[CF-metadata] CMIP6 Sea Ice MIP: Ice thickness

From: martin.juckes at stfc.ac.uk <martin.juckes>
Date: Tue, 2 Aug 2016 11:37:04 +0000

Dear Dirk,

OK. For the snow thickness, do you want a weighted time mean? For example, if there is a 1m snow layer over 50% of a grid cell for half the month and 3m over 25% for the 2nd half of the month (the rest of the cell being snow free), do you want a weighted time mean of 1.6667m or a simple mean of 2m?

regards,
Martin
________________________________________
From: Dirk Notz [dirk.notz at mpimet.mpg.de]
Sent: 02 August 2016 11:59
To: Juckes, Martin (STFC,RAL,RALSP); j.m.gregory at reading.ac.uk; Pamment, Alison (STFC,RAL,RALSP)
Cc: cf-metadata at cgd.ucar.edu
Subject: Re: [CF-metadata] CMIP6 Sea Ice MIP: Ice thickness

Dear Martin,

> It looks as though you have resolved all the issues nicely, but I have one concern about Jonathan's suggestion for dealing with the thickness of ice floating in melt ponds on the surface of sea ice. The suggestion is to use the existing standard name "floating_ice_thickness" with a cell methods term of the form "area: mean where sea_ice_melt_pond". My concern is that there is some ambiguity here between the thickness of the ice floating in the melt pond and the thickness of the sea ice underneath the melt pond, which is floating in the sea. This might be resolved if "floating_ice_thickness" was to be defined in such a way as to exclude sea ice, but the current definition does not do this. (It states that '"Floating ice" means any ice that is floating on water, e.g. on a sea or lake surface.')

This is an important point, thanks for raising it. It might then be
necessary to indeed introduce a new variable as initially suggested to
avoid this ambiguity. This variable was initially called
"thickness_of_sea_ice_melt_pond_refrozen_ice ".

> I have a 2nd comment about the suggested area type of snow_covered_sea_ice: this has been proposed for use with variables which represent the amount of snow on sea ice (m) and the heat content of that snow (J m-2). In both cases these are quantities which can be considered as zero on snow-free sea ice, in which case there is no need to mask them by the area of snow cover (I believe Jonathan has advanced this argument in another thread recently). In this case, amount of snow on sea ice would just be represented by "surface_snow_thickness" with cell methods "area: mean where sea_ice over all_area_types" (or "area: mean where sea_ice over sea").

This is certainly true for the heat content. For snow thickness,
however, we would like to be able to consider partial snow coverage. We
would like to record the thickness of the actual snow, not the average
thickness including those parts of sea ice that are ice free. Hence, for
snow thickness, I believe that the area type "snow_covered_sea_ice"
should still be applied.

Thanks in any case for your helpful feedback,

best,

 Dirk
Received on Tue Aug 02 2016 - 05:37:04 BST