1-01-01 addition of new variable "relative attenuated backscatter"

[fstuerzl]

Issue Summary

Proposal

~ 9 June 2022

notation	path/tags	name	description
12251	particle phase, optical properties, atmosphere	Relative attenuated backscatter	The product of the total volume backscatter coefficient and the atmospheric two-way transmission of the layers between the instrument and the observed volume presented in arbitrary units. The total volume backscatter coefficient is the sum of all backscatter coefficients from the atmosphere (including from molecules and particles) accounting for all polarizations. The atmospheric two-way transmission accounts for the total volume (molecular and particulate and all polarizations) extinction along the path from the instrument to the observed volume and back. The value of this variable in presented in relative, arbitrary units to account for observations of attenuated backscatter lacking final absolute calibrations.

Reason

Addition of a missing variable.

Stakeholder(s)

@ejwelton (MPLNET)

Consultations

@ejwelton, @AlexanderHaefele, @markusfiebig, @JohnEyre  

Context

MPLNET refers to this as normalized relative backscatter (NRB, which is NASA terminology). This variable is also used by the NASA ICESat and ICESat-2 satellite missions, and the US Depart of Energy ARM program.

Campbell, J.R. J.R., D.L. Hlavka, E.J. Welton, C.J. Flynn, D.D. Turner, J.D. Spinhirne, V.S. Scott, and I.H. Hwang, 2002. Full-time, Eye-Safe Cloud and Aerosol Lidar Observation at Atmospheric Radiation Measurement Program Sites: Instrument and Data Processing, J. Atmos. Oceanic Technol., 19, 431-442.

Expected Impact of Change

MEDIUM

[ejwelton]

Per discussion in issue #363 I recommend the following description for relative attenuated backscatter:

The product of the total volume backscatter coefficient and the atmospheric two-way transmission of the layers between the instrument and the observed volume presented in arbitrary units. The total volume backscatter coefficient is the sum of all backscatter coefficients from the atmosphere (including from molecules and particles) accounting for all polarizations. The atmospheric two-way transmission accounts for the total volume (molecular and particulate and all polarizations) extinction along the path from the instrument to the observed volume and back. The value of this variable in presented in relative, arbitrary units to account for observations of attenuated backscatter lacking final absolute calibrations.

In addition, the same references from #363 are also applicable for this variable: Wandinger U. (2005) Introduction to Lidar. In: Weitkamp C. (eds) Lidar. Springer Series in Optical Sciences, vol 102. Springer, New York, NY. https://doi.org/10.1007/0-387-25101-4_1 Sassen K. (2005) Polarization in Lidar. In: Weitkamp C. (eds) Lidar. Springer Series in Optical Sciences, vol 102. Springer, New York, NY. https://doi.org/10.1007/0-387-25101-4_2

[joergklausen]

@AlexanderHaefele, John Sullivan (john.t.sullivan@nasa.gov), Amin Nehrir (amin.r.nehrir@nasa.gov), Ina Mattis (@inamattis), @gaochen-larc Please confirm or comment this proposal.

[AlexanderHaefele]

I support support the proposal as presented in https://github.com/wmo-im/wmds/issues/367#issuecomment-991240129.

[markusfiebig]

Sorry to come in late into this discussion. For other names, we included the entities involved in observing the named property, here "atmosphere" and "light". Should we do this here as well? This would make the name to "atmosphere light attenuated backscatter".

[joergklausen]

Sorry to come in late into this discussion. For other names, we included the entities involved in observing the named property, here "atmosphere" and "light". Should we do this here as well? This would make the name to "atmosphere light attenuated backscatter".

I don't know about including 'light' in the name. 'atmosphere' will be added as a tag. We won't include the 'domains' in the variable names any more if this can be expressed with the 'feature-of-interest' (aka, domain).

[joergklausen]

I would like to tag this a 'ready' for FT-2022-2. For that, I need to know your preference between 'Relative attenuated backscatter' and 'Relative particle light attenuated backscatter coefficient'. FYI: Particle light backscatter coefficient exists already, so my preference would be for the second option.

@markusfiebig @ejwelton @AlexanderHaefele @gaochen-larc Thanks for your feedback.

[ejwelton]

It should remain as relative attenuated backscatter. Attenuated backscatter is already in the fast track, and this is very closely related. These two terms are used for the lidar signals not retrieved properties like backscatter.

[AlexanderHaefele]

it has to remain as relative attenuated backscatter since it is the result of all sorts of scattering and not only particle as explained in the description in https://github.com/wmo-im/wmds/issues/367#issuecomment-991240129. I am also not in favor of the notion "light" since I do not see a need to limit it to a certain wavelength region.

[joergklausen]

@markusfiebig @ejwelton @AlexanderHaefele @gaochen-larc

I have created the branch and need 2 final confirmations that this is okay. Thanks!

[joergklausen]

@markusfiebig @ejwelton @AlexanderHaefele @gaochen-larc

I have created the branch (https://github.com/wmo-im/wmds/commit/62621e1f818126ad8c0134f025e1db20ee0619ee) and need 2 final confirmations that this is okay. Thanks!

[joergklausen]

Branch (https://github.com/wmo-im/wmds/commit/fbb0087d47cc1f96e7e10940a2d7c42551974796) updated/corrected after comment from @AlexanderHaefele.

[amilan17]

Is this the same or different from issue #363, which was recently introduced during FT22-1?

I updated a new branch which includes both terms/definitions. https://github.com/wmo-im/wmds/tree/issue367new

[amilan17]

I commented too quick, I see in the comment history that this is not a dupe of #363.

[amilan17]

@fstuerzl or @joergklausen can you please update the issue summary to include the individuals that were consulted?

[JohnEyre]

@joergklausen @amilan17 . This proposal raises a more general issue, which I highlighted in my initial review of WMDR (Nov 2021). A class of variables appears to be missing from WMDR. This is the class of Level I observed variables, such as radiances, brightness temperatures, reflectivities, backscatter coefficients, bending angles, etc. Observations of these variables are used directly in NWP and other applications. As a consequence, WMO messages containing such data are widely disseminated. (In fact, in terms of data volumes, they dominate the observational data on the GTS). They do not appear in OSCAR/Requirements because requirements are stated in terms of Level II geophysical variables. In OSCAR/Space, the capabilities are stated both in these Level I terms and in Level II terms for comparison with OSCAR/Requirements.”
Although the variables do not appear in /wmdr, similar variables appear (to some extent) in other parts of the WMO codes database – in /grib2, /bufr4, etc. So, my comment of this proposal is: if we accept "relative attenuated backscatter" in the "observed variables" list in /wmdr, should we also add the other variables listed above? Alternatively, one could argue that "observed variables" in /wmdr should be restricted to geophysical variables and that the "engineering" (Level 1) variables belong elsewhere in /wmdr.

[markusfiebig]

Apologies for not speaking up during today's telecon. I came across another fundamental issue that touches again the question of including "atmosphere" and "light".

We need to ask ourselves what these code lists are supposed to be, a glossary or a vocabulary. In a glossary, you can use all kinds of terms in different meanings, as long as the definition explains what the term means. In a glossary, it is good practice to always use a term in the same definition, and to have the same grammar for terms, but you get away with not following good practice. This is different with a vocabulary. In a vocabulary, you need to:

• always use the same grammar. For example, in "aerosol particle light scattering coefficient", you specify all entities involved in the observation ("aerosol particle", "light") and the observed property defining their interaction (scattering coefficient). If that is the grammar, then all concepts should use that grammar.
• In a vocabulary, a term is used always in the same definition in all concepts.
• The target group of a vocabulary includes the scientifically educated non-expert, not only the core experts. Core expert groups have a tendency to define their own "tribal language", which makes their own communication very efficient, but difficult to understand even for educated people just slightly outside the field. For example, "liquid water content" is readily understood by a cloud expert, but a scientifically educated non-expert will understand "liquid droplet mass concentration" better.

In this light, I would still argue for using "atmosphere light relative attenuated backscatter", since it makes explicit which entities are involved. With "relative attenuated backscatter", you need to be a core expert to know that it implicitly refers to "atmosphere" and "light". This should also take care of the comment by @AlexanderHaefele .

[joergklausen]

@JohnEyre I agree these Level I observed variables should be added. Can you please open a new issue for this, ideally you would already include a table with name, definition, and tags.

@markusfiebig I also agree with your comment, in principle. I am not sure I fully agree with your distinction between glossary and vocabulary, but the purpose of the code lists is clearly to provide a vocabulary of well defined terms for a well defined concept. The concept we want to evolve a vocabulary for is the 'variable'. I-ADOPT has provided a fairly consistent approach and I believe, we should converge towards that also for WMD vocabularies. We're on our way, not quite there yet, and still have some clean-up to do. Now, wrt the example above, I tend to disagree with your reasoning. The term 'atmosphere light relative attenuated backscatter' is indeed what we would call 'observed variable', but in our (and the I-ADOPT) information model, this can be atomized/normalized and/or tagged:

• variable = relative attenuated backscatter
• domain = atmosphere
• tags = Optical properties, atmosphere, particle phase, light (@AlexanderHaefele has argued why 'light' should not be part of the name, but I don't see a problem to include light as a tag)

[joergklausen]

@amilan17 I would like to declare this issue as 'sufficiently discussed', the branch is validated. Have you included it in the amendment?

[amilan17]

ready for FT