values are unique, > numeric, and monotonic (as required for a coordinate variable), but the time > bounds all overlap: > > variables: > float stratiform_rainfall_amount(dim0, grid_latitude, > grid_longitude) ; > stratiform_rainfall_amount:standard_name = > "stratiform_rainfall_amount" ; > ... > int time(time) ; > time:bounds = "time_bnds" ; > ... > int time_bnds(time, bnds) ; > data: > > time = 1, 2, 3, 4, 5, 6 ; > > time_bnds = > 0, 2, > 0, 4, > 0, 6, > 0, 8, > 0, 10, > 0, 12 ; > > Is it recommended to encode these overlapping accumulations as a coordinate > variable? > > In the most relevant prior discussion I have found > (http://mailman.cgd.ucar.edu/pipermail/cf-metadata/2010/053823.html) the > decision seems to hinge on whether the data have any "particular order" or > whether it would "make sense to do an operation which depended on their being > in order". In our case there definitely is an order and it does make sense to > operate on them in that order. > > Richard Hattersley > _______________________________________________ > CF-metadata mailing list > [email protected] > http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata ----- End forwarded message ----- _______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
