Revise Ground Sampling Distance (GSD) definition

[strobpr]

I came across STAC's description of GSD here and here. Allow me a few remarks as since a while I try to clarify and harmonise fundamental geospatial and EO terminology:

gsd to my knowledge (going back 35 years) is not 'Ground Sample Distance at the sensor' but 'Ground Sampling Distance'. As the name says, it refers to 'on ground' NOT to 'at sensor'. You can define it as 'the nominal horizontal distance of adjacent samples at a reference altitude (usually sea level) in (usually) the center of the sensor's swath (or at Nadir)'. Sometimes it is not the same in x (along-track) and y (across-track) direction, so would be useful to allow two values as an option.

I very much appreciate the statement: "The value of this field should be related to the spatial resolution at the sensor, rather than the pixel size of images after orthorectification, pansharpening, or scaling." Only the 'should be' actually must read 'is' as this is a fact and it is not 'at the sensor' but 'of the sensor'.

Regarding the 'pixel size of images' I'll open a separate issue.