[CF-metadata] CF-1.0 registration of new names for SST

[Jonathan Gregory]

Dear Frank

> > By the way, "sea_water_temperature with a depth attribute" is fine;
> 
> does this also account for the cases where the temperature is 
> the (weighted?) mean value of the first 5, 10, 100, or 1000 
> centimeters?

That would be descripted by an entry in cell_methods indicating a mean over
the vertical coordinate, with the thickness of the layer being indicated by the
bounds of the vertical coordinate.

Cheers

Jonathan