taylor13
commented
6 months ago For CMIP6, the requested variable has the standard name "sea_ice_area_fraction" (not concentration), and it was supposed to be reported as a fraction of the grid cell (with a variable named siconca) and also as a fraction of the sea area (with a variable named siconc). These two variables have different cell methods (but the same standard_name): siconca: cell_methods = "area: time: mean" siconc: cell_methods = "area: mean where sea time: mean"
I agree that the description of sea_ice_area_fraction needs revision.
Bruce-T-T-Calvert
The CF definition of sea_ice_area_fraction is as follows:
"Area fraction" is the fraction of a grid cell's horizontal area that has some characteristic of interest. It is evaluated as the area of interest divided by the grid cell area. It may be expressed as a fraction, a percentage, or any other dimensionless representation of a fraction. Sea ice area fraction is area of the sea surface occupied by sea ice. It is also called "sea ice concentration".
This definition is confusing for grid cells that contain both land and sea ice. The first two sentences suggest that the denominator of fraction includes the land area. The fourth sentence has a typo, as a fraction is not an area. The fifth sentence relates the definition to sea ice concentration. However, the AR6 definition of sea ice concentration would exclude the land area from the denominator of the fraction. Here is the AR6 definition: Sea ice concentration is the fraction of the ocean covered by ice.
It should be noted that AR6 does not have a definition of "sea ice area fraction", while there is no CF definition of "sea_ice_concentration".
To make things more confusing, many sea ice concentration datasets will use the standard name sea_ice_area_fraction in their netcdf files. Here are some examples: HadISST2 G10010v2 G02202v4 OSI-450-a
I also noticed that model output for CCSM4 available at earthsystemgrid.org uses sea_ice_area_fraction as the standard name for the sea ice concentration variable; I suspect this is used across climate models.
All five above cases define their land masks and grid cells in such a way that there are no grid cells that contain both land and sea ice, which means that the treatment of land area in the denominator is not relevant for the calculation of the fraction.
However, more generally, for grid cells that contain both land and sea ice, should sea_ice_area_fraction include or not the land fraction? Is the AR6 definition of sea ice concentration correct? It would help if the definition available at CFconventions.org were modified to clarify things. If the definitions of sea ice concentration and sea_ice_area_fraction are not the same, then it could be helpful if a new standard name of sea_ice_concentration were introduced.