Table 1-02, revise radiation units

[JREyre]

Initial request

Revise units for radiation variables

Amendment details

Current notations, names, definitions of units relevant to radiation

(internal) id	notation	name	definition
183	W.m-1.sr-1	watt per metre per steradian	SI derived unit of spectral intensity; SI base unit equivalent: kg.m.s-3
184	W.m-2	watt per square metre	SI derived unit of heat flux density or irradiance; SI base unit equivalent: kg.m-1.s-3
185	W.m-2.sr-1	watt per square metre per steradian	SI derived unit of radiance; SI base unit equivalent: kg.s-3
186	W.m-2.sr-1.cm	watt per square metre per steradian per inverse centimetre	SI derived unit of spectral radiance per wave number, wave number expressed in 100 m-1; SI base unit equivalent: 0.01 kg.m.s-3
187	W.m-2.sr-1.m	watt per square metre per steradian per inverse metre	SI derived unit of spectral radiance per wave number; SI base unit equivalent: kg.m.s-3
188	W.m-3.sr-1	watt per cubic metre per steradian	SI derived unit of spectral radiance; SI base unit equivalent: 100 m-1.kg.s-3

Updates to descriptions

(internal) id	notation	name	definition
183	W.m-1.sr-1	watt per metre per steradian	SI derived unit of spectral intensity, a directional quantity; SI base unit equivalent: kg.m.s-3
184	W.m-2	watt per square metre	SI derived unit of heat flux density or irradiance; SI base unit equivalent: kg.s-3
186	W.m-2.sr-1.cm	watt per square metre per steradian per inverse centimetre	SI derived unit of spectral radiance per unit wavenumber, wavenumber expressed in 100 m-1; equivalent to W.m-2.sr-1.(cm-1)-1; SI base unit equivalent: 0.01 kg.m.s-3
187	W.m-2.sr-1.m	watt per square metre per steradian per inverse metre	SI derived unit of spectral radiance per inverse metre; SI base unit equivalent: kg.m.s-3

Update to name and description

(internal) id	notation	name	definition
188	W.m-3.sr-1	watt per square metre per steradian per metre	SI derived unit of spectral radiance per unit wavelength; SI base unit equivalent: kg.m-1.s-3

New code

(internal) id	notation	name	definition
191	W.m-2.sr-1.Hz-1	watt per square metre per steradian per hertz	SI derived unit of spectral radiance per unit frequency; SI base unit equivalent: kg.s-2

Comments (external reviewers)

No response

Requestor(s)

John Eyre, on behalf of TT-WIGOSMD

Stakeholder(s)

Enter list of stakeholder(s).

Publication(s)

http://codes.wmo.int/common/unit
http://codes.wmo.int/wmdr/unit

Expected impact of change

LOW

Collaborators

@joergklausen @JREyre @amilan17

References

http://codes.wmo.int/common/unit
https://en.wikipedia.org/wiki/Irradiance#SI_radiometry_units

Validation

Internal review

[JREyre]

@joergklausen @amilan17
I have created this new Issue to revise the radiation units. It will need careful checking as these units are quite tricky.

[amilan17]

https://github.com/wmo-im/tt-wigosmd/wiki/2022.12.08-TT-WIGOSMD notes:
@gaochen-larc will review

[joergklausen]

I advise against deprecating recently added or updated terms unless there is compelling reason. Material errors need of course to be corrected. Mere duplication doesn't seem to be a problem for me in this case, as different communities may use either unit. @JREyre has a point about (cm-1)-1 being more meaningful, yet we had decided to simplify notations as much as possible.

• 183: is correct, see also https://en.wikipedia.org/wiki/Irradiance#SI_radiometry_units
• 184: description needs to be corrected to read "SI derived unit of heat flux density or irradiance; SI base unit equivalent: kg.s-3"
• 185: no problem raised.
• 186: is correct, changing a notation is not advisable unless it is wrong. Description could be updated to read "SI derived unit of spectral radiance per wave number, wave number expressed in 100 m-1; equivalent to W.m-2.sr-1.(cm-1)-1; SI base unit equivalent: 0.01 m.kg.s-3"
• 187: is correct as stated.
• 188: description needs to be corrected to read "SI derived unit of spectral radiance; SI base unit equivalent: kg.m-1.s-3"

[JREyre]

@joergklausen
Thanks, Joerg.
I've checked your corrections to SI units for 184 and 188.

[amilan17]

https://github.com/wmo-im/tt-wigosmd/wiki/2023.01.12-TT-WIGOSMD notes:
@JREyre will double check bullet points in proposal and in discussions and then update proposal summary; @gaochen-larc will review and update branch

[JREyre]

@amilan17 @joergklausen
I have reviewed the state of this proposal. I suggest that we withdraw it from this fast track because it needs more discussion. My suggestions:
183. No change (i.e. remove proposal to delete or deprecate). However, its units are the same as 187, which we have agreed is defined as "spectral radiance". Are we happy for 183 to be called "spectral intensity"?
184. Description needs to be corrected to read "SI derived unit of heat flux density or irradiance; SI base unit equivalent: kg.s-3". (Joerg's proposal)
185. No change.
186. Description should be updated to read "SI derived unit of spectral radiance per wave number, wave number expressed in 100 m-1; equivalent to W.m-2.sr-1.(cm-1)-1; SI base unit equivalent: 0.01 kg.m.s-3". (Joerg's proposal)
187. This is formally correct as it stands, except in the definition: "per wavenumber" should be "per inverse metre". However, this is never used - no-one uses "inverse metre", only "inverse centimetre" = "wavenumber". I suspect that this is a mistake, and that it was the original intention here to define spectral radiance per metre, as used by the short-wave community. In this case, it would be: "187. W.m-2.sr-1.m-1. Watt per square metre per steradian per metre. SI derived unit of spectral radiance per metre. SI base unit equivalent: kg.m-2.s-3".
188. Reconsider following discussion of 187. 188 and proposed 187 are effectively the same. Units of 188 should be kg.m-2.s-3. Name is not meaningful - should be "Watt per square metre per steradian per metre".

[gaochen-larc]

Following up with @JREyre, @joergklausen, and @amilan17:

183. I am happy with "spectral intensity", which is consistent with the link cited by @joergklausen. We may want to change the notation from "W.m-1.sr-1" to "W.sr-1.m-1" and the name from "watt per metre per steradian" to "watt per steradian per metre". This reflects that "intensity" is a directional quantity. This is a rather minor and subjective change. For definition, we may consider changing to "SI derived unit of spectral intensity per unit wavelength; SI base unit equivalent: kg.m.s-3" to be clearer.

184. Also agree with the correction proposed by @joergklausen to definition: "kg.m-1.s-3" to "kg.s-3"

185. Agree with "no change".

186. Also agree with @joergklausen proposal. The updated version should read: "SI derived unit of spectral radiance per unit wavenumber, wavenumber expressed in 100 m-1; equivalent to W.m-2.sr-1.(cm-1)-1; SI base unit equivalent: 0.01 kg.m.s-3". Made minor changes to definition: adding "unit" and changing to "wavenumber".

187. Agree with @joergklausen. I did some checking found the wavenumber unit described as either "inverse metre" or "reciprocal metre". Also, "cm-1" is the conventional unit. "m-1" is the SI unit. We may consider adding a unit based on "cm-1" if someone makes a request. It is true that "cm-1" is widely used in the literature. The "per wave number" should be changed to "per unit wavenumber".

188. This is unit is similar to 187, except it is per unit wavelength. Considering further minor change to @joergklausen proposal: "SI derived unit of spectral radiance per unit wavelength; SI base unit equivalent: kg.m-1.s-3"

These units are a little tricky. It is necessary to have someone double check these again... Sorry for the delay

[amilan17]

@joergklausen @JREyre @gaochen-larc if three of you come to an agreement this week, then I believe that Gao or I can update this branch for FT2023-1. Otherwise, I'm ok taking more time for discussions and changing the milestone.

[joergklausen]

@JREyre @gaochen-larc This terminology is truly mind-boggling ...

183. Agree with 'spectral intensity', and @gaochen-larc 's suggestion for the name and definition is fine with me. To accomodate the directional aspect, I have added '[...], a directional quantity'. I am less happy with changing the notation, because (a) it was only recently introduced and, (b) we should be in alignment with http://codes.wmo.int/common/unit. Proposal updated.
184. non-controversial. Proposal updated.
185. non-controversial. No change
186. 'unit wavenumber' is fine. Proposal updated.
187. 'per inverse meter' is fine. Proposal updated.
188. Correcting my earlier comment, the description must include '[...] per metre; [...]'. The name suggested by @JREyre makes sense to me. Proposal updated.

I hope this covers it.

[gaochen-larc]

I concur with @joergklausen, We can leave 183 notation unchanged.

187: could we make a small change to definition: "per inverse metre" to "per unit wavenumber"? This would make things more consistent.

[JREyre]

@gaochen-larc @joergklausen
I'm not clear what you are suggesting about "spectral intensity". Are you agreed that "spectral intensity" and "spectral radiance" are the same variable? Is it OK to call it by one name in one place (183) and by another name in the other places (186, 187, 188)?

[gaochen-larc]

@JREyre. I think "Spectral Intensity" is different from "Spectral Radiance". "Spectral Intensity" can be viewed as "Spectral Radiance" per area. Does this make sense? https://en.wikipedia.org/wiki/Radiant_flux

[JREyre]

@gaochen-larc @joergklausen
Thanks for the link. You will see that "spectral intensity" appears twice is this table - once as equivalent to spectral radiance and once without "per area". In our table, at present, they both have the same units and so are equivalent. I have to admit that I have always used "spectral radiance" but I have been happy to talk also about it as "intensity", meaning spectral intensity, particularly within the radiative transfer equation. In this case, both have units of W.m-2.sr-1.(cm-1)-1 in wavenumber space and W.m-2.sr-1.m-1 in wavelength space.

[gaochen-larc]

I will check my radiation textbook to see if we can avoid the confusion... Thank you, @JREyre !

[gaochen-larc]

@JREyre We may want to have a Teams meeting. This is a little too hard, haha

"You will see that "spectral intensity" appears twice is this table" I guess you meant in the left column right? I can find only one "Spectral intensity" in the left column. The one equivalent to "Spectral radiance" is "Specific intensity", not "Spectral intensity". Maybe I read it wrong, haha.

[joergklausen]

@JREyre @gaochen-larc According to wikipedia, there is a difference and I hope, I have used the terminology correctly. I think this could be resolved with the help of ISO 9288:2022, but I don't have access to that (@PeterBlattner: do you?). Overall, I am a little concerned that there are no users engaged in this discussion. They should know ... we don't want this to be of theoretical interest but of practical importance.

[JREyre]

@joergklausen I've been a user of "spectral radiances" for more years than I wish to remember, in the context of retrievals of atmospheric variables from and NWP-assimilation of satellite observations. The problem with users (like myself) is that we are often very sloppy with terminology - my community always says "radiances" when strictly we should say "spectral radiances". I am happy to speak for the IR and MW communities here. I have less experience of what the shortwave community does. The thermal IR community uses W.m-2.sr-1.(cm-1)-1. The shortwave community will use a "per wavelength" unit, but I don't know which one they prefer or whether they all use the same one. As I said above, it is normal to calibrate visible and near-IR radiances into reflectances.

[joergklausen]

Dear @JREyre I know you are a power user yourself, and if you are confident you can speak for the community in terms of terminology, then I trust your word completely. In the wikipedia reference to SI units, the terms 'spectral intensity' and 'specific intensity' are different quantities with different units. It is only in the notes where the wikipedia article speaks of 'confusingly called "spectral intensity"'. We may want to use the terminology 'spectral radiance (=specific intensity)' instead of 'spectral radiance' to bring this to a conclusion.
Please let me know if this is acceptable and I will update the proposal, ask @amilan17 to update the branch, and then create a PR.

[JREyre]

@joergklausen @gaochen-larc @amilan17
Thanks for the link to the wikipedia table, which is interesting. It gives 2 units for spectral radiance:

• one appropriate to the microwave: W.m-2.sr-1.Hz-1
• one appropriate to the shortwave: W.m-2.sr-1.m-1
  It does not give:
• one appropriate to the thermal IR: W.m-2.sr-1.(cm-1)-1.
  The last one is linearly related to the first one: wavenumber = cm-1 = frequency in Hz / speed of light in cm/s
  It also has "specific intensity" as equivalent to "spectral radiance". Noone has suggested that we use "specific intensity" and I am not aware that it is used in our community. Therefore, I suggest we stay with "spectral radiance" and define it in the 3 ways described above. We could put "spectral intensity (=specific intensity)" if you wish, but I don't think this adds anything helpful for our community.

[gaochen-larc]

I agree with @JREyre. We should consider adding W.m-2.sr-1.(cm-1)-1 as it is commonly used in research and literature. If we don't someone may request later anyways. I also checked with a Polarimeter expert about the use of "spectral radiance" and "specific intensity". He prefers "spectral radiance. We can add a phrase after the "spectral radiance" in the definition: (also known as specific intensity). So it would read like: "SI derived unit of spectral radiance (also known as specific intensity) per..." Will this work?

[amilan17]

Decision:
All agree that the summary in the issue is ready

[amilan17]

@JohnEyre @gaochen-larc can you find a reference comparable to the wikipedia article that's not wikipedia?