[CF-metadata] 4. Re: New Standard Names for Satellite Data

From: Aleksandar Jelenak - NOAA Affiliate <aleksandar.jelenak>
Date: Tue, 30 Apr 2013 15:50:07 -0400

Dear Martin,

Thanks for taking the time to review the proposed names.

On Wed, Apr 17, 2013 at 2:36 AM, Schultz, Martin
<m.schultz at fz-juelich.de> wrote:
> "relative_platform_azimuth_angle" and "relative_platform_azimuth_angle": in my understanding "relative" denotes a (percent) fraction rather than a difference. Therefore, I think this term could be misleading. Wouldn't "platform_azimuth_angle_difference" be more precise?

The term "relative" is also used to describe difference, for example
"relative velocity" is a difference between two velocity vectors.
Using "relative" also keeps those names more familiar to the satellite
data community.

> Out of curiosity: isn't there a need to describe several other platform/sensor/solar/viewing angles?

Several such angles are already in the official table. My proposal
complements that and I have initially used the definitions from those
names as templates.

> I am wondering if "sensor zenith angle" and "sensor look angle" aren't the same, only shifted by 180 degrees.

No, the relationship is not that simple because these angles are
calculated approximating the Earth as an oblate spheroid. Below is a
new definition for sensor_zenith_angle that is hopefully more
graphical and less confusing:

Standard name: sensor_zenith_angle

Definition:
The angle between the line of sight to the sensor and the local zenith
at the observation target; a value of zero is directly overhead the
observation target. Local zenith is a line perpendicular to the
Earth?s surface at a given location. Observation target is a location
on the Earth defined by the sensor performing the observations.

> And, please excuse my ignorance, but "covariance_between_constant_and_linear_terms_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration" does sound rather specific and incomprehensible to me

I agree this standard name is unusual.

> Is this a general concept, or are these variables needed for one specific satellite sensor?

It is a general concept for one way of expressing the uncertainty of
satellite sensor inter-calibration. The CF convention does not
currently have a similar concept and the NetCDF-U convention does not
seem to have matured enough for use.

> Do we run the "danger" to see many more such proposals with other (inter)calibration concepts?

I don't see the "danger"; look how many standard names for atmospheric
chemical compounds are in the table now.

> I think, these definitions require a far more extensive description (i.e. definition), again, for example pointing to a web reference where the sensor concept and/or viewing geometry is described. It may thus be better from the standard_name perspective to try and find a somewhat more generalized term (if this is indeed sensor specific) and require/request a comment attribute which would then detail the exact procedure used. As an extreme (and perhaps unrealistic) suggestion, one could think about "covariance_between_correction_terms_due_to_intercalibration". The comment attribute would then have to specify that these are corrections of "radiance_per_unit_wavenumber", and that the covariance refers to "constant and linear terms".

The approach you outline appears to me as similar to what the NetCDF-U
convention tries to achieve. If this standard name is deemed too
unusual I am fine to drop it from the proposal.

       -Aleksandar
Received on Tue Apr 30 2013 - 13:50:07 BST