Updating the oxygen and hydrogen isotope ratios in the PMIP part of Emon

[jypeter]

This thread will be used to update the oxygen and hydrogen isotope ratios in the PMIP part of Emon, aka Roche variables because they were requested by Didier Roche (@dmr-dj) and the content of the description column was Roche - LSCE up to DR 01.00.22

This will address the missing units problem mentioned in #275

Note that there are 10 Roche variables in DR 01.00.22, but 2 variables are missing: we need a total of 12 variables in order to cover { 18O, 2H, 17O } x { precip, solid precip, water vapor, sea water }
=> 3 x 4 variables

The list of updated variables is available in CMIP6_MIP_tables_Roche-O-H-Isotope.xlsx and I will try to find a clean way to list the required changes in the next part of this issue

[jypeter]

The table below gives alternating lines of current variable definitions in DR 01.00.22 and the new requested definitions, and the first column gives either the row number of the variable in the DR 01.00.22 Emon variables, or updated to show the new values, or NEW for the 2 new variables

The following columns are not changed (or were replicated, in the case of the 2 new variables) and were not included in the summary table below: priority, cell_methods, positive, type, dimensions, modeling_realm, frequency, cell_measures

The columns from prov to MIPs (by experiment) that are more internal DR stuff were not included either

The xlsx version of the table below is available in CMIP6_MIP_tables_DR22vsRoche_subset.xlsx (and a version without the extra columns removed is available in CMIP6_MIP_tables_DR22vsRoche.xlsx)

Line in DR22 Emon	Long name	units	description	comment	Variable Name	CF Standard Name	CMOR Name
280	18O in total precipitation		Roche - LSCE		O18p	missing	O18p
updated	H218O in precipitation	kg m-2 s-1	Ratio of 18O in precipitation, includes both liquid and solid phases	at surface; includes both liquid and solid phases from all types of clouds (both large-scale and convective)	pr18O	precipitation_flux_H218O	pr18O
281	18O in solid precipitation		Roche - LSCE		O18s	missing	O18s
updated	H218O in snowfall flux	kg m-2 s-1	Ratio of 18O in solid precipitation, includes solid phases only	at surface; includes precipitation of all forms of water in the solid phase	prsn18O	snowfall_flux_H218O	prsn18O
282	2H in total precipitation		Roche - LSCE		H2p	missing	H2p
updated	1H2HO in precipitation	kg m-2 s-1	Ratio of 2H in precipitation, includes both liquid and solid phases	at surface; includes both liquid and solid phases from all types of clouds (both large-scale and convective)	pr2h	precipitation_flux_1H2HO	pr2h
283	2H in solid precipitation		Roche - LSCE		H2s	missing	H2s
updated	1H2HO in snowfall flux	kg m-2 s-1	Ratio of 2H in solid precipitation, includes solid phases only	at surface; includes precipitation of all forms of water in the solid phase	prsn2h	snowfall_flux_1H2HO	prsn2h
284	17O in total precipitation		Roche - LSCE		O17p	missing	O17p
updated	H217O in precipitation	kg m-2 s-1	Ratio of 17O in precipitation, includes both liquid and solid phases	at surface; includes both liquid and solid phases from all types of clouds (both large-scale and convective)	pr17O	precipitation_flux_H217O	pr17O
285	17O in solid precipitation		Roche - LSCE		O17s	missing	O17s
updated	H217O in snowfall flux	kg m-2 s-1	Ratio of 17O in solid precipitation, includes solid phases only	at surface; includes precipitation of all forms of water in the solid phase	prsn17O	snowfall_flux_H217O	prsn17O
286	O18 in water vapor		Roche - LSCE		O18wv	missing	O18wv
updated	H2O18 in water vapor	kg m-2	Ratio of 18O in atmospheric water vapor	ideally, provided on all vertical levels	prw18O	atmosphere_water_vapor_content_H218O	prw18O
287	O17 in water vapor		Roche - LSCE		O17wv	missing	O17wv
updated	H2O17 in water vapor	kg m-2	Ratio of 17O in atmospheric water vapor	ideally, provided on all vertical levels	prw17O	atmosphere_water_vapor_content_H217O	prw17O
288	H2 in water vapor		Roche - LSCE		H2wv	missing	H2wv
updated	1H2HO in water vapor	kg m-2	Ratio of 2H in atmospheric water vapor	ideally, provided on all vertical levels	prw2H	atmosphere_water_vapor_content_1H2HO	prw2H
318	O18 in sea water		Roche - LSCE		O18sw	missing	O18sw
updated	H2O18 in sea water	1	Ratio of 18O in sea water		sw18O	sea_water_H218O	sw18O
NEW	H2O17 in sea water	1	Ratio of 17O in sea water		sw17O	sea_water_H217O	sw17O
NEW	1H2HO in sea water	1	Ratio of 2H in sea water		sw2H	sea_water_1H2HO	sw2H

[taylor13]

Do the units imply a mass flux of water molecules that contain the given isotope (e.g., mass of H2O containing an O18 atom) or is it the mass flux of the atom itself (e.g., mass of O18 contained in precipitation)?

If these really are ratios, I presume it is the ratio of water containing a given isotope to all the water (e.g., ratio of mass of precipitation containing O18 to total precipitation).

The descriptions and possibly the units need revision to make it clear.

In addition the standard names don't follow current CF patterns, which include, for example:
litter_mass_content_of_13C
and
mass_content_of_13C_in_vegetation_and_litter_and_soil_and_forestry_and_agricultural_products

Finally, the "Name in Data Request", and the "Variable Name" should almost always be the same.

thanks.

[dmr-dj]

I presume from the above discussion that the name is wrong, not the unit itself.
Indeed, what I had in mind and what is done in most models (to my knowledge) is to output the water mass flux (hence kg.m-2.s-1) times the ratio.

In short, the ratio is generally computed as (using precipitation as an example):
18R_prc = pr18O / pr

Hence the unit of pr18O should be a mass flux. I thus conclude that the correct name should be "a mass flux of water molecules that contain the given isotope".

For the ocean, it is a ratio of the given isotope in seawater, since there is nothing like the mass of water in a given cell and hence the unit is correctly "None" or "1".

From the above, I would keep the short name and unit but adjust the long name I think.

Best wishes,
Didier

[martinjuckes]

Thanks @dmr-dj , that makes the intentions clear.

For pr18O we could use "Precipitation mass flux of water molecules that contain the oxygen-18 isotope (180), including solid and liquid phases. Sometimes known as heavy-oxygen water."

I'd like to change the long name, because the term "H218O" could be confusing to someone who doesn't know the context. Can we use the term "heavy-oxygen water", e.g. "Heavy-Oxygen Water Precipitation Flux"?

Similarly for "1H2HO", can we refer to "Semi-heavy water", e.g. "Semi-heavy Water Precipitation Flux"?

As Karl has mentioned, we also need to develop some new standard names for these variables, but I'd like to discuss that with Alison Pamment before discussing it further here.
regards,
Martin

[dmr-dj]

On the long name point, @martinjuckes :
I do not like the long name suggested. It would not appeal at all to the water isotope community I think. Potentially, we could discuss in the future double-substituted water isotopes such as HD18O, what would then be the name? Also, how do you place the H217O in this frame? If H218O is heavy and HDO is semi-heavy, then H217O is ... "almost heavy"? This is way too imprecise I think.

I am not attached to a particular form of the long name, but it should be clear that we are referring to water with a single heavier substitute atom, being 18O, 17O or 2H (=D).

Best wishes,
Didier

[martinjuckes]

OK, @dmr-dj , in other long names we have put chemical formulae in brackets for clarification, but I would like to have some words to convey the meaning as well. e.g.
(i) Precipitation Flux of Water containing Oxygen-18 (H2 18O)
or
(ii) Oxygen-18 Water Precipitation Flux (H2 18O)

Do either of these make sense?

[martinjuckes]

I've put some modified long names and descriptions into https://docs.google.com/spreadsheets/d/1vcHdcfhW423aqg8LCuMTBACB5-JUrVuxDV7s52XJm6I/edit#gid=0 .. based on our discussion above ... please check that they are OK. I've also added some draft standard names, which follow existing standard name patterns as far as possible. I'll discuss these with Alison next week and get them posted to the CF discussion list for approval.

One question for @taylor13 : the proposed revision of the variable names (O18p --> pr18O etc) would go against the WIP decision to fix all variable names, but these variables are highly specialised and unlikely to have been produced by the initial model runs completed so far -- so we could consider allowing an exception here (I've checked with @senesis , and he has confirmed that IPSL have not produced these variables yet). Should we stick to the old names or allow an exception? There are also two new variables which somehow got omitted from the original request ... I think we could add these (sw17O, sw2H -- the last 2 rows).

[jypeter]

The updated Description and CF Standard Name fields can be found in the PMIP Standard names: isotopic fluxes, mass contents and ratios proposal that @martinjuckes sent to the CF-metadata mailing list

@martinjuckes can we leave the Long name and Comment fields as they were proposed in #316 (comment) ?

[martinjuckes]

Hello @jypeter , I'd prefer to use the long names I've suggested in my comment on April 29th above [ https://github.com//issues/316#issuecomment-377205460 ]. The one that you have suggested in #316 (comment) are a bit too brief and technical for my taste.

[martinjuckes]

@dmr-dj : we need some clarification on sw2h : you have long name implying that it is the ration of 1H.2H.O molecules to 1H2.O molecules (i.e. ratio of number of molecules with a single deuterium to the number of molecules with two base state hydrogen atoms) and a description which suggests that it is the ratio of 2H atoms to 1H atoms in sea water: which of these interpretations is correct?

[martinjuckes]

Hello @dmr-dj , @jypeter : we need to finalise definition of variables for CMIP6: unless I hear otherwise I'll assume that s2wh is an isotope ratio (i.e. ratio of 2H atoms to 1H atoms) rather than a molecular composition ratio (i.e. ratio of 2H.H.O to H2O).