NOTE: The cf-satellite mailing list is no longer active. The list archives are made available for historical reasons.
Tom, John Caron had encourage us to consider "band" to be a dimension and a > coordinate variable, so in my summary note, I went in that direction, > thinking that this would also allow us to define an array of > wavelengths, say, in one entity. For example, if I use the name > "bands" as a dimension, it might look like this: > > dimensions: > bands = 6; > variables: > double bands(bands) ; > bands:long_name = "center wavelength" ; > bands:standard_name = "wavelength"; > bands:units = "micrometer"; > bands:_CoordinateAxisType = "wavelength"; > > data: > bands = 2.5, 3.9, 7.8, 10.4.....; > > > I believe the goal here is to establish a pattern that applications > can easily recognize as being "satellite image data", which is why > we've stayed away from things like navigation and engineering data. I > think it would be literally impossible to define a convention that > encompasses all those things...but if we can get data providers to > structure the contents of the "image" data in their files according to > a CF Convention, then application developers have a good opportunity > when opening a file to know what they are dealing with (as opposed to > things like gridded data from, say, models). > I agree with your assessment. It will pretty impossible to cover all cases but I am sure we can come up with some common denominators that work for the majority of the satellite data around. What I like about your approach here describing the bands in one entity. There is probably a lot more metadata that could be dealt with that way. It would be nice to get some structure into those parameter by having some conventions around to define a block for generic parameters such as satellite, sensors and other acquisition relevant information. The solution using attributes for all variables is certainly working to some degree but does not help with organizing the metadata in some structured way. -- Best regards, Rudi
cf-satellite archives: