How to handle viscosity in standard names.

[ChrisBarker-NOAA]

Question

I'm working with a group on developing standard names for oil spill modeling.

There are lots of questions -- but the topic at hand is how to handle viscosity:

Viscosity is commonly represented as either Dynamic or Kinematic viscosity -- the difference is that one includes the density -- so if you know the density, you can convert.

So how to represent this with standard names? Options I see are:

1. Only have a standard name for one of these (usually kinematic)
2. Have one standard name, and let the units make it clear which on it is.
3. Have two standard names, e.g. kinematic_viscosity, dynamic_viscosity

I will say that in my work, I have to deal with both frequently for lab-derived oil property data -- different lab protocols measure it differently, and so the natural units are different, and so both get reported. But it's a pain, as then my code has to look for both, and convert one way or the other. So I'd rather have only one.

That being said, a standard name is about "if you have this thing, call it this name", and if you have a particular one, you should not have to convert to write it out -- and maybe even cannot do so.

With a quick search, I didn't find viscosity in the existing standard names.

[*] More info, if interested: https://www.rheosense.com/basics/viscosity-units

[JonathanGregory]

Dear @ChrisBarker-NOAA

I agree with, "if you have this thing, call it this name". That is, we should provide standard names for the quantities people want to use. That's also consistent with principle 8 in section 1.2, "Conventions are provided to allow data-producers to describe the data they wish to produce, rather than attempting to prescribe what data they should produce." Also, the two quantities have different units, so they can't share a standard name.

I can never remember which is kinematic and which dynamic viscosity - they don't strike me as self-explanatory terms. I've just looked it up. Dynamic viscosity is in Pa s. Kinematic viscosity is dynamic viscosity divided by density, giving it the unit m2 s-1, which is the unit of diffusivity. For any quantity X in units u, the flux of X is X multiplied by velocity, which has units u m s-1. If the flux is diffusive, it's proportional to the gradient of X, in units u m-1. The diffusivity is the ratio of the flux to the gradient.

Accordingly, Wikipedia says "kinematic viscosity" is sometimes called "momentum diffusivity", and CF has followed this with a few existing standard names e.g. atmosphere_momentum_diffusivity (canonical unit m2 s-1). Therefore I'd suggest we continue with momentum_diffusivity for kinematic viscosity, and define dynamic_viscosity standard names for which you have a use case.

Best wishes

Jonathan

[ChrisBarker-NOAA]

Side note:

atmosphere_momentum_diffusivity has no definition in the standard name table:

ocean_vertical_momentum_diffusivity has one that's a bit lacking

ocean_vertical_momentum_diffusivity_due to *

Has more -- so we can probably borrow text from the ocean ones.

But: "momentum diffusivity" usually refers to what is sometimes called "turbulent viscosity" or "eddy viscosity". As opposed to molecular viscosity.

As I learned it, molecular viscosity s the "real" viscosity, and "eddy viscosity" is called that because it has the same units, and fits into the equations in the same way, even though it's a distinct physical phenomenon.

So using "momentum_diffusivity" in the existing standard names was a good choice, rather than "eddy viscosity" -- as that's what it really IS.

But in this case, I'm trying to capture actual molecular viscosity, and "viscosity" is the correct term, and I think it's better that the names ARE different, even though they are the same units -- if you think about it, there must be many names that are different, as they capture a different phenomenon, even thought they use the same units.

And in practice, you would confuse the heck of oil spill practitioners if you called it "momentum_diffusivity".

But it seems you do think we should have separate names for dynamic and kinematic viscosity even though they are essentially the same physical phenomenon.

While I have at least one person's attention, another question coming up is how specific to be about the oil part of these names. For instance:

When we model oil spills, we keep track of the oil itself, and oil also tends to form emulsions with sea water.

So there are two different sets of physical properties, such as density and viscosity:

The properties of the oil itself.

The properties of the emulsion.

The question is -- how specific should we make the names?

density_of_emulsion
and
density_of_oil

vs
density_of mixture
and
density_of_oil. -- we probably need "oil" there, not just plain density ....

Hmm -- I should probably start a new question for these.....

[JonathanGregory]

Dynamic and kinematic viscosity are different quantities, with different units, so they cannot have the same standard name, although they're related concepts. This is like e.g. rainfall_flux and rainfall_rate. I see that "momentum diffusivity" has a lot of Google hits, not just Wikipedia, and I think it's preferable to "kinematic viscosity". We should say something like "This is often called kinematic viscosity" in the description of the standard name. That's an approach we've often used, so that people can find things in the table by searching including the descriptions.

I think that "emulsion" isn't sufficiently informative. That would not be enough to tell whether two quantities are comparable. For instance, the density of an oil-in-water emulsion isn't the same geophysical quantity as the density of milk, which is also an emulsion. 😃 I would suggest density_of_emulsion_of_oil_in_sea_water.

[ChrisBarker-NOAA]

I see that "momentum diffusivity" has a lot of Google hits, not just Wikipedia,

So does "kinematic viscosity" ?

and I think it's preferable to "kinematic viscosity".

They are actually two different things -- I have never seen a discussion of viscosity itself that mentions "momentum diffusivity" at all. Viscosity is:

"viscosity is the measure of a fluid's resistance to flow"

it is a bulk property of the fluid itself -- and that is how it's used, and what it means, in the context at hand.

"momentum diffusivity" is also what it describes, yes, but it's a term that (I think) was coined for - and certainly used for - turbulent flows, where momentum is diffused, much like viscosity, but it is a property of the flow, not of the fluid.

It is also often called "eddy viscosity" or "turbulent viscosity" and is absolutely the right term where it's currently used in CF.

That is: "kinematic viscosity" is "momentum diffusivity", but "momentum diffusiivty" is not (necessarily) kinematic viscosity.

Anyway, the important thing I thing here is that "kinematic viscosity" is absolutely the term of art in the field at hand -- people would be very confused if we didn't use it.

I'll be collecting a bunch of this all together -- we can argue the details later :-)

Dynamic and kinematic viscosity are different quantities, with different units, so they cannot have the same standard name, although they're related concepts. This is like e.g. rainfall_flux and rainfall_rate.

Ahh yes, that's a good analogy.

I think that "emulsion" isn't sufficiently informative.

well, I know we'll need to come up with longer names, yes, but I was wondering about how general to be: is this the density of the whole mixture, whatever that mixture may be -- or specifically this one?

e.g. The density of milk, which is also an emulsion

well, yes, if we say "emulsion" we need to be more specific -- though I'm not totally sure -- density is density -- the goal here is to capture particle tracking model results, and particle tracking models can model lots of things -- if it's the density of whatever the particle represents, then do we need to be more specific? But I suppose it would be more in keeping with the rest of the quite specific standard names...

density_of_emulsion_of_oil_in_sea_water

Well, it's

density_of_emulsion_of_sea_water_in_oil

but yes, that's appropriately clear and specific.

and then:

kinematic_viscosity_of_emulsion_of_sea_water_in_oil
and
dynamic_viscosity_of_emulsion_of_sea_water_in_oil

This does get to be a long list!

The other trick is that in practice, the properties are that of the "substance" the particle represents, that may or may not be an emulsion at any given time, but it always has a total density. If the water fraction is 0 -- is it an emulsion? well, no, but we want to store it in the same variable ....

I'll close this for now -- more later ....