NOTE: The cf-satellite mailing list is no longer active. The list archives are made available for historical reasons.
Hi John....
Just (finally) catching up....
> this seems like a good starting proposal to discuss. here are some issues
> from my POV:
>
> 1) "units(band)". I think if there are really different units, its probably
> a different variable. Do you have a concrete example where the units vary?
> Note also that "bands with different resolution" would also be a different
> variable (martin's question).
We have discussed this and are inclined to agree that if the units are
different, it should probably be a separate variable. So be it.
>
> 2) what is the mechanism for associating the scale/offset to the imageData?
The original idea was that scale/offset would be used to create
physical units from data values ("calibration"). The problem we
faced is that for each "band", these values would be different...but
there was no mechanism to have an array of values as an attribute of a
variable. Given that, we took the narrow view that the scale/offset
would apply to all variables that had the associated "band". I don't
really like that, but am not sure what else can be done...
> 3) you need
>
> ImageData:coordinates = "band lat lon time"
Yes.
>
> maybe even
>
> ImageData:coordinates = "band lat lon time wavelength"
Pardon my lack of understanding, but for a variable shaped as, say,
'time, band, x, y', how can one have a coordinate of both 'band' and
'wavelength'?
> 4) what is the meaning of band(band) ??
It was my naive interpretation from CF grid conventions of a
"coordinate variable" -- that's all....
tom
--
Tom Whittaker
University of Wisconsin-Madison
Space Science & Engineering Center (SSEC)
Cooperative Institute for Meteorological Satellite Studies (CIMSS)
1225 W. Dayton Street
Madison, WI 53706 USA
ph: +1 608 262 2759
cf-satellite archives: