Reply to Bob Glahn's "Comments on 'The value of a variable resolution approach to numerical weather prediction"'