search

cf-convention / vocabularies

Issues and source files for CF controlled vocabularies
3 stars 1 forks source link

use of "specific" in standard names #30

Open taylor13 opened 2 years ago

taylor13 commented 2 years ago

Does anyone know why the word "specific" was included in the standard names land_ice_specific_mass_flux_due_to_calving and land_ice_specific_mass_flux_due_to_calving_and_ice_front_melting (but not, for example, "land_ice_runoff_flux")?

JonathanGregory commented 2 years ago

The word "specific" in terms relating to land ice mass balance means per unit area. It was included because "surface mass balance" is often given as an area-integral quantity e.g. in Gt yr-1 for an ice sheet. By contrast, "specific surface mass balance" means in kg m-2 yr-1. Hence for clarity specific in included land_ice_surface_specific_mass_balance_flux and ..._rate. For consistency it was therefore also included in land_ice_specific_mass_flux_due_to_calving[_and_ice_front_melting] but I agree that makes them inconsistent with land_ice_surface_melt_flux and land_ice_runoff_flux, which are also in kg m-2 s-1. I think it would be OK to delete specific in land_ice_specific_mass_flux[_due_to_calving[_and_ice_front_melting] because these aren't customary terms anyway.

taylor13 commented 2 years ago

I understand now why "specific" has been used.

It might be worth revisiting the various terms related to gain and loss of land_ice. For example, consider

tendency_of_land_ice_mass_due_to_calving (kg s-1)  land_ice_calving_rate (m s-1) land_ice_lwe_calving_rate (m s-1) land_ice_specific_mass_flux_due_to_calving (kg m-2 s-1)

The first name, I think is clear. The last refers to a flux (which throughout the standard_name list is expressed as per unit area, so "specific" is unnecessary), or it might be replaced by the easy-to-understand tendency_of_land_ice_mass_content_due_to_calving (where in CF , "mass_content" is consistently used to express a mass per unit area). The other 2 names are confusing to me. I supposeland_ice_calving_rate, which has units of m s-1, is a measure of the rate that the land ice thickness is changing due to calving, but calving occurs at the boundary of the ice, so would seem to change its breadth, rather than its depth. How is this variable calculated? I also suppose for the "lwe" varible that a liquid-water-equivalent (depth) is calculated as the volume of water in the ice that has calved spread uniformly across the grid cell (assuming water at some nominal density). Is this correct? This seems unnecessarily complicated, duplicative, and imprecise. Why not in place of either/both of these just report the tendency_of_land_ice_mass_due_to_calving or the tendency_of_land_ice_mass_content_due_to_calving, which are easy to understand and make no assumptions about density of water or ice. Thus, the number of options would be reduced from 4 to the following 2:

tendency_of_land_ice_mass_due_to_calving (kg s-1), which already exists, and tendency_of_land_ice_mass_content_due_to_calving (kg m-2 s-1)

I would also prefer the following changes:

land_ice_basal_specific_mass_balance_flux (kg m-2 s-1) >> tendency_of_land_ice_mass_content_due_to_basal_loss

tendency_of_land_ice_mass_due_to_basal_mass_balance (kg s-1) >> tendency_of_land_ice_mass _due_to_basal_loss

land_ice_surface_specific_mass_balance_flux (kg m-2 s-1) >>  tendency_of_land_ice_mass_content_due_to_surface_gain

tendency_of_land_ice_mass_due_to_surface_mass_balance (kg s-1) >> tendency_of_land_ice_mass_due_to_surface_gain

and I would deprecate: land_ice_lwe_surface_specific_mass_balance_rate (m s-1) and land_ice_surface_specific_mass_balance_rate(m s-1)

JonathanGregory commented 2 years ago

Dear Karl

I agree we could rename land_ice_specific_mass_flux_due_to_calving as land_ice_mass_flux_due_to_calving, without specific, since that would make it consistent with other standard names. This is not the name it usually has in the literature anyway, so it probably wouldn't cause any problems. Instead of land_ice_surface|basal_specific_mass_balance_flux we could have land_ice_surface|basal_mass_balance_flux_per_unit_area. We've used per_unit_area in other CF standard names. The terms surface and basal mass balance mean the net addition of mass.

The standard names you mention have been agreed and in use for a number of years. They result from a compromise between customary terminology and consistency with CF, as is often the case, and for various reasons I don't think they'd be improved by changing the words. Also, CF doesn't prescribe what quantities people should store (principle 8 in section 1.2). CMIP does that, of course, but CF has to describe whatever data people commonly want to exchange and archive. Yes, you are correct that when a local tendency of mass change is expressed in canonical units of m s-1 (in practice mm yr-1 or m yr-1) a density has been assumed. In ice-sheet models the density is known. In reality the observation is often thickness, from altimetry, rather than mass.

Best wishes

Jonathan

github-actions[bot] commented 1 year ago

This issue has had no activity in the last 30 days. This is a reminder to please comment on standard name requests to assist with agreement and acceptance. Standard name moderators are also reminded to review @feggleton @japamment

taylor13 commented 1 year ago

With Jonathan's careful review, I defer to his superior expertise in this area; Jonathan, in your post above (https://github.com/cf-convention/vocabularies/issues/30) you say we could change some of the standard names involving "specific", but then you say we shouldn't. Did I get that right?

In any case, I do think for two of the existing names, land_ice_calving_rate (m s-1) land_ice_lwe_calving_rate (m s-1), we should amend the descriptions to read (if I understand it correctly):

land_ice_calving_rate: "Land ice" means glaciers, ice-caps and ice-sheets resting on bedrock and also includes ice-shelves. The land ice calving rate is the rate at which ice is lost through calving into the ocean. The measure of this rate in a region containing land ice is the mass loss per unit time divided by the product of the land ice area in the region considered and some appropriate density for the ice.

Similar wording should apply to land_ice_lwe_calving_rate: Either: "lwe" means liquid water equivalent. "Land ice" means glaciers, ice-caps and ice-sheets resting on bedrock and also includes ice-shelves. The land ice calving rate is the rate at which ice is lost through calving into the ocean. The measure of this rate in a region containing land ice is the mass loss per unit time divided by the product of the land ice area in the region considered and some appropriate density for liquid water.

The last sentence in each of these descriptions is new and should be checked for correctness.

JonathanGregory commented 1 year ago

Dear Karl @taylor13

Sorry to be confusing. I think we could rename the following three names for quantities in kg m-2 s-1:

i.e. delete specific. We don't need "per unit area" because that's what flux always means in CF standard names.

I agree with clarifying the definitions you mention. Those quantifies are treated as intensive in area in models, like snowfall and surface mass balance, for instance. Hence I would phrase the last sentence as "The measure of this rate is the mass loss per unit area per unit time divided by the density of the ice."

Best wishes

Jonathan

taylor13 commented 1 year ago

Yes, thanks for the good summary. We're in agreement, but I suggest we be more explicit regarding "per unit area". I think it is per unit area of land_ice, but perhaps someone might take it to mean per unit area of the grid cell containing the ice.

In the case of "per unit area of land_ice", how does one calculate it? Is it (mass2-mass1)/(dtlandIceArea1) or (mass2-mass1)/(dtlandIceArea2) or (mass2-mass1)/(dt*(landIceArea1+landIceArea2)/2) or what? where dt = the time-interval = time2-time1. It matters quite a lot when during the time step the only remaining land_ice in a grid cell completely calves into the ocean. If we're not careful, the ice-sheet calving rate could become infinite (undefined).

If the area is the grid-cell area, there would be no problem because cell area does not depend on time and possibly vanish.

Perhaps considerations of this kind would encourage reporting "tendency_of_land_ice_mass_due_to_calving", which has no area in the denominator.

JonathanGregory commented 1 year ago

Dear Karl @taylor13

Since it's an intensive quantity, it "really" means the rate at a point, although it is typically meaned over an area. That would be a land-ice area, I agree, but implicitly such quantities only exist over land ice anyway. This is similar to sea ice thickness, for example, which is an intensive quantity that exists at a point, but only over sea ice. It can reported as a mean over the grid box or a mean where sea ice.

The calving flux is a bit unusual, I think, because that isn't really at a point. In reality it is across a line, and in models it is within a grid cell. But we calculate land-ice mass balance assuming all these quantities to be collocated.

Best wishes

Jonathan

taylor13 commented 1 year ago

I agree that "flux" isn't quite the right term to use for calving loss. I think we should have constructed names using "tendency" (as in 100's of other standard name). So, for example, instead of land_ice_mass_flux_due_to_calving, we should have defined tendency_of_land_ice_mass_content_due_to_calving (or for the opposite sign, minus_tendency_of_land_ice_mass_content_due_to_calving). Guess we could still consider that.

Regarding the area issue, I think we have never made it clear how to handle variables that only exist over a certain portion of a grid cell. For example, sea_ice_temperature is defined only where there is sea ice, so when in cell_methods you find the phrase "area: mean", you probably should only average the temperature over the sea-ice portion of the grid cell (even if the cell_methods doesn't include "where sea_ice"). So in this case with or without "where sea_ice" you get the same area-mean. For a variable like "sea_ice_depth", on the other hand, it might reasonably be considered as 0 where there is no sea ice, so then the area-mean would be different depending on whether or not cell method includes "where sea_ice".

For land_ice_mass_flux_due_to_calving, should we assume that the flux is zero except where land_ice exists, or should we assume that it is undefined, in which case we report the same value whether or not "where land_ice" is included in cell_methods?

JonathanGregory commented 1 year ago

Dear Karl @taylor13

We could rename land_ice_specific_mass_flux_due_to_calving (kg m-2 s-1) as tendency_of_land_ice_amount_due_to_calving, along the lines you suggest, where amount means mass per unit area, as in many standard names. Similarly for ...calving_and_ice_front_melting. I would support doing that. We could ask ISMIP6 for an opinion.

I agree that we have not addressed the ambiguity of means for quantities which exist only for certain area types. Perhaps that should be a separate issue.

I also think we should rename tendency_of_change_in_land_ice_amount (kg m-2 s-1) as tendency_of_land_ice_amount, which I believe is what it means, given the units i.e. it's not an acceleration.

Best wishes

Jonathan

taylor13 commented 1 year ago

Anyone from the ISMIP community want to weigh in?

I agree with Jonathan that both calving variables should be treated similarly.

I also agree that "change_in" should be removed from the tendency_of_change_in_land_ice_amount because the current name is inconsistent with the specified units.

github-actions[bot] commented 1 year ago

This issue has had no activity in the last 30 days. This is a reminder to please comment on standard name requests to assist with agreement and acceptance. Standard name moderators are also reminded to review @feggleton @japamment

efisher008 commented 2 months ago

Hi Karl @taylor13 and Jonathan @JonathanGregory,

I'm revisiting this issue as it has been open for over a year with no conclusion reached. Did you have any success with contacting the ISMIP community about the changes to the names you proposed?

Best wishes, Ellie

taylor13 commented 2 months ago

I didn't try, and obviously no one responded by commenting on this issue. On the suggestions that Jonathan and I have expressed agreement, perhaps action can now be taken. My quick read of the above says a few issues could probably benefit from more thought/input. Do you (@efisher008) want to talk a crack at summarizing, or is the input too muddled for that?

efisher008 commented 2 months ago

Hi Karl @taylor13,

I will have to look through the issue carefully but will try my best to summarise. Please bear with me while I do this!

Best, Ellie

JonathanGregory commented 2 months ago

I've consulted Sophie Nowicki, who chairs the ice-sheet model intercomparison project (ISMIP). She and I together proposed several of the existing ice-sheet standard names, for use in CMIP6. We agree that these changes would be fine, to remove "specific", which means "per unit area" in some glaciological terms, but that's not self-explanatory:

land_ice_specific_mass_flux_due_to_calving to tendency_of_land_ice_amount_due_to_calving

land_ice_specific_mass_flux_due_to_calving_and_ice_front_melting to tendency_of_land_ice_amount_due_to_calving_and_ice_front_melting

land_ice_surface_specific_mass_balance_flux to land_ice_surface_mass_balance_flux

land_ice_lwe_surface_specific_mass_balance_rate to land_ice_lwe_surface_mass_balance_rate

land_ice_basal_specific_mass_balance_flux to land_ice_basal_mass_balance_flux

There are some other changes that Karl suggested to use the syntax tendency_of_land_ice..._due_to... I agree with those.

taylor13 commented 2 months ago

Thanks for consulting with Sophie. I agree with the changes you suggest immediately above.

efisher008 commented 2 months ago

Hi @JonathanGregory,

I have a question about the use of amount in the names suggested above. Is this defined according to the CF standard name construction guidelines (i.e. units of kg m-2)?

Best, Ellie

JonathanGregory commented 2 months ago

Yes, land_ice_amount is intended to mean kg m-2. This may be misleading in the context of land ice, though. Perhaps tendency_of_land_ice_mass_per_unit_area_due_to_calving would be better.

efisher008 commented 2 months ago

Hi @JonathanGregory @taylor13,

I think given that amount has a defined unit in CF, it should be sufficient to call these names tendency_of_land_ice_amount_due_to_calving etc. I have made these changes in the editor and the names are currently listed as "under discussion".

I will address Karl's earlier comments in a separate post.

Best, Ellie

taylor13 commented 2 months ago

maybe "amount" is o.k., but is "mass_balance" well understood? It reminds me of a mechanical scale used to weigh objects by matching their mass with a known masses. What is proposed for defining what is meant by "mass_balance", or is it so obvious to everyone but me that no further explanation is needed. (Don't be afraid to say so, if it is.)

JonathanGregory commented 2 months ago

Dear Karl

"Mass balance" is well-understood in glaciology, but that may not be sufficient for standard names. It's in the UNESCO glaciological glossary and the IPCC WG1 AR6 glossary, for instance. It means the difference between accumulation (addition of mass) and ablation (loss of mass). "Mass budget" is a synonym, less commonly used in the literature. Do you think "mass budget" would be preferable for standard names?

taylor13 commented 2 months ago

Perhaps, then, "mass balance" is o.k., but it may not be obvious what the sign convention is, so could the description include something like "Mass balance is a measure of the change in ice mass over time and is taken to be positive when the addition of mass (accumulation) exceeds loss (ablation)."

JonathanGregory commented 2 months ago

Dear Karl @taylor13, Ellie @efisher008

The standard names we're discussing are renamings, rather than new entries. The existing descriptions of the two standard names containing specific_mass_balance_flux, either surface or basal, contain the text

"Specific mass balance" means the net rate at which ice is added per unit area. A negative value means loss of ice.

For the renamed quantities (deleting specific), I think this should be changed to

"Mass balance flux" means etc. unchanged

The descriptions of the two standard names containing surface_specific_mass_balance_rate current includes

Specific mass balance means the net rate at which ice is added per unit area at the land ice surface due to all processes of surface accumulation and ablation. A negative value means loss of ice.

which should be changed to

"Surface mass balance" means etc. unchanged

The descriptions of the two tendencies are OK as they stand.

Ellie, I can't view the entries with the links you gave. They give

You need to be logged in to see this page. Use the link below to login.
Login

and if I click Login, I get

403 Forbidden
nginx

Best wishes

Jonathan

efisher008 commented 2 months ago

Hi Jonathan @JonathanGregory,

That is interesting - I've been including these links as standard practice in GitHub issues for a while, and nobody has said before that they cannot actually access the names!

It seems the 'edit' suffix in the link URL requires an admin login. Please try using the links without this last 'edit', i.e. https://cfeditor.ceda.ac.uk/proposal/5421. This should bring you into 'view' mode. I tested this by logging out of the admin interface and then trying the new link, so I think it should work.

Best wishes, Ellie

efisher008 commented 2 months ago

I also noticed that there was a mass_balance_rate name missing from this comment which @JonathanGregory refers later: land_ice_surface_mass_balance_rate.

I have added it here: https://cfeditor.ceda.ac.uk/proposal/5426

PS. I have also changed (shortened) the CF editor links so they should take you to view the entry, not edit it.

JonathanGregory commented 2 months ago

Thanks for adjusting the links. These don't work either for me, unfortunately. They give

Not Found
The requested resource was not found on this server.

Edit: The ones you've adjusted in your earlier post work OK.

taylor13 commented 2 months ago

The modified to the descriptions suggested by Jonathan in https://github.com/cf-convention/vocabularies/issues/30#issuecomment-2298547175 are now clear to me. Thanks!

github-actions[bot] commented 1 month ago

This issue has had no activity in the last 30 days. Accordingly:

Standard name moderators are also reminded to review @feggleton @japamment @efisher008