The behaviour of large-scale model cloud schemes under idealized forcing scenarios

The behaviour of widely used large-scale cloud schemes is assessed in the idealized framework of a single grid box forced by uniform warming or cooling. The general advantages and disadvantages Of using diagnostic and prognostic schemes are compared, as are characteristics of observational relationships and specific parametrization schemes used in general-circulation models. It is noted that all parametrization schemes fail to produce, in some way, the Correct behaviour when the grid box is forced and initial conditions altered. The reasons for this fundamental result are discussed. Closures for prognostic cloud schemes under uniform forcing are investigated, and a closure developed which closely meets the theoretically ideal behaviour.