There is apparently yet another metadata standard for Earth data: The Generic Earth Observation Metadata Standard. This is being used by the EVCD, but this is supported by NASA too, so we may want to implement it. Most of this is just most attributes, but the two things that I see that we will need to be aware of:
- They require "classic" mode netCDF4 files, which forbids use of features not in the netCDF3 API. I think this means no groups or string variables.
- Times are supposed to be in days since Jan 1, 2000 without leap seconds. I'm not terribly thrilled replacing our time variable with this (because a Unix-like timestamp is easier to handle IMO) but maybe we can add an alternate time variable. N.B.: Python generally ignores leap seconds too.
For all the metadata, we'll probably need an additional file that specifies the attribute values that vary per-site (PI name, PI institution, etc.).
There is apparently yet another metadata standard for Earth data: The Generic Earth Observation Metadata Standard. This is being used by the EVCD, but this is supported by NASA too, so we may want to implement it. Most of this is just most attributes, but the two things that I see that we will need to be aware of:
For all the metadata, we'll probably need an additional file that specifies the attribute values that vary per-site (PI name, PI institution, etc.).