I've noticed I sent this (on 15 Aug) to Michael only, not the CF list.
Dear Michael
Sorry for the delay. I was busy with another urgent task (IPCC report first
draft).
We agree on these:
> * partial_pressure_of_carbon_dioxide_in_air:Pa
> * partial_pressure_of_carbon_dioxide_in_sea_water:Pa
> * speed_of_sound_in_sea_water:m s-1
> * height_above_sea_floor:m
> * volume_absorption_coefficient_of_radiation_in_sea_water:m-1
> * volume_attenuation_coefficient_of_radiation_in_sea_water:m-1
> * volume_scattering_function_of_radiation_in_sea_water:m-1 sr-1
Following Roy's points (1) I presume that the absorption and attenuation are
integrals over all wavelengths, unless there is a coordinate variable
specifying range(s) of wavelengths (2) scattering function is scattering due
to all scatterers.
I'm putting the "platform" business in a separate email.
> If you accept (and interpret as I do) the NIST SI Unit rules and style
> conventions (http://physics.nist.gov/cuu/Units/checklist.html) #22, which
> mentions molality, then we also agree on the following:
> * molality_of_oxygen_in_sea_water:mol kg-1
> * molality_of_nitrate_in_sea_water:mol kg-1
> * molality_of_phosphate_in_sea_water:mol kg-1
> * molality_of_silicate_in_sea_water:mol kg-1
I don't disagree that these are acceptable and conventional terms, but
although we should depart from common usage only with care, I'd repeat that I
don't think they are ideal terms. I'm using only my own ignorance as a guide,
but I suspect that "molality" isn't a term that most climate modellers would
be sure about, for example, but yet they may well come across these quantities
in the output of their ocean models soon. People may have heard of "molarity";
the similarity between these terms is another source of confusion. It's pretty
difficult even to pronounce "molarity" and "molality" distinctly! Therefore I
would propose the more cumbersome but explicit "moles_of_oxygen_per_unit_
mass_in_sea_water" etc. instead of "molality". What do you think?
> You had proposed "Omnidirectional spherical irradiance" for 4-pi spherical
> or "scalar" irradiance. The ocean optics literature tends to use the word
> "total", rather than "omnidirectional" to indicate a 4-pi measurement.
> Incidentally, such a 4-pi measurement is often the sum of upwelling and
> downwelling 2-pi spherical irradiances, where the difference (downwelling
> minus upwelling) is called "net downward", and which is already used in the
> CF convention.
> So rather than:
> omnidirectional_photosynthetic_spherical_irradiance_in_sea_water:W m-2
> omnidirectional_spectral_spherical_irradiance_in_sea_water:W m-3
> I would like to propose:
> total_photosynthetic_spherical_irradiance_in_sea_water:W m-2
> total_spectral_spherical_irradiance_in_sea_water:W m-3
Again while not common usage, I still prefer my "omnidirectional". That is
because it says what it means, whereas "total" does not. You would have to
know what the other possibilities are to understand "total". In fact the same
issue arises with the plane irradiance quantities that you mention. In common
usage, "total downward" is sometimes contrasted with "net downward", or just
"downward" (understood to mean "net"). But this is also unclear, so we adopted
"downwelling" instead of "total downward", and we do not have any name which
says just "downward" radiation; it has to be "downwelling" or "net downward",
to leave minimal possibility for misinterpretation by people not experts in
the domain. The point of standard names is that they facilitate exchange of
information, and hence they have to be comprehensible by as wide a group of
potential users as possible.
> For my proposed backscattering variable, you had suggested using just
> "scattering" rather than "back-scattering", and wondered if "back" was
> important. I would have to that "back" is indeed important, since the
> majority of scattering measurements are either 1)backscatter 2)small-angle
> forward scattering, or 3)difference of attenuation and absorption, and while
> any of the three measurement techniques -can- be scaled to estimate total
> scattering, the calculations are generally not performed because they are
> very approximate, and the resulting total scattering values are not
> comparable across techniques. Hence, I'd like to stick with:
> * volume_backscattering_coefficient_of_radiation_in_sea_water:m-1
I think this indicates I don't know what "scattering" means to you, so I
couldn't write a proper definition of it. I assumed that "scattering" means
that part of attenuation which isn't due to absorption. The angle into which
the scattering occurs is not relevant to the attenuation. I assumed
"back-scattering" was scattering into particular range of angles (those which
are nearly opposite to the incident beam). Is that what you mean? If so, your
back-scattering coefficient is an integral of the scattering function over
that range of angles, which ought to be defined (as a coordinate, I imagine),
and is a small part of the scattering coefficient. I'd prefer to give it a
more generic term indicating that's what it is. Your practical distinction is
in terms of measurement techniques, but CF names describe the physical
quantities measured, so as to make them comparable for data from different
sources (models and observations, particularly).
On this subject, we need a convention for the angle in the scattering
function. I assume it is the opening angle between the incident beam and the
scattered beam, pi for backward scattering. Is that right?
> Also, you mentioned that you didn't understand the comments about being
> "relative to air or pure water". The issue is how 'zero' is defined, and is
> somewhat of a calibration issue -- "relative to air" is essentially
> "relative to nothing", so I prefer it.
Well, indeed. I assumed the coefficients were defined in absolute terms,
telling you how much attenuation there is. Zero means no attenuation. That
doesn't seem relative to anything, or it's relative to nothing if you like!
> sea_water_bioluminescent_photon_rate:s-1 m-3
How about
bioluminescent_photon_rate_in_sea_water:s-1 m-3
to use the usual "in_MEDIUM" convention.
> sea_water_dissolved_oxygen_saturation:1
How about
fractional_saturation_of_oxygen_in_sea_water:1
for similar reasons. Is "dissolved" important? It is assumed in other cases
such as the mol kg-1 names. I think it is probably obvious in the context of
"saturation".
> sea_water_dissolved_oxygen_volume_fraction:1
I tend to think "volume fraction" is puzzling in this term, since once it's
dissolved it doesn't really occupy a volume of the sea water. How about
volume_mixing_ratio_of_oxygen_in_sea_water:1
which (like mass mixing ratios) might more obviously refer to the volume it
had before mixing. There must be a convention for the measurement of that
original volume. Is it STP? If so we could insert "at_stp" after oxygen. We
have used that before.
> Regarding (surface_carbon_dioxide_partial_pressure_anomaly:1), 'anomaly'
> indicates the difference between air and water carbon dioxide partial
> pressure. I see that this is different than how anomaly is used for all the
> other CF variables, so how about:
> surface_carbon_dioxide_partial_pressure_difference:Pa
This is a headache. It's the difference between two quantities and we could
name it as such, but that would be very lengthy. I think it would be clear
if we had yours with "between_sea_water_and_air" on the end, but would even
that be too lengthy? This is a routinely recorded quantity, presumably?
Since you put "surface" here, do you also want it on the other two partial
pressure names?
> water_volume_transport:m3 s-1
I suppose that should be sea_water_volume_transport, given the application.
Flows through estuaries and through instruments are very different, and it
seems odd to give them the same name! The former is something that should
have a standard name, although it would often need geographical definition.
When you use the latter, how do you associate it with other quantities?
Thanks for your help. Best wishes
Jonathan
Received on Sun Sep 11 2005 - 01:42:52 BST
