Dear Martin, Karl and Jonathan,
As has been pointed out already, this conversation started because land_ice and ice_on_land seemed to offer too much potential for confusion between the area_types. I accept that the choice of terminology could have been better, but it is very clear from the definitions that the two are intended to represent different areas. Hence, we need something that is linguistically sufficiently different from "land_ice" to avoid mixing the two up.
I agree with Martin that we should keep a distinction between ice and snow in area_types. We already distinguish between them in standard names, and I would argue that the terms "ice" and "snow" should be used consistently in both vocabularies, otherwise it will become very complicated to use them in conjunction with one another, not to mention a headache to maintain! Furthermore, we already have two standard names that refer explicitly to the conversion of snow to sea ice, tendency_of_sea_ice_amount_due_to_conversion_of_snow_to_sea_ice and tendency_of_surface_snow_amount_due_to_conversion_of_snow_to_sea_ice, and we may eventually have similar names for snow to ice conversion on land. I worry that if we define ice and snow to be the "same thing" based solely on CMIP5 and CMIP6 usage of area_types, we will have difficulty trying to (re)create the distinction if it is needed at a later stage. It would be difficult to retrospectively unpick which ice terms meant only ice, as opposed to ice and snow, just as w
e're now looking back at some older terms like ice_free_land and wondering exactly what they mean in the context of more recent developments (more on this later). As far back as 2010 we were discussing what to do about sea_ice albedo:
> It probably should be total sea ice albedo, though in NCAR
> model bare ice albedo is changeable due to black carbon and
> multi-scattering in internal layers.
(This is an extract from an email from Siobhan O'Farrell relating to CMIP5 sea ice quantities. It wasn't sent to the mailing list, which is why I couldn't find it in the archives, but I think it lends further support to the case for differentiating between snow and ice in area_types).
Jonathan has suggested:
ice_and_snow_on_land -> land_surface_ice_and_snow;
ice_on_land -> land_surface_ice_except_snow;
and Martin has suggested
ice_and_snow_on_land -> land_surface_ice_and_snow;
ice_on_land -> land_surface_ice (omitting the "except snow").
I'm afraid I have some issues with both of these. I think land_surface_ice suffers from the original problem - it sounds too much like land_ice. On the other hand, land_surface_ice_except_snow sounds like we are treating snow as though it's a subtype of surface ice rather than something distinct. In both cases, I'm not keen on referring to "surface_ice" . In standard names, the "surface" in surface_snow is used to distinguish between lying snow and falling snow. We've never needed to make the same distinction with ice because we use specific terms such as "cloud ice", "hail", and "graupel" to refer to ice that isn't already on the surface. If we now start talking about "surface_ice" we will introduce a difference in terminology between area_types and standard names. I wonder also if we are adding "surface" simply because we are searching for something that "sounds" different to land_ice.
There is another approach which may be less elegant, but more explicit. In an earlier post I wrote:
> I think we can regard the three existing land ice area_types as nested:
> ice_sheets = Grounded ice sheets + Floating ice shelves;
> land_ice = ice_sheets + Glaciers + Ice caps;
> ice_on_land = land_ice + River ice + Lake ice + Other ice on land, e.g frozen flood water.
>
> In addition we have:
> ice_and_snow_on_land = snow overlying ice_on_land + snow overlying
> bare ground or vegetation
Evan Manning (offlist) queried the last definition. In fact I think it should say ice_and_snow_on_land = ice_on_land + snow overlying_ice_on_land + snow overlying bare ground or vegetation.
Based on the above, I suggest:
ice_and_snow_on_land ->land_ice_and_freshwater_ice_and_surface_snow
ice_on_land -> land_ice_and_freshwater_ice
(N.B.This implies that Freshwater_ice = River ice + Lake ice + Other ice on land, e.g frozen flood water).
I think this has the advantage of showing how the land based ice area_types are related, and at the same time it would be difficult to mistake them for one another. We have sometimes taken a similar approach with standard names, e.g. joining process names with more than one _and_, when we have needed to be very clear about what is included/excluded in a particular term.
In his most recent post, Martin wrote:
>Firstly, I think lake_and_sea_ice should be aliassed to fresh_water_and_sea_ice, or have a change of definition to exclude river ice).
Because of my suggestions above, we'd have to go with the option of clarifying the definition to exclude river_ice. I think it makes sense to regard lake_ice as part of freshwater ice and it's preferable to implying that 'fresh water' only means lakes.
Karl pointed out that with the introduction of the new area_types we need to provide a definition of ice_free_land (currently it doesn't have one). Karl wrote:
> I suggest we also provide some text guidance as to what ice_free_land includes. I think it should probably be the complement of land_ice, but one could argue that it should be the complement of ice_and_snow_on_land. In either case we need to make this
> clear in the CF area type listing. I would note that ice_free_land as a type has not been used in CMIP5 or CMIP6 and in addition it is not used in any of the current standard_names.
To take the last point first, ice_free_land was in fact proposed by Jonathan in 2011:
http://mailman.cgd.ucar.edu/pipermail/cf-metadata/003998.html. It was proposed in conjunction with standard names for modelling ice-sheet mass balance,
surface_snow_and_ice_melt_flux, surface_snow_and_ice_refreezing_flux and land_ice_surface_specific_mass_balance_flux. We didn't discuss exactly what ice_free_land should mean. That wasn't an omission, it's simply that we didn't have definitions for area types at that time - it's only as we have added to the list that it has become necessary to manage them in a more careful way. However, it's clear that the only other relevant entry in the area_types table at the time was land_ice, so Karl's first interpretation is correct: ice_free_land and land_ice are the complements of one another. I think the standard names proposed at the same time also suggest that only snow overlying ice was under consideration, not snow on ground or vegetation. I suggest that we add the following description of ice_free_land:
'The area_type ice_free_land is the complement of the area_type land_ice, i.e. it indicates areas where glaciers, ice-caps, grounded ice sheets resting on bedrock and floating ice-shelves do not occur.'
For further clarity I suggest we also add a sentence to the definition of land_ice:
' "Land ice" means glaciers, ice-caps, grounded ice sheets resting on bedrock and floating ice-shelves. The area_type land_ice is the complement of the area_type ice_free_land.'
Karl has also asked how we should interpret standard names that say 'land_ice', i.e., we know what *area* they cover, but do (at least some of them) also include the overlying snow? This question affects names such as tendency_of_change_in_land_ice_amount - does it include the mass of the snow plus ice, or strictly the ice? These questions haven't arisen before in this level of detail, not least because some of the names have been around for a long time. I agree that it would be useful to clarify the standard name definitions wherever possible.
Best wishes,
Alison
------
Alison Pamment Tel: +44 1235 778065
NCAS/Centre for Environmental Data Archival Email: alison.pamment at stfc.ac.uk
STFC Rutherford Appleton Laboratory
R25, 2.22
Harwell Oxford, Didcot, OX11 0QX, U.K.
-----Original Message-----
From: CF-metadata <cf-metadata-bounces at cgd.ucar.edu> On Behalf Of Martin Juckes - UKRI STFC
Sent: 22 October 2018 13:13
To: Jonathan Gregory <j.m.gregory at reading.ac.uk>; cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
Dear Jonathan,
Thanks ... that is very clear, but it looks as though I haven't expressed my view clearly enough yet: I don't agree that snow is a specific form of ice.
In the atmosphere everything is reasonably clear, and I do agree that, within the atmospheric context, snow is just a form of ice. However, things get more complicated when the snow hits the ground and piles up.
In glaciology people talk about the transitions from snow through n?v? and firn and finally into ice. This transition is also of importance in the study of sea-ice. I'm not an expert in this area, but it is clear that there is an extensive community of people who see "snow" and "ice" as distinct. In the glaciological context we could make things unambiguous by referring to a transition from firn to "glacial ice", but that would not work in the sea-ice context, where we have exactly the same transition. A term such as "solid_ice" might work (solidity appears to be the main criteria for the transition, e.g. expressed in terms of the connectivity of air bubbles:
http://www.iceandclimate.nbi.ku.dk/research/drill_analysing/cutting_and_analysing_ice_cores/analysing_gasses/firn_zone/ ).
Trying to understand this distinction takes us on a diversion away from your original question, but it may help us to understand why people are interested in areas of "snow" as distinct from "ice" (though we may not want to call it "snow" in the end).
The other problem is that piles of snow may contain some liquid water, and the behaviour of this component will be reflected, to a limited extent, in CMIP6 diagnostics. Accumulated layers of snow can be referred to as "snowpack", but I'm not sure that the term is appropriate for snow which might have fallen today and may be gone tomorrow. Perhaps something like "surface_snow_layer"? Snow flakes and hail stones in the atmosphere may also have some liquid water present, of course, but I'm not aware of anyone wanting to refer to that liquid mass fraction in the standard names -- at the surface, on the other hand, we do have a need to distinguish between frozen and liquid parts of the snow layer.
You ask why anyone would care about "all forms of solid water except snow" .. and the answer is, I think, that, as far as CMIP6 is concerned, people only care about the binary partition snow (i.e. surface snow layer) vs. ice (i.e. solid ice). As far as I can tell, it is primarily yourself that is interested in a comprehensive terminology that takes into account all possible forms of frozen water ... with the best possible motivation of course, as we would lie to have terminology which is robust and usable in the future.
Coming back to your suggestion of using "ice_except_snow" to refer to parts of the surface cryosphere which are not snow: I'm concerned that this does not deal with the issue of liquid water in the surface snow layer, which would not be a problem when we restrict attention to area types, but which would become a problem as soon as we look at extending the terminology into standard names. I'm also concerned that this approach is driven by a belief in a specific interpretation of "snow" and "ice" which is far from universal ... hence the long diversion above.
This conversation started because of the potential confusion between "land_ice" and "ice_on_land". Are we clear about the definition of land_ice? It has been in area type table for a long time, and carries the description 'Land ice means glaciers, ice-caps, grounded ice sheets resting on bedrock and floating ice-shelves'. According to AMS an ice-cap should be dome shaped and extensive, but they don't say how extensive. Wikipedia and NSIDC suggests that it should be less than 50,000km2, but neither of them give a lower size limit or refer to the dome shape. There is a suggestion in the AMS definition that it should be perennial. Part of the problem, it seems to me, is to understand what we want to exclude from "land_ice". Would a phrase such as "Ice masses formed on land through the transformation of snow into solid ice"?
We might them consider "fresh water ice" as a concept complementary to "sea ice", although we already have "lake_and_sea_ice" which, according to its definition, must include river ice (it is defined as "floating ice excluding ice-shelves and icebergs"). Thus, we appear to have 3 main categories of surface ice: "land ice" (formed through consolidation of snow), "sea ice" (formed through freezing of sea water) and "fresh water ice", or "lake and river ice" (formed through freezing of fresh water). Both sea ice and lake etc ice can also have some material added through the consolidation of snow .... but that is not the mechanism for the formation of the areal coverage.
I think it will be clearer overall if we don't treat surface snow as a sub-category of surface ice: it just conflicts too much with common usage and usage in the relevant scientific communities.
Firstly, I think lake_and_sea_ice should be aliassed to fresh_water_and_sea_ice, or have a change of definition to exclude river ice);
We could follow the same pattern for "ice_on_land", and replace it with "fresh_water_and_land_ice", but that approach does not extend well to the snow and ice case.
How about:
ice_and_snow_on_land -> land_surface_ice_and_snow (as you suggested);
ice_on_land -> land_surface_ice (omitting the "except snow");
regards,
Martin
________________________________
From: CF-metadata <cf-metadata-bounces at cgd.ucar.edu> on behalf of Jonathan Gregory <j.m.gregory at reading.ac.uk>
Sent: 19 October 2018 13:21
To: cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
Dear Martin
I agree that snow is a specific form of ice, and that we have and need many standard names which refer to it. I'm questioning only the need for the area type of ice_on_land, because I'm wondering in which circumstances one needs to consider the area occupied by all conceivable forms of frozen water with the sole exception of snow.
However, the reason for starting this thread was a concern that land_ice is confusingly similar to ice_on_land. I tried to combine these concerns by suggesting making ice_on_land an alias of ice_and_snow_on_land. If we need to keep them distinct, we could alias ice_on_land as ice_on_land_except_snow, to be clear what it means, and less easily confused with land_ice.
Perhaps even better, to be consistent with the many standard_names that use the phrase surface_snow, could we change both
ice_and_snow_on_land -> land_surface_ice_and_snow
ice_on_land -> land_surface_ice_except_snow There is one existing standard name which contains the phrase surface_ice, by analogy with surface_snow. However, its definition does not state whether it includes snow or not! Hence for consistency I think we ought to clarify this standard name as well: to tendency_of_atmosphere_mass_content_of_water_vapor_
due_to_sublimation_of_surface_ice we should append ether _and_snow, or _except_snow, according to what it is intended to include.
Best wishes
Jonathan
----- Forwarded message from Martin Juckes - UKRI STFC <martin.juckes at stfc.ac.uk> -----
> Date: Fri, 19 Oct 2018 09:53:59 +0000
> From: Martin Juckes - UKRI STFC <martin.juckes at stfc.ac.uk>
> To: "Taylor, Karl E." <taylor13 at llnl.gov>, "cf-metadata at cgd.ucar.edu"
> <cf-metadata at cgd.ucar.edu>, Jonathan Gregory
> <j.m.gregory at reading.ac.uk>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear Karl, Jonathan,
>
>
> I appreciate where you are coming from with the assertions that snow is a form of ice, but it ain't necessarily so, at least not in the current CF names list. Karl has made the point that there are multiple issues to consider when comparing usage in area types with usage in standard names, but it is surely important to be consistent at this level, and decide whether we want to deprecate all the names that treat ice and snow as distinct (some of them agreed quite recently).
>
>
> As far as CMIP6 is concerned, the land-ice modellers (which means modellers of ice sheets and shelves, as glaciers and ice caps don't yet figure in the active realms of CMIP models) do not appear to care much about the distinction between ice and snow ... once it has touched the ground it adds to the mass of the ice and that is all you need to know (apart from perhaps albedo modelling). For sea-ice, on the other hand, there will be diagnostics representing the transition from snow to ice, clearly implying that snow is "distinct" from "ice" in this context.
>
>
> In LS3MIP also has an understanding of "snow" which conflicts with the simple idea that "snow is ice". For this community, "snow" is a matrix of ice crystals, air and liquid water which lies on the ground.
>
>
> I appreciate that these issues are somewhat tangential to the specific discussion here, but adopting the rule that "snow is ice" does have consequences beyond the present discussion, and it would be a significant change from the status quo. This goes beyond the use of area types in standard names: the question is whether we are using the concepts of "snow" and "ice" consistently.
>
>
> It is clear, I think, that the distinction between solid ice and the matrix of ice, air an liquid commonly known as "snow", or perhaps "snow-pack", is important in a range of land and sea-ice surface modelling contexts -- including for some CMIP6 diagnostics. That doesn't change the fact that we appear to have some lack of clarity in current naming conventions.
>
>
> regards,
>
> Martin
>
>
>
>
>
>
> ________________________________
> From: CF-metadata <cf-metadata-bounces at cgd.ucar.edu> on behalf of
> Taylor, Karl E. <taylor13 at llnl.gov>
> Sent: 18 October 2018 19:27
> To: cf-metadata at cgd.ucar.edu
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear all,
>
> I tend to agree that ice_on_land will confuse many users, especially with another surface type being named "land_ice".
>
> I also think that incorporating "area type" into a standard_name can sometimes lead to further complications (and potential confusion). I'll restrict my initial discussion here to "area type" used as an allowed value for a variable with standard_name="area_type" or used in cell_methods.
>
> I think "ice_and_snow_on_land" is fairly descriptive and is easy to define in a few words: "all frozen surface water on land (plus ice shelves) but excluding water in frozen vegetation".
>
> I suppose "ice_on_land" could be defined as: "all frozen surface water on land (plus ice shelves) but excluding snow and water in frozen vegetation."
>
> Like Jonathan, I wonder why there is a need for ice_on_land, and it is sometimes difficult to distinguish between surface snow and other types of frozen water, so I would favor either eliminating it or making it a synonym for "ice_and_snow_on_land" (since snow is just a form of ice).
>
> I suggest we also provide some text guidance as to what ice_free_land includes. I think it should probably be the complement of land_ice, but one could argue that it should be the complement of ice_and_snow_on_land. In either case we need to make this clear in the CF area type listing. I would note that ice_free_land as a type has not been used in CMIP5 or CMIP6 and in addition it is not used in any of the current standard_names.
>
> Turning to standard_names that include an "area type" as part of the
> name (e.g., land_ice_thickness is a standard name that includes
> "land_ice", which is an area type). Our definition of land_ice
> doesn't mention whether it includes any snow that might lie on its
> surface; it definitely excludes snow regions that are free of land_ice
> and it excludes some other types of frozen surface water (e.g. frozen
> puddles on bare soil). Because we don't say whether land_ice includes
> the snow resting on it, the definition of
> tendency_of_change_in_land_ice_amount, for example, is unclear. In
> many models where budgets of the land_ice are of interest, one would
> naturally include "snow" in the tendency calculation, but there could
> be a model that has a budget for the snow layer amount and a separate
> budget for the more compact "solid" ice below it. For the purpose of
> standard_names, should we explicitly state that land_ice includes the
> snow on top of it? [This statement is unneeded if we limit
the use of land_ice to identifying a surface type in cell_methods or as a value allowed for variables with the standard_name=area_type (because in those cases it is irrelevant whether the "solid ice" is or is not covered by snow).]
>
> best regards,
> Karl
>
>
> On 10/18/18 10:18 AM, Martin Juckes - UKRI STFC wrote:
>
> Dear Alison, Jonathan,
>
>
> we do need to be careful about the distinction between snow and ice.
>
>
> The generally reliable AMS glossary is not much help here: *firn<http://glossary.ametsoc.org/wiki/Firn><http://glossary.ametsoc.org/wiki/Firn>" is part of the process of transformation from *snow<http://glossary.ametsoc.org/wiki/Snow><http://glossary.ametsoc.org/wiki/Snow>* to *land ice* , *land ice<http://glossary.ametsoc.org/wiki/Land_ice><http://glossary.ametsoc.org/wiki/Land_ice>* is a layer of *ice* formed on land and *ice<http://glossary.ametsoc.org/wiki/Ice><http://glossary.ametsoc.org/wiki/Ice>* includes snow, which leaves the definition of "firn" a bit wanting.
>
>
> For the glaciologists, the transformation snow --> firn --> ice appears to be well established, but there are other usages in which ice includes snow. In cloud physics it seems to be clear that snow flakes are ice crystals.
>
>
> regards,
>
> Martin
>
>
> ________________________________
> From: CF-metadata
> <cf-metadata-bounces at cgd.ucar.edu><mailto:cf-metadata-bounces at cgd.ucar
> .edu> on behalf of Jonathan Gregory
> <j.m.gregory at reading.ac.uk><mailto:j.m.gregory at reading.ac.uk>
> Sent: 18 October 2018 14:52
> To: Pamment, Alison (STFC,RAL,RALSP)
> Cc: CF-metadata
> (cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>)
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear Alison
>
> Yes, I realise I'm being a bit provocative and perhaps rash in trying
> to get rid of ice_on_land (because I think it's terribly confusing).
> Distinctions can be made for practical purposes (e.g. in models)
> between ice and snow, although in reality it's a continuum. I'm
> wondering who has a need for an area type including *all* kinds of
> frozen water on land (ice sheets, glaciers, firn/neve, rivers, lakes,
> ponds, frozen flood water, snowfall which has melted and refrozen as
> ice, hailstones, frost) *except* snow - and if so, how do they distinguish snow from the rest.
>
> Best wishes
>
> Jonathan
>
> On Thu, Oct 18, 2018 at 12:29:01PM +0000, Alison Pamment - UKRI STFC wrote:
>
>
> Date: Thu, 18 Oct 2018 12:29:01 +0000
> From: Alison Pamment - UKRI STFC
> <alison.pamment at stfc.ac.uk><mailto:alison.pamment at stfc.ac.uk>
> To: "CF-metadata
> (cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>)"
> <cf-metadata at cgd.ucar.edu><mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear Jonathan and Karl,
>
> I agree with Karl's suggestion to change "lying snow" to "surface snow" in the definitions of ice_on_land and ice_and_snow_on_land. That would be more consistent with standard names and their definitions.
>
> We seem to agree also that changes to area_type strings should be treated in the same way as changes to standard_names, i.e. using aliases. There is no formal description of the area_type table in the Conventions document. The XML schema would look very much like that of the standard name table, minus the units and the amip and grib tags. In fact, I think it has been agreed in principle to remove the AMIP and GRIB columns from the standard name table itself (CF trac #116). I will submit a GitHub issue to encompass trac #116 and some proper documentation of the area_type table. The CF-checker would need to cope with the possibility of aliases for area_types, so it probably needs to go in the conformance document too.
>
> Regarding Jonathan's assertion that "snow *is* ice", once again I am a little cautious. We certainly have standard names that *don't* regard them as being the same thing. For example, the two names I mentioned previously: change_over_time_in_amount_of_ice_and_snow_on_land and change_over_time_in_thermal_energy_content_of_ice_and_snow_on_land. I was under the impression that models treat ice and freshly fallen snow as different variables which sometimes co-exist in the same grid cell and sometimes don't. I haven't yet managed to track it down in the mailing list archives, but I also have a vague recollection of a discussion some years ago about the surface albedo being affected by factors such as the age of ice and snow, which would be important when considering an area_type. Perhaps I'm wrong about the last point, so I'd be interested to know what others think about the suggestion to turn ice_on_land into an alias of ice_and_snow_on_land.
>
> Best wishes,
> Alison
>
> ------
> Alison Pamment Tel: +44 1235 778065
> NCAS/Centre for Environmental Data Archival Email: alison.pamment at stfc.ac.uk<mailto:alison.pamment at stfc.ac.uk>
> STFC Rutherford Appleton Laboratory
> R25, 2.22
> Harwell Oxford, Didcot, OX11 0QX, U.K.
>
>
> -----Original Message-----
> From: CF-metadata
> <cf-metadata-bounces at cgd.ucar.edu><mailto:cf-metadata-bounces at cgd.ucar
> .edu> On Behalf Of Jonathan Gregory
> Sent: 18 October 2018 05:45
> To: cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear Alison
>
> I agree that an alias mechanism would be better than abolishing
> something we have introduced. Although it might sound inaccurate,
> "land ice" is a term that is used in the literature to mean ice sheets
> and glaciers, rather than all ice on land. It contrasts with sea ice,
> as has been remarked,
>
> I think that ice_on_land could be confused with land_ice. In addition, ice_on_land could be confusing because snow *is* ice; there isn't a clear distinction between snow and non-snow ice, and ice_and_snow_on_land could mean the same as ice_on_land. Therefore I suggest that we made ice_on_land into an alias of ice_and_snow_on_land.
>
> Best wishes
>
> Jonathan
>
> ----- Forwarded message from Alison Pamment - UKRI STFC
> <alison.pamment at stfc.ac.uk><mailto:alison.pamment at stfc.ac.uk> -----
>
>
>
> Date: Wed, 17 Oct 2018 15:05:06 +0000
> From: Alison Pamment - UKRI STFC
> <alison.pamment at stfc.ac.uk><mailto:alison.pamment at stfc.ac.uk>
> To: Karl Taylor <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov>, Martin Juckes - UKRI STFC
> <martin.juckes at stfc.ac.uk><mailto:martin.juckes at stfc.ac.uk>, "CF-metadata (cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>)"
> <cf-metadata at cgd.ucar.edu><mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Dear Karl et al.,
>
> Thank you all for the comments in this discussion, which I have been watching with interest.
>
> I think we can regard the three existing land ice area_types as nested:
> ice_sheets = Grounded ice sheets + Floating ice shelves;
> land_ice = ice_sheets + Glaciers + Ice caps;
> ice_on_land = land_ice + River ice + Lake ice + Other ice on land, e.g frozen flood water.
>
> In addition we have:
> ice_and_snow_on_land = snow overlying ice_on_land + snow overlying
> bare ground or vegetation
>
> As Martin says, ice_on_land and ice_and_snow_on_land were designed to work with LS3MIP standard names. They include all frozen terrestrial water and are therefore wider than the other two categories. I can't comment on whether or not they are currently being used in the CMIP6 archive, but certainly that was the intention. The reason was to enable the use of the surface_albedo standard name along with specifying an area_type, instead of introducing lots of separate albedo standard names for different surface types. This approach received support in the mailing list discussions of LS3MIP names. We also introduced some standard names: change_over_time_in_amount_of_ice_and_snow_on_land and change_over_time_in_amount_of_ice_and_snow_on_land. The definition of "ice_and_snow_on_land" in these names follows that of the area_type.
>
> Martin has supported Karl's suggestion to modify the description of ice_sheet. In addition, Martin and Jonathan have suggested adding Greenland and Antarctica as examples rather than part of the basic definition so that the area_type can also be used for paleoclimate models. That seems like a good approach, hence I suggest:
> 'An area type of "ice_sheet" indicates where ice sheets are present, for example, in the present climate this would refer to the Greenland and Antarctic ice sheets. It includes both the grounded portion of those ice sheets (i.e., the portion resting on bedrock either above or below sea level) and the portion that is floating as ice shelves. It excludes all other ice on land (in contrast to land_ice, which includes, for example, small mountain glaciers and in contrast to ice_on_land, which is comprehensively inclusive of all types of ice on land).'
>
> Karl has asked whether ice_on_land includes snow. I think it doesn't, because as already mentioned we have ice_and_snow_on_land as a separate area_type. Therefore, I support Karl's suggestion to modify the description of ice_on_land to make that point clear:
> 'The area type "ice_on_land" means ice in glaciers, ice caps, grounded ice sheets (grounded and floating shelves), river and lake ice, and any other ice on a land surface, such as frozen flood water (but excluding snow). This is distinct from the area type 'land ice' which has a narrower definition. The area_type ice_and_snow_on_land is defined similarly, but includes lying snow.'
>
> It would also make sense to add a corresponding cross-reference in the description of ice_and_snow_on_land:
> 'The area type "ice_and_snow_on_land" means ice in glaciers, ice caps, ice sheets (grounded and floating shelves), river and lake ice, any other ice on a land surface, such as frozen flood water, and snow lying on such ice or on the land surface. The area_type ice_on_land is defined similarly, but excludes lying snow.'
>
> I am cautious about Jonathan's suggestion to remove ice_on_land - it was introduced specifically to cope with CMIP6, so might it not be needed in due course? Also, I don't know that the Conventions have anything to say about simply removing an area_type once it's gone into the table. I have been managing the area_types vocabulary following a parallel procedure to standard names. It would be nice if we could think of a better term, so as to cause less confusion with land_ice. We could then turn ice_on_land into an alias, just as we would with a standard name.
>
> I agree with Martin that Evan will probably need to request some new area_types to work with his microwave data. Evan's suggestion of land_without_snow_or_ice sounds like a good starting point. Similarly we can discuss new area types for lakes with or without snow and/or ice. The key thing with all of these, as with standard names, is to describe them clearly. Where categories sound similar, or perhaps overlap, we need to be very clear about what is included or excluded in each area_type.
>
> Best wishes,
> Alison
>
> ------
> Alison Pamment Tel: +44 1235 778065
> NCAS/Centre for Environmental Data Archival Email: alison.pamment at stfc.ac.uk<mailto:alison.pamment at stfc.ac.uk>
> STFC Rutherford Appleton Laboratory
> R25, 2.22
> Harwell Oxford, Didcot, OX11 0QX, U.K.
>
>
> -----Original Message-----
> From: CF-metadata
> <cf-metadata-bounces at cgd.ucar.edu><mailto:cf-metadata-bounces at cgd.ucar
> .edu> On Behalf Of Martin Juckes - UKRI STFC
> Sent: 17 October 2018 12:44
> To: Taylor, Karl E. <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov>;
> cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Hello All,
>
>
> I agree with Karl's suggestion that it is useful to mention Greenland and Antarctica to clarify the intended meaning of "ice_sheet", and also with with Jonathan point that there needs to be a caveat (perhaps "present era", rather than "modern world" -- the latter is often used to describe a much shorter timescale than we want here).
>
>
> The CMIP approach to dividing the world is a little different from the approach Evan : the term "land_ice" has been introduced long ago and includes floating ice shelves. This could be described as a process driven approach: "land_ice" includes ice formed on land which has moved out to sea and has very different characteristics to "sea_ice", which is ice that has formed at sea.
>
>
> In CMIP6 "land" is interpreted as including floating ice shelves when it refers to the surface. In CMIP5 the models did not include a physical representation of floating ice shelves, so areas such as the Ross Sea would generally be represented as grounded ice sheets, I believe. For CMIP6, we did discuss restricting "land" to exclude floating ice shelves and introducing a new area type for the broader meaning, but in the end opted for continuity with CMIP5. "land" is also taken to include lakes -- the fact that we have a small number of lakes and inland seas resolved in CMIP models is not yet reflected in the area types.
>
>
> Consequently, Evan's requirements will need some new area types which will need to be named carefully to avoid confusion with existing ones.
>
> "ice_on_land" appears to have been introduced following a discussion of LS3MIP variables, one of which was originally an albedo of ice and snow on land but later got changed to an albedo of snow on land, hence this area type is not used.
>
> regards,
> Martin
> ________________________________
> From: CF-metadata
> <cf-metadata-bounces at cgd.ucar.edu><mailto:cf-metadata-bounces at cgd.ucar
> .edu> on behalf of Taylor, Karl E.
> <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov>
> Sent: 17 October 2018 05:38
> To: cf-metadata at cgd.ucar.edu<mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet/land_ice confusion
>
> Hi all,
>
> In CMIP5 only one of the three terms under discussion here was used:
> "land_ice" (in the standard_name "land_ice_area_fraction"), which was described as "fraction of grid cell occupied by "permanent" ice (i.e., glaciers)." This was a "fixed" (time-independent) field.
>
> As far as I can tell, "ice_on_land" isn't needed by CMIP6 (and it wasn't needed or used in CMIP5). I don't know (or have forgotten) what led it to be introduced as a valid surface type.
>
> best regards,
> Karl
>
> On 10/14/18 7:30 AM, Jonathan Gregory wrote:
>
>
> Reposting this, which didn't get to the list.
>
> Dear Karl, Sophie, Alison
>
> If we define ice_sheet to mean those of Greenland and Antarctica, it
> won't be applicable for palaeoclimate, so I think it's too
> restrictive. Although it's a continuum, there is a distinction between "ice sheet" and "glacier"
> that refers to size, with "ice-cap" being in the middle (and not used
> in IPCC to make things simpler). Ice sheets are big enough to bury the
> bedrock topography, so that the surface shape is determined by mass
> balance and dynamics. Glaciers are smaller, and confined within
> bedrock topography, which strongly influences their shape.
>
> If we want to mention Greenland and Antarctica explicitly, it would be
> a good idea to say "for example, in the modern world".
>
> No doubt it was discussed and I have forgotten, but being confronted
> with it now, I feel rather uncomfortable about there being distinct
> area_types of land_ice and ice_on_land. These types are not
> self-describing, in that the difference in wording does not convey anything about the difference in meaning.
>
> When and why was ice_on_land introduced?
>
> Best wishes
>
> Jonathan
>
> ----- Forwarded message from Karl Taylor
> <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov> -----
>
>
>
> Date: Tue, 9 Oct 2018 11:44:53 -0700
> From: Karl Taylor <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov>
> To: "Nowicki, Sophie (GSFC-6150)" <sophie.nowicki at nasa.gov><mailto:sophie.nowicki at nasa.gov>,
> "cf-metadata at cgd.ucar.edu"<mailto:cf-metadata at cgd.ucar.edu>
> <cf-metadata at cgd.ucar.edu><mailto:cf-metadata at cgd.ucar.edu>
> CC: Jonathan Gregory
> <j.m.gregory at reading.ac.uk><mailto:j.m.gregory at reading.ac.uk>
> Subject: Re: ice_sheet/land_ice confusion
> User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:52.0)
> Gecko/20100101 Thunderbird/52.9.1
>
> Thanks, Sophie, for your quick response. Given your clarification,
> perhaps we might replace the description of ice_sheet, which currently
> reads:
>
> > ice_sheet: An area type of "ice sheet" indicates where ice sheets are
> > present. It includes both grounded ice sheets resting over bedrock and
> > ice shelves flowing over the ocean, but excludes ice-caps and glaciers
> > (in contrast to land_ice, which includes all components).
>
> with this description:
>
> ice_sheet: An area type of "ice_sheet" indicates where the Greenland
> and Antarctic ice sheets are present. It includes both the grounded
> portion of those ice sheets (i.e., the portion resting on bedrock
> either above or below sea level) and the portion that is floating as
> ice shelves. It excludes all other ice on land (in contrast to
> land_ice, which includes, for example, small mountain glaciers and in
> contrast to ice_on_land, which is comprehensively inclusive of all
> types of ice on land).
>
> Also I think it should be clarified whether "snow" is considered to be
> "ice_on_land". If not, I think the descriptive phrase "any other ice
> on a land surface" should be modified to read "any other ice on a land
> surface (except snow)".
>
> Best regards,
> Karl
>
>
>
> On 10/9/18 11:03 AM, Nowicki, Sophie (GSFC-6150) wrote:
>
>
> Hi Karl,
>
> I am responding to your question about ice_sheet/land_ice (CF-metadata Digest, Message 2, Vol 186, Issue11), and deleted the other topics from the thread.
>
> ?ice_sheet would be the Greenland and Antarctic ice sheets. It contains both the grounded_ice_sheet (part of the ice sheet flowing over bedrock, and you are technically right that an ice sheet is a combination of many many glaciers) and floating_ice_shelf (the part that only flows on water).
>
> land_ice is much bigger as it includes the polar ice sheets, glaciers in non-polar regions (glaciers are considered small body of ice: for example in the Alps, or the US), and the small ice caps. The ice caps are also a large combinations of glaciers, but too small to be considered an ice sheets. For example the Svartissen Ice Cap in northern Norway.
>
> For ISMIP6, we are interested in ice_sheet, but some climate models may also include glaciers and ice caps (which ISMIP6 does not care about). Hence the use of both ice_sheet and land_ice in the ISMIP6 protocol (and I cant recall if land_ice was already present in CMIP5, but I think that it was).
>
> I don't know the origin of ice_on_land.
>
> Jonathan: please help me make my answers less confusing...
>
> I hope that this helps,
>
> Sophie
> Message: 2
> Date: Tue, 9 Oct 2018 17:19:36 +0000
> From: "Taylor, Karl E." <taylor13 at llnl.gov><mailto:taylor13 at llnl.gov>
> To: "cf-metadata at cgd.ucar.edu"<mailto:cf-metadata at cgd.ucar.edu> <cf-metadata at cgd.ucar.edu><mailto:cf-metadata at cgd.ucar.edu>
> Subject: Re: [CF-metadata] ice_sheet / land_ice confusion
> Message-ID: <ec366da6-0f45-0c3a-0ebe-d7b20f7cfb55 at llnl.gov><mailto:ec366da6-0f45-0c3a-0ebe-d7b20f7cfb55 at llnl.gov>
> Content-Type: text/plain; charset="utf-8"
> HI all,
> Can anyone provide any guidance on the difference between ice_sheet and
> land_ice (see below)?? It has a bearing on metadata to be stored with
> CMIP6 model output.
> thanks and best regards,
> Karl
> On 10/4/18 10:29 AM, Taylor, Karl E. wrote:
> > Hi all,
> >
> > I think there might be a mistake in the descriptions of "ice_sheet"
> > and/or "land_ice" in the "area type" table at
> > http://cfconventions.org/Data/area-type-table/current/build/area-type-table.html
> > .
> >
> > I find there the following definitions:
> >
> > ice_sheet: An area type of "ice sheet" indicates where ice sheets are
> > present. It includes both grounded ice sheets resting over bedrock and
> > ice shelves flowing over the ocean, but excludes ice-caps and glaciers
> > (in contrast to land_ice, which includes all components).
> >
> > land_ice: "Land ice" means glaciers, ice-caps, grounded ice sheets
> > resting on bedrock and floating ice-shelves.
> >
> > ice_on_land: The area type "ice_on_land" means ice in glaciers, ice
> > caps, grounded ice sheets (grounded and floating shelves), river and
> > lake ice, and any other ice on a land surface, such as frozen flood
> > water. This is distinct from the area type 'land ice' which has a
> > narrower definition.
> >
> > Are "ice-caps" and "glaciers" really excluded from "ice_sheet".? I would
> > have thought that "ice-cap" would be an ice_sheet located over a pole
> > (or something to that effect).? And i thought ice_sheets were just big
> > glaciers.
> >
> > ice_on_land is pretty clearly any frozen water, except sea ice,
> > icebergs, and ice particles in clouds, that is exposed to the atmosphere.
> >
> > So, I guess I'm trying to understand the difference between ice_sheet
> > and land_ice, and why do we need both of these?
> >
> > thanks and best regards,
> > Karl
> End of CF-metadata Digest, Vol 186, Issue 11
> ********************************************
>
>
>
> ----- End forwarded message -----
>
> ----- End forwarded message -----
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
>
> ----- End forwarded message -----
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
>
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu<mailto:CF-metadata at cgd.ucar.edu>
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
>
>
----- End forwarded message -----
_______________________________________________
CF-metadata mailing list
CF-metadata at cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
_______________________________________________
CF-metadata mailing list
CF-metadata at cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
Received on Wed Oct 24 2018 - 07:02:00 BST
