[CF-metadata] New Standard Names for Satellite Data

From: alison.pamment at stfc.ac.uk <alison.pamment>
Date: Thu, 6 Jun 2013 13:36:14 +0000

Dear Aleksander, All,

I have completed a review of the discussion of Aleksander's proposals (24 names in all) as a number of these are now in a state where they can be accepted for inclusion into the standard name table and others are very close to being agreed. Thank you to all those who have contributed to the discussion so far and thank you also to Aleksander for providing the wiki page at http://wiki.esipfed.org/index.php/Standard_Names_For_Satellite_Observations#Proposal_.232 which lists the names and definitions.

My comments and questions on each proposed name are given below. In some cases I have suggested cosmetic changes to the proposed definition to improve readability or to achieve consistency with the wording of existing definitions. In other cases I have suggested bigger changes to the text but always with the intention of increasing clarity rather than altering the substantive meaning of the terms. Names (6), (9) , (10), (11) and (12) are very close to being agreed and can be accepted if we can iron out the details of the definitions; most of them simply require a statement of the sign convention for 'angle' names. For names (13) - (18) I would like Aleksander to check the definitions, where I have added quite a bit of text for consistency between names, but the main point to be resolved is how the canonical units should appear in the standard name table. Names (7) and (8) are still under discussion ('view_angle' versus 'look_angle') but they can probably be agreed quite quickly. Proposals (19) - (22) are fo
r correction_due_to_intercalibration names and these are still some way from being agreed. Names (1) - (5) and (23) are accepted for inclusion in the standard name table.

The discussion of these proposals has raised two points that I think are of wider significance for standard names. Firstly, the use of the term 'spectral' which appears in a number of existing names. Currently this term is defined as follows: ' "spectral" means per unit wavelength or as a function of wavelength; spectral quantities are sometimes called "monochromatic". However, it was pointed out in this discussion that 'spectral' could equally well apply to quantities expressed as functions of wavenumber or frequency. Aleksander has neatly side stepped this issue by removing 'spectral' from his original proposals and replacing them with specific phrases such as 'per_unit_wavelength' and 'per_unit_wavenumber' which in turn allows the correct canonical units to be assigned to each quantity. I think this approach has significant advantages and I therefore propose to turn the existing 26 spectral radiance names into aliases of 'per_unit_wavelength' names. (The change would not apply to the existing sea_surface_
wave_spectral names which are defined differently.) Are there any objections to this change?

The second important point concerns the clear definition of angles, such as platform_azimuth_angle. For such quantities one needs to be clear about the reference direction and also about the location of the apex of the angle - is it at the satellite or at the observed point? Several contributors to the discussion were confused about the intended definitions until Aleksander provided a link to a diagram. Martin Schultz suggested that there should be a references to similar diagrams that would clearly illustrate the intended meaning of all angle names. I think this is an excellent idea and it would be useful to try to assemble such a document that could reside on the CF website and could then be referred to by the standard name definitions. Does anyone else think this would be useful? We already refer to external documents in some definitions, for example, many salinity names refer to TEOS-10 documentation and we refer to other vocabularies such as the area_type table and the FAO land cover classification syst
em.

---
1) sensor_band_identifier (Units: N/A)
'A variable with the standard name of sensor_band_identifier contains strings which give the alphanumeric identifier of a sensor band. These strings have not yet been standardised.'
Definition modified for consistency with those of other string valued standard names. This name is accepted for inclusion in the standard name table.
2) sensor_band_central_radiation_wavelength (m)
'sensor_band_central_radiation_wavelength is the central wavelength of a sensor's band, calculated as the first moment of the band's normalized spectral response function.'
This name is accepted for inclusion in the standard name table.
3) sensor_band_central_radiation_wavenumber (m-1)
'sensor_band_central_radiation_wavenumber is the central wavenumber of a sensor's band, calculated as the first moment of the band's normalized spectral response function.'
This name is accepted for inclusion in the standard name table.
4) sensor_band_central_radiation_frequency (s-1)
'sensor_band_central_radiation_frequency is the central frequency of a sensor's band, calculated as the first moment of the band's normalized spectral response function.'
N.B. The proposed unit of this name is Hertz, which in the standard name table would normally be written as s-1 (see, for example, the existing radiation_frequency name). It is fine to write Hz in the netCDF file because udunits identifies this with s-1.
This name is accepted for inclusion in the standard name table.
5) time_sample_difference_due_to_collocation (s)
' time_sample_difference_due_to_collocation  is the difference in time between two events that are collocated. Two events are deemed to be collocated based on some set of spatial, temporal, and viewing geometry criteria.'
This name is accepted for inclusion in the standard name table.
6) sensor_zenith_angle (degree)
'sensor_zenith_angle is 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" means a location on the Earth defined by the sensor performing the observations.  A standard name also exists for platform_zenith_angle, where "platform" refers to the vehicle from which observations are made e.g. aeroplane, ship, or satellite. For some viewing geometries the sensor and the platform cannot be assumed to be close enough to neglect the difference in calculated zenith angle.'
I have added a cross reference to the existing platform_zenith_angle name in the definition. The explanation of the distinction between the two names is taken from Aleksander's comments in the discussion of the proposal. I propose also to add a corresponding cross-reference to the definition of platform_zenith_angle. Also, I will modify the definitions of all existing platform names to include 'satellite' in the list of examples. Is this OK?
One further question: do we need to specify in which sense the angle is measured, e.g., 'The angle is measured clockwise starting from directly overhead'? We have done this sort of thing with some existing 'angle' standard names such as angle_of_rotation_from_east_to_x and it is included in the platform_azimuth_angle proposal. Incidentally, we don't seem to have included it in any of the existing platform names. Is this an omission?
This name is still under discussion, pending agreement on the details of the definition.
7) platform_look_angle (degree)
' "platform" refers to the vehicle from which observations are made e.g. airplane, ship, or satellite. Platform look angle is the angle between the line of sight from the platform and the direction straight vertically down. Zero look angle means looking directly beneath the platform.'
Aleksander and Jonathan have discussed an alternative name of 'platform_view_angle'. Aleksander, are you OK to use that version of the name?
Jonathan has also asked for the sign convention of the angle to be clarified in the definition (similar to my comments above on zenith angle).
This name is still under discussion.
8) sensor_look_angle (degree)
'The angle between the line of sight from the sensor and the direction straight vertically down. Zero look angle means looking directly beneath the sensor.'
As with (7), do we want to use 'look' or 'view'? What is the sign convention for the angle?
This name is still under discussion.
9) platform_azimuth_angle (degree)
' "platform" refers to the vehicle from which observations are made e.g. aeroplane, ship, or satellite. Platform azimuth angle is the horizontal angle between the line of sight from the observation point to the platform and a reference direction at the observation point, which is often due north. The angle is measured clockwise positive, starting from the reference direction. A comment attribute should be added to a data variable with this standard name to specify the reference direction. A standard name also exists for sensor_azimuth_angle. For some viewing geometries the sensor and the platform cannot be assumed to be close enough to neglect the difference in calculated azimuth angle.'
Definition modified to recommend specification of reference direction. This sentence should also be added to the definition of the existing solar_azimuth_angle name. Cross-reference added to sensor_azimuth_angle.
This name is still under discussion, pending agreement on the details of the definition.
10) sensor_azimuth_angle (degree)
'sensor_azimuth_angle is the horizontal angle between the line of sight from the observation point to the sensor and a reference direction at the observation point, which is often due north. The angle is measured clockwise positive, starting from the reference direction. A comment attribute should be added to a data variable with this standard name to specify the reference direction. A standard name also exists for platform_azimuth_angle, where "platform" refers to the vehicle from which observations are made e.g. aeroplane, ship, or satellite. For some viewing geometries the sensor and the platform cannot be assumed to be close enough to neglect the difference in calculated azimuth angle.'
Definition modified to recommend specification of reference direction. Cross-reference added to platform_azimuth_angle.
This name is still under discussion, pending agreement on the details of the definition.
11) relative_platform_azimuth_angle (degree)
'relative_platform_azimuth_angle is the difference between the viewing geometries from two different platforms over the same observation target. It is the difference between the values of two quantities with standard name platform_azimuth_angle. "Observation target" means a location on the Earth defined by the sensor performing the observations. A standard name also exists for relative_sensor_azimuth_angle. For some viewing geometries the sensor and the platform cannot be assumed to be close enough to neglect the difference in calculated azimuth angle.'
Definition expanded to reflect points made in the discussion and for consistency with definitions of other proposed satellite  names. Cross-reference to relative_sensor_azimuth_angle added. Jonathan has asked whether there is a sign convention for this name. The convention, or lack of one, should be stated explicitly in the definition.
This name is still under discussion, pending agreement on the details of the definition.
12) relative_sensor_azimuth_angle (degree)
' relative_sensor_azimuth_angle is the difference between the viewing geometries from two different sensors over the same observation target. It is the difference between the values of two quantities with standard name sensor_azimuth_angle. "Observation target" means a location on the Earth defined by the sensor performing the observations. A standard name also exists for relative_platform_azimuth_angle, where "platform" refers to the vehicle from which observations are made e.g. aeroplane, ship, or satellite. For some viewing geometries the sensor and the platform cannot be assumed to be close enough to neglect the difference in calculated azimuth angle.'
Definition expanded to reflect points made in the discussion and for consistency with definitions of other proposed satellite  names. Cross-reference to relative_sensor_azimuth_angle added. Jonathan has asked whether there is a sign convention for this name. The convention, or lack of one, should be stated explicitly in the definition.
This name is still under discussion, pending agreement on the details of the definition.
13) toa_outgoing_radiance_per_unit_wavenumber (mW m-2 sr-1 (cm-1)-1)
' "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics.'
Definition modified to make wording more consistent with existing 'toa_outgoing' names. OK?
I have a question about the units. I fully accept that you want to use mW m-2 sr-1 (cm-1)-1 in the netCDF files, but for canonical units we usually stick with basic SI units which for this quantity would be W m-2 sr-1 (m-1)-1. Udunits would convert quite happily between the two, and W m-2 is used for existing toa outgoing fluxes, so is it OK to list W m-2 sr-1 (m-1)-1 in the standard name table? Is there a strong reason for sticking with the proposed units?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
14) toa_outgoing_radiance_per_unit_wavelength (mW m-2 sr-1 um-1)
' "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics.'
Definition modified to make wording more consistent with existing 'toa_outgoing' names. OK?
As with (13) I have a question about units. Can we use canonical units of W m-2 sr-1 m-1 in the standard name table?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
15) toa_outgoing_radiance_per_unit_wavenumber_mean_within_collocation_target (mW m-2 sr-1 (cm-1)-1)
' toa_outgoing_radiance_per_unit_wavenumber_mean_within_collocation_target is an average of observations of the quantity with standard name toa_outgoing_radiance_per_unit_wavenumber from a sensor's adjacent field-of-views within a collocation target. "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics. The "collocation target" is an area on the Earth's surface at which observations from at least two sensors are collected. Its size is defined by the sensor with the largest field-of-view footprint. Two events are deemed to be collocated based on some set of spatial, temporal, and viewing geometry criteria.' 
Can we use canonical units of W m-2 sr-1 (m-1)-1 in the standard name table?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
16) toa_outgoing_radiance_per_unit_wavenumber_stdev_within_collocation_target (mW m-2 sr-1 (cm-1)-1)
' toa_outgoing_radiance_per_unit_wavenumber_stdev_within_collocation_target is the standard deviation of observations of the quantity with standard name toa_outgoing_radiance_per_unit_wavenumber from a sensor's adjacent field-of-views within a collocation target. "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics. The "collocation target" is an area on the Earth's surface at which observations from at least two sensors are collected. Its size is defined by the sensor with the largest field-of-view footprint. Two events are deemed to be collocated based on some set of spatial, temporal, and viewing geometry criteria.'
Can we use canonical units of W m-2 sr-1 (m-1)-1 in the standard name table?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
17) toa_outgoing_radiance_per_unit_wavenumber_mean_within_collocation_scene (mW m-2 sr-1 (cm-1)-1)
' toa_outgoing_radiance_per_unit_wavenumber_mean_within_collocation_scene is an average of observations of the quantity with standard name toa_outgoing_radiance_per_unit_wavenumber from a sensor's adjacent field-of-views within a collocation scene. "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics. The "collocation scene" is a grouping of a sensor's adjacent field-of-views centered on a collocation target. The "collocation target" is an area on the Earth's surface at which observations from at least two sensors are collected. Its size is defined by the sensor with the largest field-of-view footprint. Two events are deemed to be collocated based on some set of spatial, temporal, and viewing geometry criteria. The collocation scene's size is typically about twice that of the collocation target.'
Can we use canonical units of W m-2 sr-1 (m-1)-1 in the standard name table?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
18) toa_outgoing_radiance_per_unit_wavenumber_stdev_within_collocation_scene (mW m-2 sr-1 (cm-1)-1)
' toa_outgoing_radiance_per_unit_wavenumber_stdev_within_collocation_scene is the standard deviation of observations of the quantity with standard name toa_outgoing_radiance_per_unit_wavenumber from a sensor's adjacent field-of-views within a collocation scene. "toa" means top of atmosphere. The TOA outgoing radiance is the upwelling radiance, i.e., toward outer space. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics. The "collocation scene" is a grouping of a sensor's adjacent field-of-views centered on a collocation target. The "collocation target" is an area on the Earth's surface at which observations from at least two sensors are collected. Its size is defined by the sensor with the largest field-of-view footprint. Two events are deemed to be collocated based on some set of spatial, temporal, and viewing geometry criteria. The collocation scene's size is typically about twice that of the collocation target.'
Can we use canonical units of  W m-2 sr-1 (m-1)-1 in the standard name table?
This name is still under discussion, pending agreement on the canonical units and cosmetic changes to the definition.
19) constant_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration  (mW m-2 sr-1 (cm-1)-1)
' constant_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration is the constant term (offset) of the formula for correcting measured spectral radiance expressed as a function of wavenumber. The correction is derived from intercalibration between the monitored and the reference sensor. The resulting corrected spectral radiance of the monitored sensor becomes comparable with measured spectral radiance of the reference sensor. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics.'
Can we use canonical units of W m-2 sr-1 (m-1)-1 in the standard name table?
Discussion between Martin and Aleksander has established that this is a quantity that is generally applicable to remote sensing observations. Martin has asked for the definition of the covariance name (22) to be expanded and I wonder whether we need some more detailed description for this name too. Are there are many 'intercalibration' algorithms and should the data variable be accompanied by a comment attribute to say which algorithm was used, or give a reference to it? Or is there one generally accepted method that we can reference in the definition? It is fine to have quite detailed explanations in the text, for example, some of the sea_surface_wave spectral names have very specific definitions. Do we need something like that for this quantity?
This name is still under discussion.
20) linear_term_of_spectral_radiance_correction_due_to_intercalibration (1)
' linear_term_of_spectral_radiance_correction_due_to_intercalibration is the linear term (slope) of the formula for correcting measured spectral
radiance. The correction is derived from intercalibration between the monitored and the reference sensor. The resulting corrected spectral radiance of the monitored sensor becomes comparable to measured spectral radiance of the reference sensor. "Spectral" means per unit wavenumber or as a function of wavenumber. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics.'
My comments for (19) also apply to this proposal.
I note that in http://wiki.esipfed.org/index.php/Standard_Names_For_Satellite_Observations#Proposal_.232 all the other 'spectral' names have been altered to use 'per_unit_wavelength' or 'per_unit_wavenumber' so that correct units can be assigned. Units of '1' for this quantity may mean that in some sense it is less necessary to specify whether the spectrum is given in terms of wavelength, wavenumber, or frequency, but I think for consistency with the other proposed names it would be better to use 'per_unit_wavenumber' for this name too. Also, the definition on the wiki page for the covariance name does refer to linear_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration, so perhaps the intention was to modify this name and it got missed accidentally?
This name is still under discussion.
21) quadratic_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration (mW-1 m2 sr cm-1)
' quadratic_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration is the quadratic term of the formula for correcting measured spectral radiance expressed as a function of wavenumber. The correction is derived from intercalibration between the monitored and the reference sensor. The resulting corrected spectral radiance of the monitored sensor becomes comparable to measured spectral radiance of the reference sensor. Radiance is the radiative flux in a particular direction, per unit of solid angle. In accordance with common usage in geophysical disciplines, "flux" implies per unit area, called "flux density" in physics.'
My comments for (19) also apply to this proposal.
This name is still under discussion.
22) covariance_between_constant_and_linear_terms_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration (mW m-2 sr-1 (cm-1)-1)
' Covariance between constant_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration and linear_term_of_radiance_per_unit_wavenumber_correction_due_to_intercalibration values.
Martin has asked for the definition of this name to be expanded (http://mailman.cgd.ucar.edu/pipermail/cf-metadata/2013/056420.html). Aleksander said that this proposal could be dropped if it is 'too unusual'. I'm not sure that we need to drop the name, but I agree with Martin that more explanation is needed. If all the intercalibration names are indeed generally used remote sensing quantities then they probably do need standard names, but I think we do need some more detail and a reference about the intercalibration method used.
This name is still under discussion.
23) toa_brightness_temperature_of_standard_scene (K)
' "toa" means top of atmosphere. The brightness temperature of a body is the temperature of a black body which radiates the same power per unit solid angle per unit area at a given wavenumber. The standard scene is a target area with typical Earth surface and atmospheric conditions that is accepted as a reference. The toa radiance of the standard scene is calculated using a radiative transfer model for a given viewing geometry. The resultant toa spectral radiance is then integrated with a sensor's spectral response function and converted to equivalent brightness temperature.'
I don't think anyone has offered comments on this name. I think the name and its definition are consistent with existing brightness temperature names and so it is accepted for inclusion in the standard name table.
24) toa_brightness_temperature_bias_at_standard_scene_wrt_intercalibration (K)
'The difference between top-of-atmosphere (TOA) brightness temperature of the reference sensor and TOA brightness temperature of the
monitored sensor. This TOA brightness temperature difference is a measure of the calibration difference between the monitored and
reference sensors. Standard scene is a target area with typical Earth surface and atmospheric conditions that is accepted as a reference. Brightness temperature of a body is the temperature of a black body which radiates the same power per unit solid angle per unit area at a given wavenumber. TOA brightness temperature of the standard scene is calculated using a radiative transfer simulation for a given viewing
geometry. The resultant top-of-atmosphere spectral radiance is then integrated with each sensor's spectral response function and converted to equivalent brightness temperature.'
Jonathan has suggested a reordering of this name to be bias_in_toa_brightness_temperature_of_standard_scene_wrt_intercalibration and has also asked whether the 'wrt_intercalibration' is really needed. Are other kinds of bias likely to be calculated? If so then we should keep 'intercalibration' in the name. Also, shouldn't it be 'due_to' rather than 'wrt' as in the other names?
This name is still under discussion.
---
I am currently preparing the next update of the standard name table. The accepted names in this thread will, of course, be included. If we can resolve the outstanding questions about the names that are close to agreement then many of these could also be accepted and included in the next update.
Best wishes,
Alison
------
Alison Pamment                          Tel: +44 1235 778065
NCAS/British Atmospheric Data Centre    Email: alison.pamment at stfc.ac.uk
STFC Rutherford Appleton Laboratory     
R25, 2.22
Harwell Oxford, Didcot, OX11 0QX, U.K.
-- 
Scanned by iCritical.

Received on Thu Jun 06 2013 - 07:36:14 BST