There are only a few variables where the cell_methods includes for the horizontal dimension a phrase of the form "where type1 over type2". These should be reviewed. In particular,
Is the cell methods correct for sisnthick? It has cell_methods "area: mean where snow over sea_ice area: time: mean where sea_ice" I can't tell how this variable is supposed to be calculated. I would think that you wouldn't want to count snow on land or land ice, but applying the first "where snow" directive would include that.
For variables like cropFrac, the cell methods is "area mean: where land over all_area_types time: mean". I would think you could integrate the fraction of crop over the entire cell area (to obtain the total area of baresoil or crop in the cell), and then divide by the entire area over which you integrated (i.e., the cell area). Couldn't you therefore simplify the cell_methods to "area: mean time: mean"? or even "area time: mean"?
There are only a few variables where the cell_methods includes for the horizontal dimension a phrase of the form "where type1 over type2". These should be reviewed. In particular,